diff --git a/cts/scheduler/8-am-then-bm-a-migrating-b-stopping.summary b/cts/scheduler/8-am-then-bm-a-migrating-b-stopping.summary index a5120dc215..f49be0a2f7 100644 --- a/cts/scheduler/8-am-then-bm-a-migrating-b-stopping.summary +++ b/cts/scheduler/8-am-then-bm-a-migrating-b-stopping.summary @@ -1,26 +1,26 @@ -1 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 2 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ 18node1 18node2 18node3 ] A (ocf::heartbeat:Dummy): Started 18node1 B (ocf::heartbeat:Dummy): Started 18node2 (disabled) Transition Summary: * Migrate A ( 18node1 -> 18node2 ) * Stop B ( 18node2 ) due to node availability Executing cluster transition: * Resource action: B stop on 18node2 * Resource action: A migrate_to on 18node1 * Resource action: A migrate_from on 18node2 * Resource action: A stop on 18node1 * Pseudo action: A_start_0 * Resource action: A monitor=60000 on 18node2 Revised cluster status: Online: [ 18node1 18node2 18node3 ] A (ocf::heartbeat:Dummy): Started 18node2 B (ocf::heartbeat:Dummy): Stopped (disabled) diff --git a/cts/scheduler/9-am-then-bm-b-migrating-a-stopping.summary b/cts/scheduler/9-am-then-bm-b-migrating-a-stopping.summary index c06ee69d45..6ee493982a 100644 --- a/cts/scheduler/9-am-then-bm-b-migrating-a-stopping.summary +++ b/cts/scheduler/9-am-then-bm-b-migrating-a-stopping.summary @@ -1,22 +1,22 @@ -1 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 2 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ 18node1 18node2 18node3 ] A (ocf::heartbeat:Dummy): Started 18node1 (disabled) B (ocf::heartbeat:Dummy): Started 18node2 Transition Summary: * Stop A ( 18node1 ) due to node availability * Stop B ( 18node2 ) due to unrunnable A start Executing cluster transition: * Resource action: B stop on 18node2 * Resource action: A stop on 18node1 Revised cluster status: Online: [ 18node1 18node2 18node3 ] A (ocf::heartbeat:Dummy): Stopped (disabled) B (ocf::heartbeat:Dummy): Stopped diff --git a/cts/scheduler/asymmetrical-order-move.summary b/cts/scheduler/asymmetrical-order-move.summary index 9c0ff74ab1..5ca5740da2 100644 --- a/cts/scheduler/asymmetrical-order-move.summary +++ b/cts/scheduler/asymmetrical-order-move.summary @@ -1,24 +1,24 @@ Using the original execution date of: 2016-04-28 11:50:29Z -1 of 3 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 3 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ sle12sp2-1 sle12sp2-2 ] st_sbd (stonith:external/sbd): Started sle12sp2-2 dummy1 (ocf::pacemaker:Dummy): Stopped (disabled) dummy2 (ocf::pacemaker:Dummy): Started sle12sp2-1 Transition Summary: * Stop dummy2 ( sle12sp2-1 ) due to unrunnable dummy1 start Executing cluster transition: * Resource action: dummy2 stop on sle12sp2-1 Using the original execution date of: 2016-04-28 11:50:29Z Revised cluster status: Online: [ sle12sp2-1 sle12sp2-2 ] st_sbd (stonith:external/sbd): Started sle12sp2-2 dummy1 (ocf::pacemaker:Dummy): Stopped (disabled) dummy2 (ocf::pacemaker:Dummy): Stopped diff --git a/cts/scheduler/asymmetrical-order-restart.summary b/cts/scheduler/asymmetrical-order-restart.summary index 4ef8b8272b..b6bdcf244b 100644 --- a/cts/scheduler/asymmetrical-order-restart.summary +++ b/cts/scheduler/asymmetrical-order-restart.summary @@ -1,28 +1,28 @@ Using the original execution date of: 2018-08-09 18:55:41Z -1 of 3 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 3 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ cesr105-p16 cesr109-p16 ] cesr104ipmi (stonith:fence_ipmilan): Started cesr105-p16 sleep_a (ocf::classe:anything): Stopped (disabled) sleep_b (ocf::classe:anything): FAILED cesr109-p16 Transition Summary: * Restart cesr104ipmi ( cesr105-p16 ) due to resource definition change * Stop sleep_b ( cesr109-p16 ) due to unrunnable sleep_a start Executing cluster transition: * Resource action: cesr104ipmi stop on cesr105-p16 * Resource action: cesr104ipmi start on cesr105-p16 * Resource action: cesr104ipmi monitor=60000 on cesr105-p16 * Resource action: sleep_b stop on cesr109-p16 Using the original execution date of: 2018-08-09 18:55:41Z Revised cluster status: Online: [ cesr105-p16 cesr109-p16 ] cesr104ipmi (stonith:fence_ipmilan): Started cesr105-p16 sleep_a (ocf::classe:anything): Stopped (disabled) sleep_b (ocf::classe:anything): Stopped diff --git a/cts/scheduler/bug-1718.summary b/cts/scheduler/bug-1718.summary index 8d25c97e26..0b74a157f4 100644 --- a/cts/scheduler/bug-1718.summary +++ b/cts/scheduler/bug-1718.summary @@ -1,41 +1,41 @@ -2 of 5 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 5 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ biggame.ds9 heartbeat.ds9 ops.ds9 ] OFFLINE: [ defiant.ds9 warbird.ds9 ] Resource Group: Web_Group Apache_IP (ocf::heartbeat:IPaddr): Started heartbeat.ds9 resource_IP2 (ocf::heartbeat:IPaddr): Stopped (disabled) resource_dummyweb (ocf::heartbeat:Dummy): Stopped Resource Group: group_fUN resource_IP3 (ocf::heartbeat:IPaddr): Started ops.ds9 resource_dummy (ocf::heartbeat:Dummy): Started ops.ds9 Transition Summary: * Stop resource_IP3 ( ops.ds9 ) due to unrunnable Web_Group running * Stop resource_dummy ( ops.ds9 ) due to required resource_IP3 start Executing cluster transition: * Pseudo action: group_fUN_stop_0 * Resource action: resource_dummy stop on ops.ds9 * Resource action: OpenVPN_IP delete on ops.ds9 * Resource action: OpenVPN_IP delete on heartbeat.ds9 * Resource action: Apache delete on biggame.ds9 * Resource action: Apache delete on ops.ds9 * Resource action: Apache delete on heartbeat.ds9 * Resource action: resource_IP3 stop on ops.ds9 * Pseudo action: group_fUN_stopped_0 Revised cluster status: Online: [ biggame.ds9 heartbeat.ds9 ops.ds9 ] OFFLINE: [ defiant.ds9 warbird.ds9 ] Resource Group: Web_Group Apache_IP (ocf::heartbeat:IPaddr): Started heartbeat.ds9 resource_IP2 (ocf::heartbeat:IPaddr): Stopped (disabled) resource_dummyweb (ocf::heartbeat:Dummy): Stopped Resource Group: group_fUN resource_IP3 (ocf::heartbeat:IPaddr): Stopped resource_dummy (ocf::heartbeat:Dummy): Stopped diff --git a/cts/scheduler/bug-5014-A-stop-B-started.summary b/cts/scheduler/bug-5014-A-stop-B-started.summary index 765f3924e2..a122602f9b 100644 --- a/cts/scheduler/bug-5014-A-stop-B-started.summary +++ b/cts/scheduler/bug-5014-A-stop-B-started.summary @@ -1,20 +1,20 @@ -1 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 2 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] ClusterIP (ocf::heartbeat:IPaddr2): Started fc16-builder (disabled) ClusterIP2 (ocf::heartbeat:IPaddr2): Started fc16-builder Transition Summary: * Stop ClusterIP ( fc16-builder ) due to node availability Executing cluster transition: * Resource action: ClusterIP stop on fc16-builder Revised cluster status: Online: [ fc16-builder ] ClusterIP (ocf::heartbeat:IPaddr2): Stopped (disabled) ClusterIP2 (ocf::heartbeat:IPaddr2): Started fc16-builder diff --git a/cts/scheduler/bug-5014-A-stopped-B-stopped.summary b/cts/scheduler/bug-5014-A-stopped-B-stopped.summary index b5a83f6b28..280f514f01 100644 --- a/cts/scheduler/bug-5014-A-stopped-B-stopped.summary +++ b/cts/scheduler/bug-5014-A-stopped-B-stopped.summary @@ -1,21 +1,21 @@ -1 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 2 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] ClusterIP (ocf::heartbeat:IPaddr2): Stopped (disabled) ClusterIP2 (ocf::heartbeat:IPaddr2): Stopped Transition Summary: * Start ClusterIP2 ( fc16-builder ) due to unrunnable ClusterIP start (blocked) Executing cluster transition: * Resource action: ClusterIP monitor on fc16-builder * Resource action: ClusterIP2 monitor on fc16-builder Revised cluster status: Online: [ fc16-builder ] ClusterIP (ocf::heartbeat:IPaddr2): Stopped (disabled) ClusterIP2 (ocf::heartbeat:IPaddr2): Stopped diff --git a/cts/scheduler/bug-5014-CLONE-A-stop-B-started.summary b/cts/scheduler/bug-5014-CLONE-A-stop-B-started.summary index 9efb43813f..0f3096cd84 100644 --- a/cts/scheduler/bug-5014-CLONE-A-stop-B-started.summary +++ b/cts/scheduler/bug-5014-CLONE-A-stop-B-started.summary @@ -1,26 +1,26 @@ -1 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 2 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] Clone Set: clone1 [ClusterIP] Started: [ fc16-builder ] Clone Set: clone2 [ClusterIP2] Started: [ fc16-builder ] Transition Summary: * Stop ClusterIP:0 ( fc16-builder ) due to node availability Executing cluster transition: * Pseudo action: clone1_stop_0 * Resource action: ClusterIP:0 stop on fc16-builder * Pseudo action: clone1_stopped_0 Revised cluster status: Online: [ fc16-builder ] Clone Set: clone1 [ClusterIP] Stopped (disabled): [ fc16-builder ] Clone Set: clone2 [ClusterIP2] Started: [ fc16-builder ] diff --git a/cts/scheduler/bug-5014-CthenAthenB-C-stopped.summary b/cts/scheduler/bug-5014-CthenAthenB-C-stopped.summary index 57068a673c..96569d51c2 100644 --- a/cts/scheduler/bug-5014-CthenAthenB-C-stopped.summary +++ b/cts/scheduler/bug-5014-CthenAthenB-C-stopped.summary @@ -1,25 +1,25 @@ -1 of 3 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 3 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] ClusterIP (ocf::heartbeat:IPaddr2): Stopped ClusterIP2 (ocf::heartbeat:IPaddr2): Stopped ClusterIP3 (ocf::heartbeat:IPaddr2): Stopped (disabled) Transition Summary: * Start ClusterIP ( fc16-builder ) due to unrunnable ClusterIP3 start (blocked) * Start ClusterIP2 ( fc16-builder ) due to unrunnable ClusterIP start (blocked) Executing cluster transition: * Resource action: ClusterIP monitor on fc16-builder * Resource action: ClusterIP2 monitor on fc16-builder * Resource action: ClusterIP3 monitor on fc16-builder Revised cluster status: Online: [ fc16-builder ] ClusterIP (ocf::heartbeat:IPaddr2): Stopped ClusterIP2 (ocf::heartbeat:IPaddr2): Stopped ClusterIP3 (ocf::heartbeat:IPaddr2): Stopped (disabled) diff --git a/cts/scheduler/bug-5014-GROUP-A-stopped-B-started.summary b/cts/scheduler/bug-5014-GROUP-A-stopped-B-started.summary index e363718bc0..eacffdc667 100644 --- a/cts/scheduler/bug-5014-GROUP-A-stopped-B-started.summary +++ b/cts/scheduler/bug-5014-GROUP-A-stopped-B-started.summary @@ -1,26 +1,26 @@ -2 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 2 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] Resource Group: group1 ClusterIP (ocf::heartbeat:IPaddr2): Started fc16-builder (disabled) Resource Group: group2 ClusterIP2 (ocf::heartbeat:IPaddr2): Started fc16-builder Transition Summary: * Stop ClusterIP ( fc16-builder ) due to node availability Executing cluster transition: * Pseudo action: group1_stop_0 * Resource action: ClusterIP stop on fc16-builder * Pseudo action: group1_stopped_0 Revised cluster status: Online: [ fc16-builder ] Resource Group: group1 ClusterIP (ocf::heartbeat:IPaddr2): Stopped (disabled) Resource Group: group2 ClusterIP2 (ocf::heartbeat:IPaddr2): Started fc16-builder diff --git a/cts/scheduler/bug-5014-GROUP-A-stopped-B-stopped.summary b/cts/scheduler/bug-5014-GROUP-A-stopped-B-stopped.summary index c6c88faf39..f61161f45b 100644 --- a/cts/scheduler/bug-5014-GROUP-A-stopped-B-stopped.summary +++ b/cts/scheduler/bug-5014-GROUP-A-stopped-B-stopped.summary @@ -1,23 +1,23 @@ -2 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 2 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] Resource Group: group1 ClusterIP (ocf::heartbeat:IPaddr2): Stopped (disabled) Resource Group: group2 ClusterIP2 (ocf::heartbeat:IPaddr2): Stopped Transition Summary: * Start ClusterIP2 ( fc16-builder ) due to unrunnable group1 running (blocked) Executing cluster transition: Revised cluster status: Online: [ fc16-builder ] Resource Group: group1 ClusterIP (ocf::heartbeat:IPaddr2): Stopped (disabled) Resource Group: group2 ClusterIP2 (ocf::heartbeat:IPaddr2): Stopped diff --git a/cts/scheduler/bug-5014-ordered-set-symmetrical-false.summary b/cts/scheduler/bug-5014-ordered-set-symmetrical-false.summary index 9d469c8556..cad1afe0e1 100644 --- a/cts/scheduler/bug-5014-ordered-set-symmetrical-false.summary +++ b/cts/scheduler/bug-5014-ordered-set-symmetrical-false.summary @@ -1,24 +1,24 @@ -1 of 3 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 3 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Started fc16-builder B (ocf::pacemaker:Dummy): Started fc16-builder C (ocf::pacemaker:Dummy): Started fc16-builder (disabled) Transition Summary: * Stop C ( fc16-builder ) due to node availability Executing cluster transition: * Resource action: C stop on fc16-builder Revised cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Started fc16-builder B (ocf::pacemaker:Dummy): Started fc16-builder C (ocf::pacemaker:Dummy): Stopped (disabled) diff --git a/cts/scheduler/bug-5014-ordered-set-symmetrical-true.summary b/cts/scheduler/bug-5014-ordered-set-symmetrical-true.summary index 516223e544..0e02f167aa 100644 --- a/cts/scheduler/bug-5014-ordered-set-symmetrical-true.summary +++ b/cts/scheduler/bug-5014-ordered-set-symmetrical-true.summary @@ -1,26 +1,26 @@ -1 of 3 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 3 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Started fc16-builder B (ocf::pacemaker:Dummy): Started fc16-builder C (ocf::pacemaker:Dummy): Started fc16-builder (disabled) Transition Summary: * Stop A ( fc16-builder ) due to required C start * Stop C ( fc16-builder ) due to node availability Executing cluster transition: * Resource action: A stop on fc16-builder * Resource action: C stop on fc16-builder Revised cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Stopped B (ocf::pacemaker:Dummy): Started fc16-builder C (ocf::pacemaker:Dummy): Stopped (disabled) diff --git a/cts/scheduler/bug-5028-bottom.summary b/cts/scheduler/bug-5028-bottom.summary index 11c719bebd..c7329022c9 100644 --- a/cts/scheduler/bug-5028-bottom.summary +++ b/cts/scheduler/bug-5028-bottom.summary @@ -1,23 +1,24 @@ +0 of 2 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ bl460g6a bl460g6b ] Resource Group: dummy-g dummy01 (ocf::heartbeat:Dummy): FAILED bl460g6a (blocked) dummy02 (ocf::heartbeat:Dummy-stop-NG): Started bl460g6a Transition Summary: * Shutdown bl460g6a * Stop dummy02 ( bl460g6a ) due to node availability Executing cluster transition: * Pseudo action: dummy-g_stop_0 * Resource action: dummy02 stop on bl460g6a Revised cluster status: Online: [ bl460g6a bl460g6b ] Resource Group: dummy-g dummy01 (ocf::heartbeat:Dummy): FAILED bl460g6a (blocked) dummy02 (ocf::heartbeat:Dummy-stop-NG): Stopped diff --git a/cts/scheduler/bug-5028-detach.summary b/cts/scheduler/bug-5028-detach.summary index 691a7fd3a9..9930c5bb3d 100644 --- a/cts/scheduler/bug-5028-detach.summary +++ b/cts/scheduler/bug-5028-detach.summary @@ -1,24 +1,25 @@ *** Resource management is DISABLED *** The cluster will not attempt to start, stop or recover services +0 of 2 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ bl460g6a bl460g6b ] Resource Group: dummy-g dummy01 (ocf::heartbeat:Dummy): Started bl460g6a (unmanaged) dummy02 (ocf::heartbeat:Dummy-stop-NG): FAILED bl460g6a (blocked) Transition Summary: * Shutdown bl460g6a Executing cluster transition: * Cluster action: do_shutdown on bl460g6a Revised cluster status: Online: [ bl460g6a bl460g6b ] Resource Group: dummy-g dummy01 (ocf::heartbeat:Dummy): Started bl460g6a (unmanaged) dummy02 (ocf::heartbeat:Dummy-stop-NG): FAILED bl460g6a (blocked) diff --git a/cts/scheduler/bug-5028.summary b/cts/scheduler/bug-5028.summary index 5a08468b28..97fe35c6ce 100644 --- a/cts/scheduler/bug-5028.summary +++ b/cts/scheduler/bug-5028.summary @@ -1,23 +1,24 @@ +0 of 2 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ bl460g6a bl460g6b ] Resource Group: dummy-g dummy01 (ocf::heartbeat:Dummy): Started bl460g6a dummy02 (ocf::heartbeat:Dummy-stop-NG): FAILED bl460g6a (blocked) Transition Summary: * Shutdown bl460g6a * Stop dummy01 ( bl460g6a ) due to unrunnable dummy02 stop (blocked) Executing cluster transition: * Pseudo action: dummy-g_stop_0 * Pseudo action: dummy-g_start_0 Revised cluster status: Online: [ bl460g6a bl460g6b ] Resource Group: dummy-g dummy01 (ocf::heartbeat:Dummy): Started bl460g6a dummy02 (ocf::heartbeat:Dummy-stop-NG): FAILED bl460g6a (blocked) diff --git a/cts/scheduler/bug-5140-require-all-false.summary b/cts/scheduler/bug-5140-require-all-false.summary index 0c52cc4276..caf56e47af 100644 --- a/cts/scheduler/bug-5140-require-all-false.summary +++ b/cts/scheduler/bug-5140-require-all-false.summary @@ -1,80 +1,80 @@ -4 of 35 resources DISABLED and 0 BLOCKED from being started due to failures +4 of 35 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Node hex-1: standby Node hex-2: standby Node hex-3: OFFLINE (standby) fencing (stonith:external/sbd): Stopped Clone Set: baseclone [basegrp] Resource Group: basegrp:0 dlm (ocf::pacemaker:controld): Started hex-2 clvmd (ocf::lvm2:clvmd): Started hex-2 o2cb (ocf::ocfs2:o2cb): Started hex-2 vg1 (ocf::heartbeat:LVM): Stopped fs-ocfs-1 (ocf::heartbeat:Filesystem): Stopped Stopped: [ hex-1 hex-3 ] fs-xfs-1 (ocf::heartbeat:Filesystem): Stopped Clone Set: fs2 [fs-ocfs-2] Stopped: [ hex-1 hex-2 hex-3 ] Clone Set: ms-r0 [drbd-r0] (promotable) Stopped (disabled): [ hex-1 hex-2 hex-3 ] Clone Set: ms-r1 [drbd-r1] (promotable) Stopped (disabled): [ hex-1 hex-2 hex-3 ] Resource Group: md0-group md0 (ocf::heartbeat:Raid1): Stopped vg-md0 (ocf::heartbeat:LVM): Stopped fs-md0 (ocf::heartbeat:Filesystem): Stopped dummy1 (ocf::heartbeat:Delay): Stopped dummy3 (ocf::heartbeat:Delay): Stopped dummy4 (ocf::heartbeat:Delay): Stopped dummy5 (ocf::heartbeat:Delay): Stopped dummy6 (ocf::heartbeat:Delay): Stopped Resource Group: r0-group fs-r0 (ocf::heartbeat:Filesystem): Stopped dummy2 (ocf::heartbeat:Delay): Stopped cluster-md0 (ocf::heartbeat:Raid1): Stopped Transition Summary: * Stop dlm:0 ( hex-2 ) due to node availability * Stop clvmd:0 ( hex-2 ) due to node availability * Stop o2cb:0 ( hex-2 ) due to node availability Executing cluster transition: * Pseudo action: baseclone_stop_0 * Pseudo action: basegrp:0_stop_0 * Resource action: o2cb stop on hex-2 * Resource action: clvmd stop on hex-2 * Resource action: dlm stop on hex-2 * Pseudo action: basegrp:0_stopped_0 * Pseudo action: baseclone_stopped_0 Revised cluster status: Node hex-1: standby Node hex-2: standby Node hex-3: OFFLINE (standby) fencing (stonith:external/sbd): Stopped Clone Set: baseclone [basegrp] Stopped: [ hex-1 hex-2 hex-3 ] fs-xfs-1 (ocf::heartbeat:Filesystem): Stopped Clone Set: fs2 [fs-ocfs-2] Stopped: [ hex-1 hex-2 hex-3 ] Clone Set: ms-r0 [drbd-r0] (promotable) Stopped (disabled): [ hex-1 hex-2 hex-3 ] Clone Set: ms-r1 [drbd-r1] (promotable) Stopped (disabled): [ hex-1 hex-2 hex-3 ] Resource Group: md0-group md0 (ocf::heartbeat:Raid1): Stopped vg-md0 (ocf::heartbeat:LVM): Stopped fs-md0 (ocf::heartbeat:Filesystem): Stopped dummy1 (ocf::heartbeat:Delay): Stopped dummy3 (ocf::heartbeat:Delay): Stopped dummy4 (ocf::heartbeat:Delay): Stopped dummy5 (ocf::heartbeat:Delay): Stopped dummy6 (ocf::heartbeat:Delay): Stopped Resource Group: r0-group fs-r0 (ocf::heartbeat:Filesystem): Stopped dummy2 (ocf::heartbeat:Delay): Stopped cluster-md0 (ocf::heartbeat:Raid1): Stopped diff --git a/cts/scheduler/bug-5143-ms-shuffle.summary b/cts/scheduler/bug-5143-ms-shuffle.summary index 826dfbd075..568f73ce08 100644 --- a/cts/scheduler/bug-5143-ms-shuffle.summary +++ b/cts/scheduler/bug-5143-ms-shuffle.summary @@ -1,75 +1,75 @@ -2 of 34 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 34 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ hex-1 hex-2 hex-3 ] fencing (stonith:external/sbd): Started hex-1 Clone Set: baseclone [basegrp] Started: [ hex-1 hex-2 hex-3 ] fs-xfs-1 (ocf::heartbeat:Filesystem): Started hex-2 Clone Set: fs2 [fs-ocfs-2] Started: [ hex-1 hex-2 hex-3 ] Clone Set: ms-r0 [drbd-r0] (promotable) Masters: [ hex-1 ] Slaves: [ hex-2 ] Clone Set: ms-r1 [drbd-r1] (promotable) Slaves: [ hex-2 hex-3 ] Resource Group: md0-group md0 (ocf::heartbeat:Raid1): Started hex-3 vg-md0 (ocf::heartbeat:LVM): Started hex-3 fs-md0 (ocf::heartbeat:Filesystem): Started hex-3 dummy1 (ocf::heartbeat:Delay): Started hex-3 dummy3 (ocf::heartbeat:Delay): Started hex-1 dummy4 (ocf::heartbeat:Delay): Started hex-2 dummy5 (ocf::heartbeat:Delay): Started hex-1 dummy6 (ocf::heartbeat:Delay): Started hex-2 Resource Group: r0-group fs-r0 (ocf::heartbeat:Filesystem): Stopped (disabled) dummy2 (ocf::heartbeat:Delay): Stopped Transition Summary: * Promote drbd-r1:1 ( Slave -> Master hex-3 ) Executing cluster transition: * Pseudo action: ms-r1_pre_notify_promote_0 * Resource action: drbd-r1 notify on hex-2 * Resource action: drbd-r1 notify on hex-3 * Pseudo action: ms-r1_confirmed-pre_notify_promote_0 * Pseudo action: ms-r1_promote_0 * Resource action: drbd-r1 promote on hex-3 * Pseudo action: ms-r1_promoted_0 * Pseudo action: ms-r1_post_notify_promoted_0 * Resource action: drbd-r1 notify on hex-2 * Resource action: drbd-r1 notify on hex-3 * Pseudo action: ms-r1_confirmed-post_notify_promoted_0 * Resource action: drbd-r1 monitor=29000 on hex-2 * Resource action: drbd-r1 monitor=31000 on hex-3 Revised cluster status: Online: [ hex-1 hex-2 hex-3 ] fencing (stonith:external/sbd): Started hex-1 Clone Set: baseclone [basegrp] Started: [ hex-1 hex-2 hex-3 ] fs-xfs-1 (ocf::heartbeat:Filesystem): Started hex-2 Clone Set: fs2 [fs-ocfs-2] Started: [ hex-1 hex-2 hex-3 ] Clone Set: ms-r0 [drbd-r0] (promotable) Masters: [ hex-1 ] Slaves: [ hex-2 ] Clone Set: ms-r1 [drbd-r1] (promotable) Masters: [ hex-3 ] Slaves: [ hex-2 ] Resource Group: md0-group md0 (ocf::heartbeat:Raid1): Started hex-3 vg-md0 (ocf::heartbeat:LVM): Started hex-3 fs-md0 (ocf::heartbeat:Filesystem): Started hex-3 dummy1 (ocf::heartbeat:Delay): Started hex-3 dummy3 (ocf::heartbeat:Delay): Started hex-1 dummy4 (ocf::heartbeat:Delay): Started hex-2 dummy5 (ocf::heartbeat:Delay): Started hex-1 dummy6 (ocf::heartbeat:Delay): Started hex-2 Resource Group: r0-group fs-r0 (ocf::heartbeat:Filesystem): Stopped (disabled) dummy2 (ocf::heartbeat:Delay): Stopped diff --git a/cts/scheduler/bug-cl-5170.summary b/cts/scheduler/bug-cl-5170.summary index 3136c121c6..e6fa234992 100644 --- a/cts/scheduler/bug-cl-5170.summary +++ b/cts/scheduler/bug-cl-5170.summary @@ -1,33 +1,34 @@ +0 of 4 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Node TCS-1: OFFLINE (standby) Online: [ TCS-2 ] Resource Group: svc ip_trf (ocf::heartbeat:IPaddr2): Started TCS-2 ip_mgmt (ocf::heartbeat:IPaddr2): Started TCS-2 Clone Set: cl_tomcat_nms [d_tomcat_nms] d_tomcat_nms (ocf::ntc:tomcat): FAILED TCS-2 (blocked) Stopped: [ TCS-1 ] Transition Summary: * Stop ip_trf ( TCS-2 ) due to node availability * Stop ip_mgmt ( TCS-2 ) due to node availability Executing cluster transition: * Pseudo action: svc_stop_0 * Resource action: ip_mgmt stop on TCS-2 * Resource action: ip_trf stop on TCS-2 * Pseudo action: svc_stopped_0 Revised cluster status: Node TCS-1: OFFLINE (standby) Online: [ TCS-2 ] Resource Group: svc ip_trf (ocf::heartbeat:IPaddr2): Stopped ip_mgmt (ocf::heartbeat:IPaddr2): Stopped Clone Set: cl_tomcat_nms [d_tomcat_nms] d_tomcat_nms (ocf::ntc:tomcat): FAILED TCS-2 (blocked) Stopped: [ TCS-1 ] diff --git a/cts/scheduler/bug-cl-5219.summary b/cts/scheduler/bug-cl-5219.summary index 97a66fb6ae..d8921c59f3 100644 --- a/cts/scheduler/bug-cl-5219.summary +++ b/cts/scheduler/bug-cl-5219.summary @@ -1,40 +1,40 @@ -1 of 9 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 9 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ ha1.test.anchor.net.au ha2.test.anchor.net.au ] child1-service (ocf::pacemaker:Dummy): Started ha2.test.anchor.net.au (disabled) child2-service (ocf::pacemaker:Dummy): Started ha2.test.anchor.net.au parent-service (ocf::pacemaker:Dummy): Started ha2.test.anchor.net.au Clone Set: child1 [stateful-child1] (promotable) Masters: [ ha2.test.anchor.net.au ] Slaves: [ ha1.test.anchor.net.au ] Clone Set: child2 [stateful-child2] (promotable) Masters: [ ha2.test.anchor.net.au ] Slaves: [ ha1.test.anchor.net.au ] Clone Set: parent [stateful-parent] (promotable) Masters: [ ha2.test.anchor.net.au ] Slaves: [ ha1.test.anchor.net.au ] Transition Summary: * Stop child1-service ( ha2.test.anchor.net.au ) due to node availability Executing cluster transition: * Resource action: child1-service stop on ha2.test.anchor.net.au Revised cluster status: Online: [ ha1.test.anchor.net.au ha2.test.anchor.net.au ] child1-service (ocf::pacemaker:Dummy): Stopped (disabled) child2-service (ocf::pacemaker:Dummy): Started ha2.test.anchor.net.au parent-service (ocf::pacemaker:Dummy): Started ha2.test.anchor.net.au Clone Set: child1 [stateful-child1] (promotable) Masters: [ ha2.test.anchor.net.au ] Slaves: [ ha1.test.anchor.net.au ] Clone Set: child2 [stateful-child2] (promotable) Masters: [ ha2.test.anchor.net.au ] Slaves: [ ha1.test.anchor.net.au ] Clone Set: parent [stateful-parent] (promotable) Masters: [ ha2.test.anchor.net.au ] Slaves: [ ha1.test.anchor.net.au ] diff --git a/cts/scheduler/bug-lf-2171.summary b/cts/scheduler/bug-lf-2171.summary index daf96bbff5..5a8be0e714 100644 --- a/cts/scheduler/bug-lf-2171.summary +++ b/cts/scheduler/bug-lf-2171.summary @@ -1,36 +1,36 @@ -3 of 4 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 4 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ xenserver1 xenserver2 ] Clone Set: cl_res_Dummy1 [res_Dummy1] Started: [ xenserver1 xenserver2 ] Resource Group: gr_Dummy res_Dummy2 (ocf::heartbeat:Dummy): Started xenserver1 res_Dummy3 (ocf::heartbeat:Dummy): Started xenserver1 Transition Summary: * Stop res_Dummy1:0 ( xenserver1 ) due to node availability * Stop res_Dummy1:1 ( xenserver2 ) due to node availability * Stop res_Dummy2 ( xenserver1 ) due to unrunnable cl_res_Dummy1 running * Stop res_Dummy3 ( xenserver1 ) due to unrunnable cl_res_Dummy1 running Executing cluster transition: * Pseudo action: gr_Dummy_stop_0 * Resource action: res_Dummy2 stop on xenserver1 * Resource action: res_Dummy3 stop on xenserver1 * Pseudo action: gr_Dummy_stopped_0 * Pseudo action: cl_res_Dummy1_stop_0 * Resource action: res_Dummy1:1 stop on xenserver1 * Resource action: res_Dummy1:0 stop on xenserver2 * Pseudo action: cl_res_Dummy1_stopped_0 Revised cluster status: Online: [ xenserver1 xenserver2 ] Clone Set: cl_res_Dummy1 [res_Dummy1] Stopped (disabled): [ xenserver1 xenserver2 ] Resource Group: gr_Dummy res_Dummy2 (ocf::heartbeat:Dummy): Stopped res_Dummy3 (ocf::heartbeat:Dummy): Stopped diff --git a/cts/scheduler/bug-lf-2358.summary b/cts/scheduler/bug-lf-2358.summary index 09ff14e01d..40db7956b3 100644 --- a/cts/scheduler/bug-lf-2358.summary +++ b/cts/scheduler/bug-lf-2358.summary @@ -1,65 +1,65 @@ -2 of 15 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 15 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ alice.demo bob.demo ] Clone Set: ms_drbd_nfsexport [res_drbd_nfsexport] (promotable) Stopped (disabled): [ alice.demo bob.demo ] Resource Group: rg_nfs res_fs_nfsexport (ocf::heartbeat:Filesystem): Stopped res_ip_nfs (ocf::heartbeat:IPaddr2): Stopped res_nfs (lsb:nfs): Stopped Resource Group: rg_mysql1 res_fs_mysql1 (ocf::heartbeat:Filesystem): Started bob.demo res_ip_mysql1 (ocf::heartbeat:IPaddr2): Started bob.demo res_mysql1 (ocf::heartbeat:mysql): Started bob.demo Clone Set: ms_drbd_mysql1 [res_drbd_mysql1] (promotable) Masters: [ bob.demo ] Stopped: [ alice.demo ] Clone Set: ms_drbd_mysql2 [res_drbd_mysql2] (promotable) Masters: [ alice.demo ] Slaves: [ bob.demo ] Resource Group: rg_mysql2 res_fs_mysql2 (ocf::heartbeat:Filesystem): Started alice.demo res_ip_mysql2 (ocf::heartbeat:IPaddr2): Started alice.demo res_mysql2 (ocf::heartbeat:mysql): Started alice.demo Transition Summary: * Start res_drbd_mysql1:1 ( alice.demo ) Executing cluster transition: * Pseudo action: ms_drbd_mysql1_pre_notify_start_0 * Resource action: res_drbd_mysql1:0 notify on bob.demo * Pseudo action: ms_drbd_mysql1_confirmed-pre_notify_start_0 * Pseudo action: ms_drbd_mysql1_start_0 * Resource action: res_drbd_mysql1:1 start on alice.demo * Pseudo action: ms_drbd_mysql1_running_0 * Pseudo action: ms_drbd_mysql1_post_notify_running_0 * Resource action: res_drbd_mysql1:0 notify on bob.demo * Resource action: res_drbd_mysql1:1 notify on alice.demo * Pseudo action: ms_drbd_mysql1_confirmed-post_notify_running_0 Revised cluster status: Online: [ alice.demo bob.demo ] Clone Set: ms_drbd_nfsexport [res_drbd_nfsexport] (promotable) Stopped (disabled): [ alice.demo bob.demo ] Resource Group: rg_nfs res_fs_nfsexport (ocf::heartbeat:Filesystem): Stopped res_ip_nfs (ocf::heartbeat:IPaddr2): Stopped res_nfs (lsb:nfs): Stopped Resource Group: rg_mysql1 res_fs_mysql1 (ocf::heartbeat:Filesystem): Started bob.demo res_ip_mysql1 (ocf::heartbeat:IPaddr2): Started bob.demo res_mysql1 (ocf::heartbeat:mysql): Started bob.demo Clone Set: ms_drbd_mysql1 [res_drbd_mysql1] (promotable) Masters: [ bob.demo ] Slaves: [ alice.demo ] Clone Set: ms_drbd_mysql2 [res_drbd_mysql2] (promotable) Masters: [ alice.demo ] Slaves: [ bob.demo ] Resource Group: rg_mysql2 res_fs_mysql2 (ocf::heartbeat:Filesystem): Started alice.demo res_ip_mysql2 (ocf::heartbeat:IPaddr2): Started alice.demo res_mysql2 (ocf::heartbeat:mysql): Started alice.demo diff --git a/cts/scheduler/bug-lf-2422.summary b/cts/scheduler/bug-lf-2422.summary index 4abb0443f2..e90deb3b70 100644 --- a/cts/scheduler/bug-lf-2422.summary +++ b/cts/scheduler/bug-lf-2422.summary @@ -1,80 +1,80 @@ -8 of 21 resources DISABLED and 0 BLOCKED from being started due to failures +4 of 21 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ qa-suse-1 qa-suse-2 qa-suse-3 qa-suse-4 ] sbd_stonith (stonith:external/sbd): Started qa-suse-2 Clone Set: c-o2stage [o2stage] Started: [ qa-suse-1 qa-suse-2 qa-suse-3 qa-suse-4 ] Clone Set: c-ocfs [ocfs] Started: [ qa-suse-1 qa-suse-2 qa-suse-3 qa-suse-4 ] Transition Summary: * Stop o2cb:0 ( qa-suse-1 ) due to node availability * Stop cmirror:0 ( qa-suse-1 ) due to node availability * Stop o2cb:1 ( qa-suse-4 ) due to node availability * Stop cmirror:1 ( qa-suse-4 ) due to node availability * Stop o2cb:2 ( qa-suse-3 ) due to node availability * Stop cmirror:2 ( qa-suse-3 ) due to node availability * Stop o2cb:3 ( qa-suse-2 ) due to node availability * Stop cmirror:3 ( qa-suse-2 ) due to node availability * Stop ocfs:0 ( qa-suse-1 ) due to node availability * Stop ocfs:1 ( qa-suse-4 ) due to node availability * Stop ocfs:2 ( qa-suse-3 ) due to node availability * Stop ocfs:3 ( qa-suse-2 ) due to node availability Executing cluster transition: * Resource action: sbd_stonith monitor=15000 on qa-suse-2 * Pseudo action: c-ocfs_stop_0 * Resource action: ocfs:3 stop on qa-suse-2 * Resource action: ocfs:2 stop on qa-suse-3 * Resource action: ocfs:0 stop on qa-suse-4 * Resource action: ocfs:1 stop on qa-suse-1 * Pseudo action: c-ocfs_stopped_0 * Pseudo action: c-o2stage_stop_0 * Pseudo action: o2stage:0_stop_0 * Resource action: cmirror:1 stop on qa-suse-1 * Pseudo action: o2stage:1_stop_0 * Resource action: cmirror:0 stop on qa-suse-4 * Pseudo action: o2stage:2_stop_0 * Resource action: cmirror:2 stop on qa-suse-3 * Pseudo action: o2stage:3_stop_0 * Resource action: cmirror:3 stop on qa-suse-2 * Resource action: o2cb:1 stop on qa-suse-1 * Resource action: o2cb:0 stop on qa-suse-4 * Resource action: o2cb:2 stop on qa-suse-3 * Resource action: o2cb:3 stop on qa-suse-2 * Pseudo action: o2stage:0_stopped_0 * Pseudo action: o2stage:1_stopped_0 * Pseudo action: o2stage:2_stopped_0 * Pseudo action: o2stage:3_stopped_0 * Pseudo action: c-o2stage_stopped_0 Revised cluster status: Online: [ qa-suse-1 qa-suse-2 qa-suse-3 qa-suse-4 ] sbd_stonith (stonith:external/sbd): Started qa-suse-2 Clone Set: c-o2stage [o2stage] Resource Group: o2stage:0 dlm (ocf::pacemaker:controld): Started qa-suse-1 clvm (ocf::lvm2:clvmd): Started qa-suse-1 o2cb (ocf::ocfs2:o2cb): Stopped (disabled) cmirror (ocf::lvm2:cmirrord): Stopped Resource Group: o2stage:1 dlm (ocf::pacemaker:controld): Started qa-suse-4 clvm (ocf::lvm2:clvmd): Started qa-suse-4 o2cb (ocf::ocfs2:o2cb): Stopped (disabled) cmirror (ocf::lvm2:cmirrord): Stopped Resource Group: o2stage:2 dlm (ocf::pacemaker:controld): Started qa-suse-3 clvm (ocf::lvm2:clvmd): Started qa-suse-3 o2cb (ocf::ocfs2:o2cb): Stopped (disabled) cmirror (ocf::lvm2:cmirrord): Stopped Resource Group: o2stage:3 dlm (ocf::pacemaker:controld): Started qa-suse-2 clvm (ocf::lvm2:clvmd): Started qa-suse-2 o2cb (ocf::ocfs2:o2cb): Stopped (disabled) cmirror (ocf::lvm2:cmirrord): Stopped Clone Set: c-ocfs [ocfs] Stopped: [ qa-suse-1 qa-suse-2 qa-suse-3 qa-suse-4 ] diff --git a/cts/scheduler/bug-lf-2453.summary b/cts/scheduler/bug-lf-2453.summary index c9e43f2b35..894e5ecf23 100644 --- a/cts/scheduler/bug-lf-2453.summary +++ b/cts/scheduler/bug-lf-2453.summary @@ -1,38 +1,38 @@ -2 of 5 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 5 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ domu1 domu2 ] PrimitiveResource1 (ocf::heartbeat:IPaddr2): Started domu1 Clone Set: CloneResource1 [apache] Started: [ domu1 domu2 ] Clone Set: CloneResource2 [DummyResource] Started: [ domu1 domu2 ] Transition Summary: * Stop PrimitiveResource1 ( domu1 ) due to required CloneResource2 running * Stop apache:0 ( domu1 ) due to node availability * Stop apache:1 ( domu2 ) due to node availability * Stop DummyResource:0 ( domu1 ) due to unrunnable CloneResource1 running * Stop DummyResource:1 ( domu2 ) due to unrunnable CloneResource1 running Executing cluster transition: * Resource action: PrimitiveResource1 stop on domu1 * Pseudo action: CloneResource2_stop_0 * Resource action: DummyResource:1 stop on domu1 * Resource action: DummyResource:0 stop on domu2 * Pseudo action: CloneResource2_stopped_0 * Pseudo action: CloneResource1_stop_0 * Resource action: apache:1 stop on domu1 * Resource action: apache:0 stop on domu2 * Pseudo action: CloneResource1_stopped_0 Revised cluster status: Online: [ domu1 domu2 ] PrimitiveResource1 (ocf::heartbeat:IPaddr2): Stopped Clone Set: CloneResource1 [apache] Stopped (disabled): [ domu1 domu2 ] Clone Set: CloneResource2 [DummyResource] Stopped: [ domu1 domu2 ] diff --git a/cts/scheduler/bug-lf-2606.summary b/cts/scheduler/bug-lf-2606.summary index 22103e41fe..13eb7dbe76 100644 --- a/cts/scheduler/bug-lf-2606.summary +++ b/cts/scheduler/bug-lf-2606.summary @@ -1,43 +1,43 @@ -1 of 5 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 5 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Node node2: UNCLEAN (online) Online: [ node1 ] rsc_stonith (stonith:null): Started node1 rsc1 (ocf::pacemaker:Dummy): FAILED node2 (disabled) rsc2 (ocf::pacemaker:Dummy): Started node2 Clone Set: ms3 [rsc3] (promotable) Masters: [ node2 ] Slaves: [ node1 ] Transition Summary: * Fence (reboot) node2 'rsc1 failed there' * Stop rsc1 ( node2 ) due to node availability * Move rsc2 ( node2 -> node1 ) * Stop rsc3:1 ( Master node2 ) due to node availability Executing cluster transition: * Pseudo action: ms3_demote_0 * Fencing node2 (reboot) * Pseudo action: rsc1_stop_0 * Pseudo action: rsc2_stop_0 * Pseudo action: rsc3:1_demote_0 * Pseudo action: ms3_demoted_0 * Pseudo action: ms3_stop_0 * Resource action: rsc2 start on node1 * Pseudo action: rsc3:1_stop_0 * Pseudo action: ms3_stopped_0 * Resource action: rsc2 monitor=10000 on node1 Revised cluster status: Online: [ node1 ] OFFLINE: [ node2 ] rsc_stonith (stonith:null): Started node1 rsc1 (ocf::pacemaker:Dummy): Stopped (disabled) rsc2 (ocf::pacemaker:Dummy): Started node1 Clone Set: ms3 [rsc3] (promotable) Slaves: [ node1 ] Stopped: [ node2 ] diff --git a/cts/scheduler/bug-rh-1097457.summary b/cts/scheduler/bug-rh-1097457.summary index e06badb07d..30eaae510b 100644 --- a/cts/scheduler/bug-rh-1097457.summary +++ b/cts/scheduler/bug-rh-1097457.summary @@ -1,123 +1,123 @@ -2 of 26 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 26 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ lama2 lama3 ] GuestOnline: [ lamaVM1:VM1 lamaVM2:VM2 lamaVM3:VM3 ] restofencelama2 (stonith:fence_ipmilan): Started lama3 restofencelama3 (stonith:fence_ipmilan): Started lama2 VM1 (ocf::heartbeat:VirtualDomain): Started lama2 FSlun1 (ocf::heartbeat:Filesystem): Started lamaVM1 FSlun2 (ocf::heartbeat:Filesystem): Started lamaVM1 VM2 (ocf::heartbeat:VirtualDomain): FAILED lama3 VM3 (ocf::heartbeat:VirtualDomain): Started lama3 FSlun3 (ocf::heartbeat:Filesystem): FAILED lamaVM2 FSlun4 (ocf::heartbeat:Filesystem): Started lamaVM3 FAKE5-IP (ocf::heartbeat:IPaddr2): Stopped (disabled) FAKE6-IP (ocf::heartbeat:IPaddr2): Stopped (disabled) FAKE5 (ocf::heartbeat:Dummy): Started lamaVM3 Resource Group: lamaVM1-G1 FAKE1 (ocf::heartbeat:Dummy): Started lamaVM1 FAKE1-IP (ocf::heartbeat:IPaddr2): Started lamaVM1 Resource Group: lamaVM1-G2 FAKE2 (ocf::heartbeat:Dummy): Started lamaVM1 FAKE2-IP (ocf::heartbeat:IPaddr2): Started lamaVM1 Resource Group: lamaVM1-G3 FAKE3 (ocf::heartbeat:Dummy): Started lamaVM1 FAKE3-IP (ocf::heartbeat:IPaddr2): Started lamaVM1 Resource Group: lamaVM2-G4 FAKE4 (ocf::heartbeat:Dummy): Started lamaVM2 FAKE4-IP (ocf::heartbeat:IPaddr2): Started lamaVM2 Clone Set: FAKE6-clone [FAKE6] Started: [ lamaVM1 lamaVM2 lamaVM3 ] Transition Summary: * Fence (reboot) lamaVM2 (resource: VM2) 'guest is unclean' * Recover VM2 ( lama3 ) * Recover FSlun3 ( lamaVM2 -> lama2 ) * Restart FAKE4 ( lamaVM2 ) due to required VM2 start * Restart FAKE4-IP ( lamaVM2 ) due to required VM2 start * Restart FAKE6:2 ( lamaVM2 ) due to required VM2 start * Restart lamaVM2 ( lama3 ) due to required VM2 start Executing cluster transition: * Resource action: FSlun1 monitor on lamaVM3 * Resource action: FSlun2 monitor on lamaVM3 * Resource action: FSlun3 monitor on lamaVM3 * Resource action: FSlun3 monitor on lamaVM1 * Resource action: FSlun4 monitor on lamaVM1 * Resource action: FAKE5-IP monitor on lamaVM3 * Resource action: FAKE5-IP monitor on lamaVM1 * Resource action: FAKE6-IP monitor on lamaVM3 * Resource action: FAKE6-IP monitor on lamaVM1 * Resource action: FAKE5 monitor on lamaVM1 * Resource action: FAKE1 monitor on lamaVM3 * Resource action: FAKE1-IP monitor on lamaVM3 * Resource action: FAKE2 monitor on lamaVM3 * Resource action: FAKE2-IP monitor on lamaVM3 * Resource action: FAKE3 monitor on lamaVM3 * Resource action: FAKE3-IP monitor on lamaVM3 * Resource action: FAKE4 monitor on lamaVM3 * Resource action: FAKE4 monitor on lamaVM1 * Resource action: FAKE4-IP monitor on lamaVM3 * Resource action: FAKE4-IP monitor on lamaVM1 * Resource action: lamaVM2 stop on lama3 * Resource action: VM2 stop on lama3 * Pseudo action: stonith-lamaVM2-reboot on lamaVM2 * Resource action: VM2 start on lama3 * Resource action: VM2 monitor=10000 on lama3 * Pseudo action: lamaVM2-G4_stop_0 * Pseudo action: FAKE4-IP_stop_0 * Pseudo action: FAKE6-clone_stop_0 * Resource action: lamaVM2 start on lama3 * Resource action: lamaVM2 monitor=30000 on lama3 * Resource action: FSlun3 monitor=10000 on lamaVM2 * Pseudo action: FAKE4_stop_0 * Pseudo action: FAKE6_stop_0 * Pseudo action: FAKE6-clone_stopped_0 * Pseudo action: FAKE6-clone_start_0 * Pseudo action: lamaVM2-G4_stopped_0 * Resource action: FAKE6 start on lamaVM2 * Resource action: FAKE6 monitor=30000 on lamaVM2 * Pseudo action: FAKE6-clone_running_0 * Pseudo action: FSlun3_stop_0 * Resource action: FSlun3 start on lama2 * Pseudo action: lamaVM2-G4_start_0 * Resource action: FAKE4 start on lamaVM2 * Resource action: FAKE4 monitor=30000 on lamaVM2 * Resource action: FAKE4-IP start on lamaVM2 * Resource action: FAKE4-IP monitor=30000 on lamaVM2 * Resource action: FSlun3 monitor=10000 on lama2 * Pseudo action: lamaVM2-G4_running_0 Revised cluster status: Online: [ lama2 lama3 ] GuestOnline: [ lamaVM1:VM1 lamaVM2:VM2 lamaVM3:VM3 ] restofencelama2 (stonith:fence_ipmilan): Started lama3 restofencelama3 (stonith:fence_ipmilan): Started lama2 VM1 (ocf::heartbeat:VirtualDomain): Started lama2 FSlun1 (ocf::heartbeat:Filesystem): Started lamaVM1 FSlun2 (ocf::heartbeat:Filesystem): Started lamaVM1 VM2 (ocf::heartbeat:VirtualDomain): FAILED lama3 VM3 (ocf::heartbeat:VirtualDomain): Started lama3 FSlun3 (ocf::heartbeat:Filesystem): FAILED[ lama2 lamaVM2 ] FSlun4 (ocf::heartbeat:Filesystem): Started lamaVM3 FAKE5-IP (ocf::heartbeat:IPaddr2): Stopped (disabled) FAKE6-IP (ocf::heartbeat:IPaddr2): Stopped (disabled) FAKE5 (ocf::heartbeat:Dummy): Started lamaVM3 Resource Group: lamaVM1-G1 FAKE1 (ocf::heartbeat:Dummy): Started lamaVM1 FAKE1-IP (ocf::heartbeat:IPaddr2): Started lamaVM1 Resource Group: lamaVM1-G2 FAKE2 (ocf::heartbeat:Dummy): Started lamaVM1 FAKE2-IP (ocf::heartbeat:IPaddr2): Started lamaVM1 Resource Group: lamaVM1-G3 FAKE3 (ocf::heartbeat:Dummy): Started lamaVM1 FAKE3-IP (ocf::heartbeat:IPaddr2): Started lamaVM1 Resource Group: lamaVM2-G4 FAKE4 (ocf::heartbeat:Dummy): Started lamaVM2 FAKE4-IP (ocf::heartbeat:IPaddr2): Started lamaVM2 Clone Set: FAKE6-clone [FAKE6] Started: [ lamaVM1 lamaVM2 lamaVM3 ] diff --git a/cts/scheduler/bug-suse-707150.summary b/cts/scheduler/bug-suse-707150.summary index b3949888ee..adb37ccfee 100644 --- a/cts/scheduler/bug-suse-707150.summary +++ b/cts/scheduler/bug-suse-707150.summary @@ -1,72 +1,72 @@ -9 of 28 resources DISABLED and 0 BLOCKED from being started due to failures +5 of 28 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ hex-0 hex-9 ] OFFLINE: [ hex-7 hex-8 ] vm-00 (ocf::heartbeat:Xen): Stopped (disabled) Clone Set: base-clone [base-group] Resource Group: base-group:0 dlm (ocf::pacemaker:controld): Started hex-0 o2cb (ocf::ocfs2:o2cb): Stopped clvm (ocf::lvm2:clvmd): Stopped cmirrord (ocf::lvm2:cmirrord): Stopped vg1 (ocf::heartbeat:LVM): Stopped (disabled) ocfs2-1 (ocf::heartbeat:Filesystem): Stopped Stopped: [ hex-7 hex-8 hex-9 ] vm-01 (ocf::heartbeat:Xen): Stopped fencing-sbd (stonith:external/sbd): Started hex-9 dummy1 (ocf::heartbeat:Dummy): Started hex-0 Transition Summary: * Start o2cb:0 ( hex-0 ) * Start clvm:0 ( hex-0 ) * Start cmirrord:0 ( hex-0 ) * Start dlm:1 ( hex-9 ) * Start o2cb:1 ( hex-9 ) * Start clvm:1 ( hex-9 ) * Start cmirrord:1 ( hex-9 ) * Start vm-01 ( hex-9 ) due to unrunnable base-clone running (blocked) Executing cluster transition: * Resource action: vg1:1 monitor on hex-9 * Pseudo action: base-clone_start_0 * Pseudo action: load_stopped_hex-9 * Pseudo action: load_stopped_hex-8 * Pseudo action: load_stopped_hex-7 * Pseudo action: load_stopped_hex-0 * Pseudo action: base-group:0_start_0 * Resource action: o2cb:0 start on hex-0 * Resource action: clvm:0 start on hex-0 * Resource action: cmirrord:0 start on hex-0 * Pseudo action: base-group:1_start_0 * Resource action: dlm:1 start on hex-9 * Resource action: o2cb:1 start on hex-9 * Resource action: clvm:1 start on hex-9 * Resource action: cmirrord:1 start on hex-9 Revised cluster status: Online: [ hex-0 hex-9 ] OFFLINE: [ hex-7 hex-8 ] vm-00 (ocf::heartbeat:Xen): Stopped (disabled) Clone Set: base-clone [base-group] Resource Group: base-group:0 dlm (ocf::pacemaker:controld): Started hex-0 o2cb (ocf::ocfs2:o2cb): Started hex-0 clvm (ocf::lvm2:clvmd): Started hex-0 cmirrord (ocf::lvm2:cmirrord): Started hex-0 vg1 (ocf::heartbeat:LVM): Stopped (disabled) ocfs2-1 (ocf::heartbeat:Filesystem): Stopped Resource Group: base-group:1 dlm (ocf::pacemaker:controld): Started hex-9 o2cb (ocf::ocfs2:o2cb): Started hex-9 clvm (ocf::lvm2:clvmd): Started hex-9 cmirrord (ocf::lvm2:cmirrord): Started hex-9 vg1 (ocf::heartbeat:LVM): Stopped (disabled) ocfs2-1 (ocf::heartbeat:Filesystem): Stopped Stopped: [ hex-7 hex-8 ] vm-01 (ocf::heartbeat:Xen): Stopped fencing-sbd (stonith:external/sbd): Started hex-9 dummy1 (ocf::heartbeat:Dummy): Started hex-0 diff --git a/cts/scheduler/clone-fail-block-colocation.summary b/cts/scheduler/clone-fail-block-colocation.summary index bdca45f0c5..04e2f2a6be 100644 --- a/cts/scheduler/clone-fail-block-colocation.summary +++ b/cts/scheduler/clone-fail-block-colocation.summary @@ -1,57 +1,58 @@ +0 of 10 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ DEM-1 DEM-2 ] Resource Group: svc ipv6_dem_tas_dns (ocf::heartbeat:IPv6addr): Started DEM-1 d_bird_subnet_state (lsb:bird_subnet_state): Started DEM-1 ip_mgmt (ocf::heartbeat:IPaddr2): Started DEM-1 ip_trf_tas (ocf::heartbeat:IPaddr2): Started DEM-1 Clone Set: cl_bird [d_bird] Started: [ DEM-1 DEM-2 ] Clone Set: cl_bird6 [d_bird6] d_bird6 (lsb:bird6): FAILED DEM-1 (blocked) Started: [ DEM-2 ] Clone Set: cl_tomcat_nms [d_tomcat_nms] Started: [ DEM-1 DEM-2 ] Transition Summary: * Move ipv6_dem_tas_dns ( DEM-1 -> DEM-2 ) * Move d_bird_subnet_state ( DEM-1 -> DEM-2 ) * Move ip_mgmt ( DEM-1 -> DEM-2 ) * Move ip_trf_tas ( DEM-1 -> DEM-2 ) Executing cluster transition: * Pseudo action: svc_stop_0 * Resource action: ip_trf_tas stop on DEM-1 * Resource action: ip_mgmt stop on DEM-1 * Resource action: d_bird_subnet_state stop on DEM-1 * Resource action: ipv6_dem_tas_dns stop on DEM-1 * Pseudo action: svc_stopped_0 * Pseudo action: svc_start_0 * Resource action: ipv6_dem_tas_dns start on DEM-2 * Resource action: d_bird_subnet_state start on DEM-2 * Resource action: ip_mgmt start on DEM-2 * Resource action: ip_trf_tas start on DEM-2 * Pseudo action: svc_running_0 * Resource action: ipv6_dem_tas_dns monitor=10000 on DEM-2 * Resource action: d_bird_subnet_state monitor=10000 on DEM-2 * Resource action: ip_mgmt monitor=10000 on DEM-2 * Resource action: ip_trf_tas monitor=10000 on DEM-2 Revised cluster status: Online: [ DEM-1 DEM-2 ] Resource Group: svc ipv6_dem_tas_dns (ocf::heartbeat:IPv6addr): Started DEM-2 d_bird_subnet_state (lsb:bird_subnet_state): Started DEM-2 ip_mgmt (ocf::heartbeat:IPaddr2): Started DEM-2 ip_trf_tas (ocf::heartbeat:IPaddr2): Started DEM-2 Clone Set: cl_bird [d_bird] Started: [ DEM-1 DEM-2 ] Clone Set: cl_bird6 [d_bird6] d_bird6 (lsb:bird6): FAILED DEM-1 (blocked) Started: [ DEM-2 ] Clone Set: cl_tomcat_nms [d_tomcat_nms] Started: [ DEM-1 DEM-2 ] diff --git a/cts/scheduler/clone-order-16instances.summary b/cts/scheduler/clone-order-16instances.summary index f3cd5547db..004952a494 100644 --- a/cts/scheduler/clone-order-16instances.summary +++ b/cts/scheduler/clone-order-16instances.summary @@ -1,69 +1,69 @@ -16 of 33 resources DISABLED and 0 BLOCKED from being started due to failures +16 of 33 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ virt-009.cluster-qe.lab.eng.brq.redhat.com virt-010.cluster-qe.lab.eng.brq.redhat.com virt-012.cluster-qe.lab.eng.brq.redhat.com virt-013.cluster-qe.lab.eng.brq.redhat.com virt-014.cluster-qe.lab.eng.brq.redhat.com virt-015.cluster-qe.lab.eng.brq.redhat.com virt-016.cluster-qe.lab.eng.brq.redhat.com virt-020.cluster-qe.lab.eng.brq.redhat.com virt-027.cluster-qe.lab.eng.brq.redhat.com virt-028.cluster-qe.lab.eng.brq.redhat.com virt-029.cluster-qe.lab.eng.brq.redhat.com virt-030.cluster-qe.lab.eng.brq.redhat.com virt-031.cluster-qe.lab.eng.brq.redhat.com virt-032.cluster-qe.lab.eng.brq.redhat.com virt-033.cluster-qe.lab.eng.brq.redhat.com virt-034.cluster-qe.lab.eng.brq.redhat.com ] virt-fencing (stonith:fence_xvm): Started virt-010.cluster-qe.lab.eng.brq.redhat.com Clone Set: dlm-clone [dlm] Started: [ virt-010.cluster-qe.lab.eng.brq.redhat.com virt-012.cluster-qe.lab.eng.brq.redhat.com ] Stopped: [ virt-009.cluster-qe.lab.eng.brq.redhat.com virt-013.cluster-qe.lab.eng.brq.redhat.com virt-014.cluster-qe.lab.eng.brq.redhat.com virt-015.cluster-qe.lab.eng.brq.redhat.com virt-016.cluster-qe.lab.eng.brq.redhat.com virt-020.cluster-qe.lab.eng.brq.redhat.com virt-027.cluster-qe.lab.eng.brq.redhat.com virt-028.cluster-qe.lab.eng.brq.redhat.com virt-029.cluster-qe.lab.eng.brq.redhat.com virt-030.cluster-qe.lab.eng.brq.redhat.com virt-031.cluster-qe.lab.eng.brq.redhat.com virt-032.cluster-qe.lab.eng.brq.redhat.com virt-033.cluster-qe.lab.eng.brq.redhat.com virt-034.cluster-qe.lab.eng.brq.redhat.com ] Clone Set: clvmd-clone [clvmd] Stopped (disabled): [ virt-009.cluster-qe.lab.eng.brq.redhat.com virt-010.cluster-qe.lab.eng.brq.redhat.com virt-012.cluster-qe.lab.eng.brq.redhat.com virt-013.cluster-qe.lab.eng.brq.redhat.com virt-014.cluster-qe.lab.eng.brq.redhat.com virt-015.cluster-qe.lab.eng.brq.redhat.com virt-016.cluster-qe.lab.eng.brq.redhat.com virt-020.cluster-qe.lab.eng.brq.redhat.com virt-027.cluster-qe.lab.eng.brq.redhat.com virt-028.cluster-qe.lab.eng.brq.redhat.com virt-029.cluster-qe.lab.eng.brq.redhat.com virt-030.cluster-qe.lab.eng.brq.redhat.com virt-031.cluster-qe.lab.eng.brq.redhat.com virt-032.cluster-qe.lab.eng.brq.redhat.com virt-033.cluster-qe.lab.eng.brq.redhat.com virt-034.cluster-qe.lab.eng.brq.redhat.com ] Transition Summary: * Start dlm:2 ( virt-009.cluster-qe.lab.eng.brq.redhat.com ) * Start dlm:3 ( virt-013.cluster-qe.lab.eng.brq.redhat.com ) * Start dlm:4 ( virt-014.cluster-qe.lab.eng.brq.redhat.com ) * Start dlm:5 ( virt-015.cluster-qe.lab.eng.brq.redhat.com ) * Start dlm:6 ( virt-016.cluster-qe.lab.eng.brq.redhat.com ) * Start dlm:7 ( virt-020.cluster-qe.lab.eng.brq.redhat.com ) * Start dlm:8 ( virt-027.cluster-qe.lab.eng.brq.redhat.com ) * Start dlm:9 ( virt-028.cluster-qe.lab.eng.brq.redhat.com ) * Start dlm:10 ( virt-029.cluster-qe.lab.eng.brq.redhat.com ) * Start dlm:11 ( virt-030.cluster-qe.lab.eng.brq.redhat.com ) * Start dlm:12 ( virt-031.cluster-qe.lab.eng.brq.redhat.com ) * Start dlm:13 ( virt-032.cluster-qe.lab.eng.brq.redhat.com ) * Start dlm:14 ( virt-033.cluster-qe.lab.eng.brq.redhat.com ) * Start dlm:15 ( virt-034.cluster-qe.lab.eng.brq.redhat.com ) Executing cluster transition: * Pseudo action: dlm-clone_start_0 * Resource action: dlm start on virt-009.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm start on virt-013.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm start on virt-014.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm start on virt-015.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm start on virt-016.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm start on virt-020.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm start on virt-027.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm start on virt-028.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm start on virt-029.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm start on virt-030.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm start on virt-031.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm start on virt-032.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm start on virt-033.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm start on virt-034.cluster-qe.lab.eng.brq.redhat.com * Pseudo action: dlm-clone_running_0 * Resource action: dlm monitor=30000 on virt-009.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm monitor=30000 on virt-013.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm monitor=30000 on virt-014.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm monitor=30000 on virt-015.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm monitor=30000 on virt-016.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm monitor=30000 on virt-020.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm monitor=30000 on virt-027.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm monitor=30000 on virt-028.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm monitor=30000 on virt-029.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm monitor=30000 on virt-030.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm monitor=30000 on virt-031.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm monitor=30000 on virt-032.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm monitor=30000 on virt-033.cluster-qe.lab.eng.brq.redhat.com * Resource action: dlm monitor=30000 on virt-034.cluster-qe.lab.eng.brq.redhat.com Revised cluster status: Online: [ virt-009.cluster-qe.lab.eng.brq.redhat.com virt-010.cluster-qe.lab.eng.brq.redhat.com virt-012.cluster-qe.lab.eng.brq.redhat.com virt-013.cluster-qe.lab.eng.brq.redhat.com virt-014.cluster-qe.lab.eng.brq.redhat.com virt-015.cluster-qe.lab.eng.brq.redhat.com virt-016.cluster-qe.lab.eng.brq.redhat.com virt-020.cluster-qe.lab.eng.brq.redhat.com virt-027.cluster-qe.lab.eng.brq.redhat.com virt-028.cluster-qe.lab.eng.brq.redhat.com virt-029.cluster-qe.lab.eng.brq.redhat.com virt-030.cluster-qe.lab.eng.brq.redhat.com virt-031.cluster-qe.lab.eng.brq.redhat.com virt-032.cluster-qe.lab.eng.brq.redhat.com virt-033.cluster-qe.lab.eng.brq.redhat.com virt-034.cluster-qe.lab.eng.brq.redhat.com ] virt-fencing (stonith:fence_xvm): Started virt-010.cluster-qe.lab.eng.brq.redhat.com Clone Set: dlm-clone [dlm] Started: [ virt-009.cluster-qe.lab.eng.brq.redhat.com virt-010.cluster-qe.lab.eng.brq.redhat.com virt-012.cluster-qe.lab.eng.brq.redhat.com virt-013.cluster-qe.lab.eng.brq.redhat.com virt-014.cluster-qe.lab.eng.brq.redhat.com virt-015.cluster-qe.lab.eng.brq.redhat.com virt-016.cluster-qe.lab.eng.brq.redhat.com virt-020.cluster-qe.lab.eng.brq.redhat.com virt-027.cluster-qe.lab.eng.brq.redhat.com virt-028.cluster-qe.lab.eng.brq.redhat.com virt-029.cluster-qe.lab.eng.brq.redhat.com virt-030.cluster-qe.lab.eng.brq.redhat.com virt-031.cluster-qe.lab.eng.brq.redhat.com virt-032.cluster-qe.lab.eng.brq.redhat.com virt-033.cluster-qe.lab.eng.brq.redhat.com virt-034.cluster-qe.lab.eng.brq.redhat.com ] Clone Set: clvmd-clone [clvmd] Stopped (disabled): [ virt-009.cluster-qe.lab.eng.brq.redhat.com virt-010.cluster-qe.lab.eng.brq.redhat.com virt-012.cluster-qe.lab.eng.brq.redhat.com virt-013.cluster-qe.lab.eng.brq.redhat.com virt-014.cluster-qe.lab.eng.brq.redhat.com virt-015.cluster-qe.lab.eng.brq.redhat.com virt-016.cluster-qe.lab.eng.brq.redhat.com virt-020.cluster-qe.lab.eng.brq.redhat.com virt-027.cluster-qe.lab.eng.brq.redhat.com virt-028.cluster-qe.lab.eng.brq.redhat.com virt-029.cluster-qe.lab.eng.brq.redhat.com virt-030.cluster-qe.lab.eng.brq.redhat.com virt-031.cluster-qe.lab.eng.brq.redhat.com virt-032.cluster-qe.lab.eng.brq.redhat.com virt-033.cluster-qe.lab.eng.brq.redhat.com virt-034.cluster-qe.lab.eng.brq.redhat.com ] diff --git a/cts/scheduler/cloned-group-stop.summary b/cts/scheduler/cloned-group-stop.summary index 9aa0449d04..2f3d163aff 100644 --- a/cts/scheduler/cloned-group-stop.summary +++ b/cts/scheduler/cloned-group-stop.summary @@ -1,88 +1,88 @@ -2 of 20 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 20 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ rhos4-node3 rhos4-node4 ] virt-fencing (stonith:fence_xvm): Started rhos4-node3 Resource Group: mysql-group mysql-vip (ocf::heartbeat:IPaddr2): Started rhos4-node3 mysql-fs (ocf::heartbeat:Filesystem): Started rhos4-node3 mysql-db (ocf::heartbeat:mysql): Started rhos4-node3 Clone Set: qpidd-clone [qpidd] Started: [ rhos4-node3 rhos4-node4 ] Clone Set: keystone-clone [keystone] Started: [ rhos4-node3 rhos4-node4 ] Clone Set: glance-clone [glance] Started: [ rhos4-node3 rhos4-node4 ] Clone Set: cinder-clone [cinder] Started: [ rhos4-node3 rhos4-node4 ] Transition Summary: * Stop qpidd:0 ( rhos4-node4 ) due to node availability * Stop qpidd:1 ( rhos4-node3 ) due to node availability * Stop keystone:0 ( rhos4-node4 ) due to unrunnable qpidd-clone running * Stop keystone:1 ( rhos4-node3 ) due to unrunnable qpidd-clone running * Stop glance-fs:0 ( rhos4-node4 ) due to required keystone-clone running * Stop glance-registry:0 ( rhos4-node4 ) due to required glance-fs:0 stop * Stop glance-api:0 ( rhos4-node4 ) due to required glance-registry:0 start * Stop glance-fs:1 ( rhos4-node3 ) due to required keystone-clone running * Stop glance-registry:1 ( rhos4-node3 ) due to required glance-fs:1 stop * Stop glance-api:1 ( rhos4-node3 ) due to required glance-registry:1 start * Stop cinder-api:0 ( rhos4-node4 ) due to required glance-clone running * Stop cinder-scheduler:0 ( rhos4-node4 ) due to required cinder-api:0 stop * Stop cinder-volume:0 ( rhos4-node4 ) due to required cinder-scheduler:0 start * Stop cinder-api:1 ( rhos4-node3 ) due to required glance-clone running * Stop cinder-scheduler:1 ( rhos4-node3 ) due to required cinder-api:1 stop * Stop cinder-volume:1 ( rhos4-node3 ) due to required cinder-scheduler:1 start Executing cluster transition: * Pseudo action: cinder-clone_stop_0 * Pseudo action: cinder:0_stop_0 * Resource action: cinder-volume stop on rhos4-node4 * Pseudo action: cinder:1_stop_0 * Resource action: cinder-volume stop on rhos4-node3 * Resource action: cinder-scheduler stop on rhos4-node4 * Resource action: cinder-scheduler stop on rhos4-node3 * Resource action: cinder-api stop on rhos4-node4 * Resource action: cinder-api stop on rhos4-node3 * Pseudo action: cinder:0_stopped_0 * Pseudo action: cinder:1_stopped_0 * Pseudo action: cinder-clone_stopped_0 * Pseudo action: glance-clone_stop_0 * Pseudo action: glance:0_stop_0 * Resource action: glance-api stop on rhos4-node4 * Pseudo action: glance:1_stop_0 * Resource action: glance-api stop on rhos4-node3 * Resource action: glance-registry stop on rhos4-node4 * Resource action: glance-registry stop on rhos4-node3 * Resource action: glance-fs stop on rhos4-node4 * Resource action: glance-fs stop on rhos4-node3 * Pseudo action: glance:0_stopped_0 * Pseudo action: glance:1_stopped_0 * Pseudo action: glance-clone_stopped_0 * Pseudo action: keystone-clone_stop_0 * Resource action: keystone stop on rhos4-node4 * Resource action: keystone stop on rhos4-node3 * Pseudo action: keystone-clone_stopped_0 * Pseudo action: qpidd-clone_stop_0 * Resource action: qpidd stop on rhos4-node4 * Resource action: qpidd stop on rhos4-node3 * Pseudo action: qpidd-clone_stopped_0 Revised cluster status: Online: [ rhos4-node3 rhos4-node4 ] virt-fencing (stonith:fence_xvm): Started rhos4-node3 Resource Group: mysql-group mysql-vip (ocf::heartbeat:IPaddr2): Started rhos4-node3 mysql-fs (ocf::heartbeat:Filesystem): Started rhos4-node3 mysql-db (ocf::heartbeat:mysql): Started rhos4-node3 Clone Set: qpidd-clone [qpidd] Stopped (disabled): [ rhos4-node3 rhos4-node4 ] Clone Set: keystone-clone [keystone] Stopped: [ rhos4-node3 rhos4-node4 ] Clone Set: glance-clone [glance] Stopped: [ rhos4-node3 rhos4-node4 ] Clone Set: cinder-clone [cinder] Stopped: [ rhos4-node3 rhos4-node4 ] diff --git a/cts/scheduler/coloc-clone-stays-active.summary b/cts/scheduler/coloc-clone-stays-active.summary index 010e537612..6fe7427daa 100644 --- a/cts/scheduler/coloc-clone-stays-active.summary +++ b/cts/scheduler/coloc-clone-stays-active.summary @@ -1,206 +1,206 @@ -12 of 87 resources DISABLED and 0 BLOCKED from being started due to failures +9 of 87 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ s01-0 s01-1 ] stonith-s01-0 (stonith:external/ipmi): Started s01-1 stonith-s01-1 (stonith:external/ipmi): Started s01-0 Resource Group: iscsi-pool-0-target-all iscsi-pool-0-target (ocf::vds-ok:iSCSITarget): Started s01-0 iscsi-pool-0-lun-1 (ocf::vds-ok:iSCSILogicalUnit): Started s01-0 Resource Group: iscsi-pool-0-vips vip-235 (ocf::heartbeat:IPaddr2): Started s01-0 vip-236 (ocf::heartbeat:IPaddr2): Started s01-0 Resource Group: iscsi-pool-1-target-all iscsi-pool-1-target (ocf::vds-ok:iSCSITarget): Started s01-1 iscsi-pool-1-lun-1 (ocf::vds-ok:iSCSILogicalUnit): Started s01-1 Resource Group: iscsi-pool-1-vips vip-237 (ocf::heartbeat:IPaddr2): Started s01-1 vip-238 (ocf::heartbeat:IPaddr2): Started s01-1 Clone Set: ms-drbd-pool-0 [drbd-pool-0] (promotable) Masters: [ s01-0 ] Slaves: [ s01-1 ] Clone Set: ms-drbd-pool-1 [drbd-pool-1] (promotable) Masters: [ s01-1 ] Slaves: [ s01-0 ] Clone Set: ms-iscsi-pool-0-vips-fw [iscsi-pool-0-vips-fw] (promotable) Masters: [ s01-0 ] Slaves: [ s01-1 ] Clone Set: ms-iscsi-pool-1-vips-fw [iscsi-pool-1-vips-fw] (promotable) Masters: [ s01-1 ] Slaves: [ s01-0 ] Clone Set: cl-o2cb [o2cb] Stopped (disabled): [ s01-0 s01-1 ] Clone Set: ms-drbd-s01-service [drbd-s01-service] (promotable) Masters: [ s01-0 s01-1 ] Clone Set: cl-s01-service-fs [s01-service-fs] Started: [ s01-0 s01-1 ] Clone Set: cl-ietd [ietd] Started: [ s01-0 s01-1 ] Clone Set: cl-dhcpd [dhcpd] Stopped (disabled): [ s01-0 s01-1 ] Resource Group: http-server vip-233 (ocf::heartbeat:IPaddr2): Started s01-0 nginx (lsb:nginx): Stopped (disabled) Clone Set: ms-drbd-s01-logs [drbd-s01-logs] (promotable) Masters: [ s01-0 s01-1 ] Clone Set: cl-s01-logs-fs [s01-logs-fs] Started: [ s01-0 s01-1 ] Resource Group: syslog-server vip-234 (ocf::heartbeat:IPaddr2): Started s01-1 syslog-ng (ocf::heartbeat:syslog-ng): Started s01-1 Resource Group: tftp-server vip-232 (ocf::heartbeat:IPaddr2): Stopped tftpd (ocf::heartbeat:Xinetd): Stopped Clone Set: cl-xinetd [xinetd] Started: [ s01-0 s01-1 ] Clone Set: cl-ospf-routing [ospf-routing] Started: [ s01-0 s01-1 ] Clone Set: connected-outer [ping-bmc-and-switch] Started: [ s01-0 s01-1 ] Resource Group: iscsi-vds-dom0-stateless-0-target-all iscsi-vds-dom0-stateless-0-target (ocf::vds-ok:iSCSITarget): Stopped (disabled) iscsi-vds-dom0-stateless-0-lun-1 (ocf::vds-ok:iSCSILogicalUnit): Stopped (disabled) Resource Group: iscsi-vds-dom0-stateless-0-vips vip-227 (ocf::heartbeat:IPaddr2): Stopped vip-228 (ocf::heartbeat:IPaddr2): Stopped Clone Set: ms-drbd-vds-dom0-stateless-0 [drbd-vds-dom0-stateless-0] (promotable) Masters: [ s01-0 ] Slaves: [ s01-1 ] Clone Set: ms-iscsi-vds-dom0-stateless-0-vips-fw [iscsi-vds-dom0-stateless-0-vips-fw] (promotable) Slaves: [ s01-0 s01-1 ] Clone Set: cl-dlm [dlm] Started: [ s01-0 s01-1 ] Clone Set: ms-drbd-vds-tftpboot [drbd-vds-tftpboot] (promotable) Masters: [ s01-0 s01-1 ] Clone Set: cl-vds-tftpboot-fs [vds-tftpboot-fs] Stopped (disabled): [ s01-0 s01-1 ] Clone Set: cl-gfs2 [gfs2] Started: [ s01-0 s01-1 ] Clone Set: ms-drbd-vds-http [drbd-vds-http] (promotable) Masters: [ s01-0 s01-1 ] Clone Set: cl-vds-http-fs [vds-http-fs] Started: [ s01-0 s01-1 ] Clone Set: cl-clvmd [clvmd] Started: [ s01-0 s01-1 ] Clone Set: ms-drbd-s01-vm-data [drbd-s01-vm-data] (promotable) Masters: [ s01-0 s01-1 ] Clone Set: cl-s01-vm-data-metadata-fs [s01-vm-data-metadata-fs] Started: [ s01-0 s01-1 ] Clone Set: cl-vg-s01-vm-data [vg-s01-vm-data] Started: [ s01-0 s01-1 ] mgmt-vm (ocf::vds-ok:VirtualDomain): Started s01-0 Clone Set: cl-drbdlinks-s01-service [drbdlinks-s01-service] Started: [ s01-0 s01-1 ] Clone Set: cl-libvirtd [libvirtd] Started: [ s01-0 s01-1 ] Clone Set: cl-s01-vm-data-storage-pool [s01-vm-data-storage-pool] Started: [ s01-0 s01-1 ] Transition Summary: * Migrate mgmt-vm ( s01-0 -> s01-1 ) Executing cluster transition: * Resource action: mgmt-vm migrate_to on s01-0 * Resource action: mgmt-vm migrate_from on s01-1 * Resource action: mgmt-vm stop on s01-0 * Pseudo action: mgmt-vm_start_0 * Resource action: mgmt-vm monitor=10000 on s01-1 Revised cluster status: Online: [ s01-0 s01-1 ] stonith-s01-0 (stonith:external/ipmi): Started s01-1 stonith-s01-1 (stonith:external/ipmi): Started s01-0 Resource Group: iscsi-pool-0-target-all iscsi-pool-0-target (ocf::vds-ok:iSCSITarget): Started s01-0 iscsi-pool-0-lun-1 (ocf::vds-ok:iSCSILogicalUnit): Started s01-0 Resource Group: iscsi-pool-0-vips vip-235 (ocf::heartbeat:IPaddr2): Started s01-0 vip-236 (ocf::heartbeat:IPaddr2): Started s01-0 Resource Group: iscsi-pool-1-target-all iscsi-pool-1-target (ocf::vds-ok:iSCSITarget): Started s01-1 iscsi-pool-1-lun-1 (ocf::vds-ok:iSCSILogicalUnit): Started s01-1 Resource Group: iscsi-pool-1-vips vip-237 (ocf::heartbeat:IPaddr2): Started s01-1 vip-238 (ocf::heartbeat:IPaddr2): Started s01-1 Clone Set: ms-drbd-pool-0 [drbd-pool-0] (promotable) Masters: [ s01-0 ] Slaves: [ s01-1 ] Clone Set: ms-drbd-pool-1 [drbd-pool-1] (promotable) Masters: [ s01-1 ] Slaves: [ s01-0 ] Clone Set: ms-iscsi-pool-0-vips-fw [iscsi-pool-0-vips-fw] (promotable) Masters: [ s01-0 ] Slaves: [ s01-1 ] Clone Set: ms-iscsi-pool-1-vips-fw [iscsi-pool-1-vips-fw] (promotable) Masters: [ s01-1 ] Slaves: [ s01-0 ] Clone Set: cl-o2cb [o2cb] Stopped (disabled): [ s01-0 s01-1 ] Clone Set: ms-drbd-s01-service [drbd-s01-service] (promotable) Masters: [ s01-0 s01-1 ] Clone Set: cl-s01-service-fs [s01-service-fs] Started: [ s01-0 s01-1 ] Clone Set: cl-ietd [ietd] Started: [ s01-0 s01-1 ] Clone Set: cl-dhcpd [dhcpd] Stopped (disabled): [ s01-0 s01-1 ] Resource Group: http-server vip-233 (ocf::heartbeat:IPaddr2): Started s01-0 nginx (lsb:nginx): Stopped (disabled) Clone Set: ms-drbd-s01-logs [drbd-s01-logs] (promotable) Masters: [ s01-0 s01-1 ] Clone Set: cl-s01-logs-fs [s01-logs-fs] Started: [ s01-0 s01-1 ] Resource Group: syslog-server vip-234 (ocf::heartbeat:IPaddr2): Started s01-1 syslog-ng (ocf::heartbeat:syslog-ng): Started s01-1 Resource Group: tftp-server vip-232 (ocf::heartbeat:IPaddr2): Stopped tftpd (ocf::heartbeat:Xinetd): Stopped Clone Set: cl-xinetd [xinetd] Started: [ s01-0 s01-1 ] Clone Set: cl-ospf-routing [ospf-routing] Started: [ s01-0 s01-1 ] Clone Set: connected-outer [ping-bmc-and-switch] Started: [ s01-0 s01-1 ] Resource Group: iscsi-vds-dom0-stateless-0-target-all iscsi-vds-dom0-stateless-0-target (ocf::vds-ok:iSCSITarget): Stopped (disabled) iscsi-vds-dom0-stateless-0-lun-1 (ocf::vds-ok:iSCSILogicalUnit): Stopped (disabled) Resource Group: iscsi-vds-dom0-stateless-0-vips vip-227 (ocf::heartbeat:IPaddr2): Stopped vip-228 (ocf::heartbeat:IPaddr2): Stopped Clone Set: ms-drbd-vds-dom0-stateless-0 [drbd-vds-dom0-stateless-0] (promotable) Masters: [ s01-0 ] Slaves: [ s01-1 ] Clone Set: ms-iscsi-vds-dom0-stateless-0-vips-fw [iscsi-vds-dom0-stateless-0-vips-fw] (promotable) Slaves: [ s01-0 s01-1 ] Clone Set: cl-dlm [dlm] Started: [ s01-0 s01-1 ] Clone Set: ms-drbd-vds-tftpboot [drbd-vds-tftpboot] (promotable) Masters: [ s01-0 s01-1 ] Clone Set: cl-vds-tftpboot-fs [vds-tftpboot-fs] Stopped (disabled): [ s01-0 s01-1 ] Clone Set: cl-gfs2 [gfs2] Started: [ s01-0 s01-1 ] Clone Set: ms-drbd-vds-http [drbd-vds-http] (promotable) Masters: [ s01-0 s01-1 ] Clone Set: cl-vds-http-fs [vds-http-fs] Started: [ s01-0 s01-1 ] Clone Set: cl-clvmd [clvmd] Started: [ s01-0 s01-1 ] Clone Set: ms-drbd-s01-vm-data [drbd-s01-vm-data] (promotable) Masters: [ s01-0 s01-1 ] Clone Set: cl-s01-vm-data-metadata-fs [s01-vm-data-metadata-fs] Started: [ s01-0 s01-1 ] Clone Set: cl-vg-s01-vm-data [vg-s01-vm-data] Started: [ s01-0 s01-1 ] mgmt-vm (ocf::vds-ok:VirtualDomain): Started s01-1 Clone Set: cl-drbdlinks-s01-service [drbdlinks-s01-service] Started: [ s01-0 s01-1 ] Clone Set: cl-libvirtd [libvirtd] Started: [ s01-0 s01-1 ] Clone Set: cl-s01-vm-data-storage-pool [s01-vm-data-storage-pool] Started: [ s01-0 s01-1 ] diff --git a/cts/scheduler/colocation_constraint_stops_slave.summary b/cts/scheduler/colocation_constraint_stops_slave.summary index ae74d017f5..c70d244c78 100644 --- a/cts/scheduler/colocation_constraint_stops_slave.summary +++ b/cts/scheduler/colocation_constraint_stops_slave.summary @@ -1,33 +1,33 @@ -1 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 2 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] Clone Set: MASTER_RSC_A [NATIVE_RSC_A] (promotable) Slaves: [ fc16-builder ] NATIVE_RSC_B (ocf::pacemaker:Dummy): Started fc16-builder (disabled) Transition Summary: * Stop NATIVE_RSC_A:0 ( Slave fc16-builder ) due to node availability * Stop NATIVE_RSC_B ( fc16-builder ) due to node availability Executing cluster transition: * Pseudo action: MASTER_RSC_A_pre_notify_stop_0 * Resource action: NATIVE_RSC_B stop on fc16-builder * Resource action: NATIVE_RSC_A:0 notify on fc16-builder * Pseudo action: MASTER_RSC_A_confirmed-pre_notify_stop_0 * Pseudo action: MASTER_RSC_A_stop_0 * Resource action: NATIVE_RSC_A:0 stop on fc16-builder * Pseudo action: MASTER_RSC_A_stopped_0 * Pseudo action: MASTER_RSC_A_post_notify_stopped_0 * Pseudo action: MASTER_RSC_A_confirmed-post_notify_stopped_0 Revised cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] Clone Set: MASTER_RSC_A [NATIVE_RSC_A] (promotable) Stopped: [ fc16-builder fc16-builder2 ] NATIVE_RSC_B (ocf::pacemaker:Dummy): Stopped (disabled) diff --git a/cts/scheduler/complex_enforce_colo.summary b/cts/scheduler/complex_enforce_colo.summary index 2b8db6cce8..427fd472bc 100644 --- a/cts/scheduler/complex_enforce_colo.summary +++ b/cts/scheduler/complex_enforce_colo.summary @@ -1,452 +1,452 @@ -3 of 132 resources DISABLED and 0 BLOCKED from being started due to failures +3 of 132 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ rhos6-node1 rhos6-node2 rhos6-node3 ] node1-fence (stonith:fence_xvm): Started rhos6-node1 node2-fence (stonith:fence_xvm): Started rhos6-node2 node3-fence (stonith:fence_xvm): Started rhos6-node3 Clone Set: lb-haproxy-clone [lb-haproxy] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] vip-db (ocf::heartbeat:IPaddr2): Started rhos6-node1 vip-rabbitmq (ocf::heartbeat:IPaddr2): Started rhos6-node2 vip-qpid (ocf::heartbeat:IPaddr2): Started rhos6-node3 vip-keystone (ocf::heartbeat:IPaddr2): Started rhos6-node1 vip-glance (ocf::heartbeat:IPaddr2): Started rhos6-node2 vip-cinder (ocf::heartbeat:IPaddr2): Started rhos6-node3 vip-swift (ocf::heartbeat:IPaddr2): Started rhos6-node1 vip-neutron (ocf::heartbeat:IPaddr2): Started rhos6-node2 vip-nova (ocf::heartbeat:IPaddr2): Started rhos6-node3 vip-horizon (ocf::heartbeat:IPaddr2): Started rhos6-node1 vip-heat (ocf::heartbeat:IPaddr2): Started rhos6-node2 vip-ceilometer (ocf::heartbeat:IPaddr2): Started rhos6-node3 Clone Set: galera-master [galera] (promotable) Masters: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: rabbitmq-server-clone [rabbitmq-server] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: memcached-clone [memcached] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: mongodb-clone [mongodb] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: keystone-clone [keystone] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: glance-fs-clone [glance-fs] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: glance-registry-clone [glance-registry] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: glance-api-clone [glance-api] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] cinder-api (systemd:openstack-cinder-api): Started rhos6-node1 cinder-scheduler (systemd:openstack-cinder-scheduler): Started rhos6-node1 cinder-volume (systemd:openstack-cinder-volume): Started rhos6-node1 Clone Set: swift-fs-clone [swift-fs] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: swift-account-clone [swift-account] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: swift-container-clone [swift-container] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: swift-object-clone [swift-object] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: swift-proxy-clone [swift-proxy] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] swift-object-expirer (systemd:openstack-swift-object-expirer): Started rhos6-node2 Clone Set: neutron-server-clone [neutron-server] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: neutron-scale-clone [neutron-scale] (unique) neutron-scale:0 (ocf::neutron:NeutronScale): Started rhos6-node3 neutron-scale:1 (ocf::neutron:NeutronScale): Started rhos6-node2 neutron-scale:2 (ocf::neutron:NeutronScale): Started rhos6-node1 Clone Set: neutron-ovs-cleanup-clone [neutron-ovs-cleanup] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: neutron-netns-cleanup-clone [neutron-netns-cleanup] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: neutron-openvswitch-agent-clone [neutron-openvswitch-agent] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: neutron-dhcp-agent-clone [neutron-dhcp-agent] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: neutron-l3-agent-clone [neutron-l3-agent] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: neutron-metadata-agent-clone [neutron-metadata-agent] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: nova-consoleauth-clone [nova-consoleauth] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: nova-novncproxy-clone [nova-novncproxy] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: nova-api-clone [nova-api] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: nova-scheduler-clone [nova-scheduler] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: nova-conductor-clone [nova-conductor] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] ceilometer-central (systemd:openstack-ceilometer-central): Started rhos6-node3 Clone Set: ceilometer-collector-clone [ceilometer-collector] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: ceilometer-api-clone [ceilometer-api] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: ceilometer-delay-clone [ceilometer-delay] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: ceilometer-alarm-evaluator-clone [ceilometer-alarm-evaluator] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: ceilometer-alarm-notifier-clone [ceilometer-alarm-notifier] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: ceilometer-notification-clone [ceilometer-notification] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: heat-api-clone [heat-api] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: heat-api-cfn-clone [heat-api-cfn] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: heat-api-cloudwatch-clone [heat-api-cloudwatch] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] heat-engine (systemd:openstack-heat-engine): Started rhos6-node2 Clone Set: horizon-clone [horizon] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Transition Summary: * Stop keystone:0 ( rhos6-node1 ) due to node availability * Stop keystone:1 ( rhos6-node2 ) due to node availability * Stop keystone:2 ( rhos6-node3 ) due to node availability * Stop glance-registry:0 ( rhos6-node1 ) * Stop glance-registry:1 ( rhos6-node2 ) * Stop glance-registry:2 ( rhos6-node3 ) * Stop glance-api:0 ( rhos6-node1 ) * Stop glance-api:1 ( rhos6-node2 ) * Stop glance-api:2 ( rhos6-node3 ) * Stop cinder-api ( rhos6-node1 ) due to unrunnable keystone-clone running * Stop cinder-scheduler ( rhos6-node1 ) due to required cinder-api start * Stop cinder-volume ( rhos6-node1 ) due to colocation with cinder-scheduler * Stop swift-account:0 ( rhos6-node1 ) * Stop swift-account:1 ( rhos6-node2 ) * Stop swift-account:2 ( rhos6-node3 ) * Stop swift-container:0 ( rhos6-node1 ) * Stop swift-container:1 ( rhos6-node2 ) * Stop swift-container:2 ( rhos6-node3 ) * Stop swift-object:0 ( rhos6-node1 ) * Stop swift-object:1 ( rhos6-node2 ) * Stop swift-object:2 ( rhos6-node3 ) * Stop swift-proxy:0 ( rhos6-node1 ) * Stop swift-proxy:1 ( rhos6-node2 ) * Stop swift-proxy:2 ( rhos6-node3 ) * Stop swift-object-expirer ( rhos6-node2 ) due to required swift-proxy-clone running * Stop neutron-server:0 ( rhos6-node1 ) * Stop neutron-server:1 ( rhos6-node2 ) * Stop neutron-server:2 ( rhos6-node3 ) * Stop neutron-scale:0 ( rhos6-node3 ) * Stop neutron-scale:1 ( rhos6-node2 ) * Stop neutron-scale:2 ( rhos6-node1 ) * Stop neutron-ovs-cleanup:0 ( rhos6-node1 ) * Stop neutron-ovs-cleanup:1 ( rhos6-node2 ) * Stop neutron-ovs-cleanup:2 ( rhos6-node3 ) * Stop neutron-netns-cleanup:0 ( rhos6-node1 ) * Stop neutron-netns-cleanup:1 ( rhos6-node2 ) * Stop neutron-netns-cleanup:2 ( rhos6-node3 ) * Stop neutron-openvswitch-agent:0 ( rhos6-node1 ) * Stop neutron-openvswitch-agent:1 ( rhos6-node2 ) * Stop neutron-openvswitch-agent:2 ( rhos6-node3 ) * Stop neutron-dhcp-agent:0 ( rhos6-node1 ) * Stop neutron-dhcp-agent:1 ( rhos6-node2 ) * Stop neutron-dhcp-agent:2 ( rhos6-node3 ) * Stop neutron-l3-agent:0 ( rhos6-node1 ) * Stop neutron-l3-agent:1 ( rhos6-node2 ) * Stop neutron-l3-agent:2 ( rhos6-node3 ) * Stop neutron-metadata-agent:0 ( rhos6-node1 ) * Stop neutron-metadata-agent:1 ( rhos6-node2 ) * Stop neutron-metadata-agent:2 ( rhos6-node3 ) * Stop nova-consoleauth:0 ( rhos6-node1 ) * Stop nova-consoleauth:1 ( rhos6-node2 ) * Stop nova-consoleauth:2 ( rhos6-node3 ) * Stop nova-novncproxy:0 ( rhos6-node1 ) * Stop nova-novncproxy:1 ( rhos6-node2 ) * Stop nova-novncproxy:2 ( rhos6-node3 ) * Stop nova-api:0 ( rhos6-node1 ) * Stop nova-api:1 ( rhos6-node2 ) * Stop nova-api:2 ( rhos6-node3 ) * Stop nova-scheduler:0 ( rhos6-node1 ) * Stop nova-scheduler:1 ( rhos6-node2 ) * Stop nova-scheduler:2 ( rhos6-node3 ) * Stop nova-conductor:0 ( rhos6-node1 ) * Stop nova-conductor:1 ( rhos6-node2 ) * Stop nova-conductor:2 ( rhos6-node3 ) * Stop ceilometer-central ( rhos6-node3 ) due to unrunnable keystone-clone running * Stop ceilometer-collector:0 ( rhos6-node1 ) due to required ceilometer-central start * Stop ceilometer-collector:1 ( rhos6-node2 ) due to required ceilometer-central start * Stop ceilometer-collector:2 ( rhos6-node3 ) due to required ceilometer-central start * Stop ceilometer-api:0 ( rhos6-node1 ) due to required ceilometer-collector:0 start * Stop ceilometer-api:1 ( rhos6-node2 ) due to required ceilometer-collector:1 start * Stop ceilometer-api:2 ( rhos6-node3 ) due to required ceilometer-collector:2 start * Stop ceilometer-delay:0 ( rhos6-node1 ) due to required ceilometer-api:0 start * Stop ceilometer-delay:1 ( rhos6-node2 ) due to required ceilometer-api:1 start * Stop ceilometer-delay:2 ( rhos6-node3 ) due to required ceilometer-api:2 start * Stop ceilometer-alarm-evaluator:0 ( rhos6-node1 ) due to required ceilometer-delay:0 start * Stop ceilometer-alarm-evaluator:1 ( rhos6-node2 ) due to required ceilometer-delay:1 start * Stop ceilometer-alarm-evaluator:2 ( rhos6-node3 ) due to required ceilometer-delay:2 start * Stop ceilometer-alarm-notifier:0 ( rhos6-node1 ) due to required ceilometer-alarm-evaluator:0 start * Stop ceilometer-alarm-notifier:1 ( rhos6-node2 ) due to required ceilometer-alarm-evaluator:1 start * Stop ceilometer-alarm-notifier:2 ( rhos6-node3 ) due to required ceilometer-alarm-evaluator:2 start * Stop ceilometer-notification:0 ( rhos6-node1 ) due to required ceilometer-alarm-notifier:0 start * Stop ceilometer-notification:1 ( rhos6-node2 ) due to required ceilometer-alarm-notifier:1 start * Stop ceilometer-notification:2 ( rhos6-node3 ) due to required ceilometer-alarm-notifier:2 start * Stop heat-api:0 ( rhos6-node1 ) due to required ceilometer-notification:0 start * Stop heat-api:1 ( rhos6-node2 ) due to required ceilometer-notification:1 start * Stop heat-api:2 ( rhos6-node3 ) due to required ceilometer-notification:2 start * Stop heat-api-cfn:0 ( rhos6-node1 ) due to required heat-api:0 start * Stop heat-api-cfn:1 ( rhos6-node2 ) due to required heat-api:1 start * Stop heat-api-cfn:2 ( rhos6-node3 ) due to required heat-api:2 start * Stop heat-api-cloudwatch:0 ( rhos6-node1 ) due to required heat-api-cfn:0 start * Stop heat-api-cloudwatch:1 ( rhos6-node2 ) due to required heat-api-cfn:1 start * Stop heat-api-cloudwatch:2 ( rhos6-node3 ) due to required heat-api-cfn:2 start * Stop heat-engine ( rhos6-node2 ) due to colocation with heat-api-cloudwatch-clone Executing cluster transition: * Pseudo action: glance-api-clone_stop_0 * Resource action: cinder-volume stop on rhos6-node1 * Pseudo action: swift-object-clone_stop_0 * Resource action: swift-object-expirer stop on rhos6-node2 * Pseudo action: neutron-metadata-agent-clone_stop_0 * Pseudo action: nova-conductor-clone_stop_0 * Resource action: heat-engine stop on rhos6-node2 * Resource action: glance-api stop on rhos6-node1 * Resource action: glance-api stop on rhos6-node2 * Resource action: glance-api stop on rhos6-node3 * Pseudo action: glance-api-clone_stopped_0 * Resource action: cinder-scheduler stop on rhos6-node1 * Resource action: swift-object stop on rhos6-node1 * Resource action: swift-object stop on rhos6-node2 * Resource action: swift-object stop on rhos6-node3 * Pseudo action: swift-object-clone_stopped_0 * Pseudo action: swift-proxy-clone_stop_0 * Resource action: neutron-metadata-agent stop on rhos6-node1 * Resource action: neutron-metadata-agent stop on rhos6-node2 * Resource action: neutron-metadata-agent stop on rhos6-node3 * Pseudo action: neutron-metadata-agent-clone_stopped_0 * Resource action: nova-conductor stop on rhos6-node1 * Resource action: nova-conductor stop on rhos6-node2 * Resource action: nova-conductor stop on rhos6-node3 * Pseudo action: nova-conductor-clone_stopped_0 * Pseudo action: heat-api-cloudwatch-clone_stop_0 * Pseudo action: glance-registry-clone_stop_0 * Resource action: cinder-api stop on rhos6-node1 * Pseudo action: swift-container-clone_stop_0 * Resource action: swift-proxy stop on rhos6-node1 * Resource action: swift-proxy stop on rhos6-node2 * Resource action: swift-proxy stop on rhos6-node3 * Pseudo action: swift-proxy-clone_stopped_0 * Pseudo action: neutron-l3-agent-clone_stop_0 * Pseudo action: nova-scheduler-clone_stop_0 * Resource action: heat-api-cloudwatch stop on rhos6-node1 * Resource action: heat-api-cloudwatch stop on rhos6-node2 * Resource action: heat-api-cloudwatch stop on rhos6-node3 * Pseudo action: heat-api-cloudwatch-clone_stopped_0 * Resource action: glance-registry stop on rhos6-node1 * Resource action: glance-registry stop on rhos6-node2 * Resource action: glance-registry stop on rhos6-node3 * Pseudo action: glance-registry-clone_stopped_0 * Resource action: swift-container stop on rhos6-node1 * Resource action: swift-container stop on rhos6-node2 * Resource action: swift-container stop on rhos6-node3 * Pseudo action: swift-container-clone_stopped_0 * Resource action: neutron-l3-agent stop on rhos6-node1 * Resource action: neutron-l3-agent stop on rhos6-node2 * Resource action: neutron-l3-agent stop on rhos6-node3 * Pseudo action: neutron-l3-agent-clone_stopped_0 * Resource action: nova-scheduler stop on rhos6-node1 * Resource action: nova-scheduler stop on rhos6-node2 * Resource action: nova-scheduler stop on rhos6-node3 * Pseudo action: nova-scheduler-clone_stopped_0 * Pseudo action: heat-api-cfn-clone_stop_0 * Pseudo action: swift-account-clone_stop_0 * Pseudo action: neutron-dhcp-agent-clone_stop_0 * Pseudo action: nova-api-clone_stop_0 * Resource action: heat-api-cfn stop on rhos6-node1 * Resource action: heat-api-cfn stop on rhos6-node2 * Resource action: heat-api-cfn stop on rhos6-node3 * Pseudo action: heat-api-cfn-clone_stopped_0 * Resource action: swift-account stop on rhos6-node1 * Resource action: swift-account stop on rhos6-node2 * Resource action: swift-account stop on rhos6-node3 * Pseudo action: swift-account-clone_stopped_0 * Resource action: neutron-dhcp-agent stop on rhos6-node1 * Resource action: neutron-dhcp-agent stop on rhos6-node2 * Resource action: neutron-dhcp-agent stop on rhos6-node3 * Pseudo action: neutron-dhcp-agent-clone_stopped_0 * Resource action: nova-api stop on rhos6-node1 * Resource action: nova-api stop on rhos6-node2 * Resource action: nova-api stop on rhos6-node3 * Pseudo action: nova-api-clone_stopped_0 * Pseudo action: heat-api-clone_stop_0 * Pseudo action: neutron-openvswitch-agent-clone_stop_0 * Pseudo action: nova-novncproxy-clone_stop_0 * Resource action: heat-api stop on rhos6-node1 * Resource action: heat-api stop on rhos6-node2 * Resource action: heat-api stop on rhos6-node3 * Pseudo action: heat-api-clone_stopped_0 * Resource action: neutron-openvswitch-agent stop on rhos6-node1 * Resource action: neutron-openvswitch-agent stop on rhos6-node2 * Resource action: neutron-openvswitch-agent stop on rhos6-node3 * Pseudo action: neutron-openvswitch-agent-clone_stopped_0 * Resource action: nova-novncproxy stop on rhos6-node1 * Resource action: nova-novncproxy stop on rhos6-node2 * Resource action: nova-novncproxy stop on rhos6-node3 * Pseudo action: nova-novncproxy-clone_stopped_0 * Pseudo action: ceilometer-notification-clone_stop_0 * Pseudo action: neutron-netns-cleanup-clone_stop_0 * Pseudo action: nova-consoleauth-clone_stop_0 * Resource action: ceilometer-notification stop on rhos6-node1 * Resource action: ceilometer-notification stop on rhos6-node2 * Resource action: ceilometer-notification stop on rhos6-node3 * Pseudo action: ceilometer-notification-clone_stopped_0 * Resource action: neutron-netns-cleanup stop on rhos6-node1 * Resource action: neutron-netns-cleanup stop on rhos6-node2 * Resource action: neutron-netns-cleanup stop on rhos6-node3 * Pseudo action: neutron-netns-cleanup-clone_stopped_0 * Resource action: nova-consoleauth stop on rhos6-node1 * Resource action: nova-consoleauth stop on rhos6-node2 * Resource action: nova-consoleauth stop on rhos6-node3 * Pseudo action: nova-consoleauth-clone_stopped_0 * Pseudo action: ceilometer-alarm-notifier-clone_stop_0 * Pseudo action: neutron-ovs-cleanup-clone_stop_0 * Resource action: ceilometer-alarm-notifier stop on rhos6-node1 * Resource action: ceilometer-alarm-notifier stop on rhos6-node2 * Resource action: ceilometer-alarm-notifier stop on rhos6-node3 * Pseudo action: ceilometer-alarm-notifier-clone_stopped_0 * Resource action: neutron-ovs-cleanup stop on rhos6-node1 * Resource action: neutron-ovs-cleanup stop on rhos6-node2 * Resource action: neutron-ovs-cleanup stop on rhos6-node3 * Pseudo action: neutron-ovs-cleanup-clone_stopped_0 * Pseudo action: ceilometer-alarm-evaluator-clone_stop_0 * Pseudo action: neutron-scale-clone_stop_0 * Resource action: ceilometer-alarm-evaluator stop on rhos6-node1 * Resource action: ceilometer-alarm-evaluator stop on rhos6-node2 * Resource action: ceilometer-alarm-evaluator stop on rhos6-node3 * Pseudo action: ceilometer-alarm-evaluator-clone_stopped_0 * Resource action: neutron-scale:0 stop on rhos6-node3 * Resource action: neutron-scale:1 stop on rhos6-node2 * Resource action: neutron-scale:2 stop on rhos6-node1 * Pseudo action: neutron-scale-clone_stopped_0 * Pseudo action: ceilometer-delay-clone_stop_0 * Pseudo action: neutron-server-clone_stop_0 * Resource action: ceilometer-delay stop on rhos6-node1 * Resource action: ceilometer-delay stop on rhos6-node2 * Resource action: ceilometer-delay stop on rhos6-node3 * Pseudo action: ceilometer-delay-clone_stopped_0 * Resource action: neutron-server stop on rhos6-node1 * Resource action: neutron-server stop on rhos6-node2 * Resource action: neutron-server stop on rhos6-node3 * Pseudo action: neutron-server-clone_stopped_0 * Pseudo action: ceilometer-api-clone_stop_0 * Resource action: ceilometer-api stop on rhos6-node1 * Resource action: ceilometer-api stop on rhos6-node2 * Resource action: ceilometer-api stop on rhos6-node3 * Pseudo action: ceilometer-api-clone_stopped_0 * Pseudo action: ceilometer-collector-clone_stop_0 * Resource action: ceilometer-collector stop on rhos6-node1 * Resource action: ceilometer-collector stop on rhos6-node2 * Resource action: ceilometer-collector stop on rhos6-node3 * Pseudo action: ceilometer-collector-clone_stopped_0 * Resource action: ceilometer-central stop on rhos6-node3 * Pseudo action: keystone-clone_stop_0 * Resource action: keystone stop on rhos6-node1 * Resource action: keystone stop on rhos6-node2 * Resource action: keystone stop on rhos6-node3 * Pseudo action: keystone-clone_stopped_0 Revised cluster status: Online: [ rhos6-node1 rhos6-node2 rhos6-node3 ] node1-fence (stonith:fence_xvm): Started rhos6-node1 node2-fence (stonith:fence_xvm): Started rhos6-node2 node3-fence (stonith:fence_xvm): Started rhos6-node3 Clone Set: lb-haproxy-clone [lb-haproxy] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] vip-db (ocf::heartbeat:IPaddr2): Started rhos6-node1 vip-rabbitmq (ocf::heartbeat:IPaddr2): Started rhos6-node2 vip-qpid (ocf::heartbeat:IPaddr2): Started rhos6-node3 vip-keystone (ocf::heartbeat:IPaddr2): Started rhos6-node1 vip-glance (ocf::heartbeat:IPaddr2): Started rhos6-node2 vip-cinder (ocf::heartbeat:IPaddr2): Started rhos6-node3 vip-swift (ocf::heartbeat:IPaddr2): Started rhos6-node1 vip-neutron (ocf::heartbeat:IPaddr2): Started rhos6-node2 vip-nova (ocf::heartbeat:IPaddr2): Started rhos6-node3 vip-horizon (ocf::heartbeat:IPaddr2): Started rhos6-node1 vip-heat (ocf::heartbeat:IPaddr2): Started rhos6-node2 vip-ceilometer (ocf::heartbeat:IPaddr2): Started rhos6-node3 Clone Set: galera-master [galera] (promotable) Masters: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: rabbitmq-server-clone [rabbitmq-server] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: memcached-clone [memcached] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: mongodb-clone [mongodb] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: keystone-clone [keystone] Stopped (disabled): [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: glance-fs-clone [glance-fs] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: glance-registry-clone [glance-registry] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: glance-api-clone [glance-api] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] cinder-api (systemd:openstack-cinder-api): Stopped cinder-scheduler (systemd:openstack-cinder-scheduler): Stopped cinder-volume (systemd:openstack-cinder-volume): Stopped Clone Set: swift-fs-clone [swift-fs] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: swift-account-clone [swift-account] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: swift-container-clone [swift-container] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: swift-object-clone [swift-object] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: swift-proxy-clone [swift-proxy] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] swift-object-expirer (systemd:openstack-swift-object-expirer): Stopped Clone Set: neutron-server-clone [neutron-server] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: neutron-scale-clone [neutron-scale] (unique) neutron-scale:0 (ocf::neutron:NeutronScale): Stopped neutron-scale:1 (ocf::neutron:NeutronScale): Stopped neutron-scale:2 (ocf::neutron:NeutronScale): Stopped Clone Set: neutron-ovs-cleanup-clone [neutron-ovs-cleanup] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: neutron-netns-cleanup-clone [neutron-netns-cleanup] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: neutron-openvswitch-agent-clone [neutron-openvswitch-agent] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: neutron-dhcp-agent-clone [neutron-dhcp-agent] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: neutron-l3-agent-clone [neutron-l3-agent] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: neutron-metadata-agent-clone [neutron-metadata-agent] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: nova-consoleauth-clone [nova-consoleauth] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: nova-novncproxy-clone [nova-novncproxy] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: nova-api-clone [nova-api] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: nova-scheduler-clone [nova-scheduler] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: nova-conductor-clone [nova-conductor] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] ceilometer-central (systemd:openstack-ceilometer-central): Stopped Clone Set: ceilometer-collector-clone [ceilometer-collector] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: ceilometer-api-clone [ceilometer-api] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: ceilometer-delay-clone [ceilometer-delay] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: ceilometer-alarm-evaluator-clone [ceilometer-alarm-evaluator] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: ceilometer-alarm-notifier-clone [ceilometer-alarm-notifier] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: ceilometer-notification-clone [ceilometer-notification] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: heat-api-clone [heat-api] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: heat-api-cfn-clone [heat-api-cfn] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] Clone Set: heat-api-cloudwatch-clone [heat-api-cloudwatch] Stopped: [ rhos6-node1 rhos6-node2 rhos6-node3 ] heat-engine (systemd:openstack-heat-engine): Stopped Clone Set: horizon-clone [horizon] Started: [ rhos6-node1 rhos6-node2 rhos6-node3 ] diff --git a/cts/scheduler/container-is-remote-node.summary b/cts/scheduler/container-is-remote-node.summary index 33a45afd0b..fc9c0ca82b 100644 --- a/cts/scheduler/container-is-remote-node.summary +++ b/cts/scheduler/container-is-remote-node.summary @@ -1,56 +1,56 @@ -3 of 19 resources DISABLED and 0 BLOCKED from being started due to failures +3 of 19 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ lama2 lama3 ] GuestOnline: [ RNVM1:VM1 ] restofencelama2 (stonith:fence_ipmilan): Started lama3 restofencelama3 (stonith:fence_ipmilan): Started lama2 Clone Set: dlm-clone [dlm] Started: [ lama2 lama3 ] Stopped: [ RNVM1 ] Clone Set: clvmd-clone [clvmd] Started: [ lama2 lama3 ] Stopped: [ RNVM1 ] Clone Set: gfs2-lv_1_1-clone [gfs2-lv_1_1] Started: [ lama2 lama3 ] Stopped: [ RNVM1 ] Clone Set: gfs2-lv_1_2-clone [gfs2-lv_1_2] Stopped (disabled): [ lama2 lama3 RNVM1 ] VM1 (ocf::heartbeat:VirtualDomain): Started lama2 Resource Group: RES1 FSdata1 (ocf::heartbeat:Filesystem): Started RNVM1 RES1-IP (ocf::heartbeat:IPaddr2): Started RNVM1 res-rsyslog (ocf::heartbeat:rsyslog.test): Started RNVM1 Transition Summary: Executing cluster transition: * Resource action: dlm monitor on RNVM1 * Resource action: clvmd monitor on RNVM1 * Resource action: gfs2-lv_1_1 monitor on RNVM1 * Resource action: gfs2-lv_1_2 monitor on RNVM1 Revised cluster status: Online: [ lama2 lama3 ] GuestOnline: [ RNVM1:VM1 ] restofencelama2 (stonith:fence_ipmilan): Started lama3 restofencelama3 (stonith:fence_ipmilan): Started lama2 Clone Set: dlm-clone [dlm] Started: [ lama2 lama3 ] Stopped: [ RNVM1 ] Clone Set: clvmd-clone [clvmd] Started: [ lama2 lama3 ] Stopped: [ RNVM1 ] Clone Set: gfs2-lv_1_1-clone [gfs2-lv_1_1] Started: [ lama2 lama3 ] Stopped: [ RNVM1 ] Clone Set: gfs2-lv_1_2-clone [gfs2-lv_1_2] Stopped (disabled): [ lama2 lama3 RNVM1 ] VM1 (ocf::heartbeat:VirtualDomain): Started lama2 Resource Group: RES1 FSdata1 (ocf::heartbeat:Filesystem): Started RNVM1 RES1-IP (ocf::heartbeat:IPaddr2): Started RNVM1 res-rsyslog (ocf::heartbeat:rsyslog.test): Started RNVM1 diff --git a/cts/scheduler/enforce-colo1.summary b/cts/scheduler/enforce-colo1.summary index 9a93723195..46f808454a 100644 --- a/cts/scheduler/enforce-colo1.summary +++ b/cts/scheduler/enforce-colo1.summary @@ -1,36 +1,36 @@ -3 of 6 resources DISABLED and 0 BLOCKED from being started due to failures +3 of 6 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ rhel7-auto1 rhel7-auto2 rhel7-auto3 ] shooter (stonith:fence_xvm): Started rhel7-auto2 engine (ocf::heartbeat:Dummy): Started rhel7-auto3 Clone Set: keystone-clone [keystone] Started: [ rhel7-auto1 rhel7-auto2 rhel7-auto3 ] central (ocf::heartbeat:Dummy): Started rhel7-auto3 Transition Summary: * Stop engine ( rhel7-auto3 ) due to colocation with central * Stop keystone:0 ( rhel7-auto2 ) due to node availability * Stop keystone:1 ( rhel7-auto3 ) due to node availability * Stop keystone:2 ( rhel7-auto1 ) due to node availability * Stop central ( rhel7-auto3 ) due to unrunnable keystone-clone running Executing cluster transition: * Resource action: engine stop on rhel7-auto3 * Resource action: central stop on rhel7-auto3 * Pseudo action: keystone-clone_stop_0 * Resource action: keystone stop on rhel7-auto2 * Resource action: keystone stop on rhel7-auto3 * Resource action: keystone stop on rhel7-auto1 * Pseudo action: keystone-clone_stopped_0 Revised cluster status: Online: [ rhel7-auto1 rhel7-auto2 rhel7-auto3 ] shooter (stonith:fence_xvm): Started rhel7-auto2 engine (ocf::heartbeat:Dummy): Stopped Clone Set: keystone-clone [keystone] Stopped (disabled): [ rhel7-auto1 rhel7-auto2 rhel7-auto3 ] central (ocf::heartbeat:Dummy): Stopped diff --git a/cts/scheduler/expire-non-blocked-failure.summary b/cts/scheduler/expire-non-blocked-failure.summary index 2c6e703846..aab1ffe621 100644 --- a/cts/scheduler/expire-non-blocked-failure.summary +++ b/cts/scheduler/expire-non-blocked-failure.summary @@ -1,20 +1,21 @@ +0 of 3 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc_stonith (stonith:null): Started node1 rsc1 (ocf::pacemaker:Dummy): FAILED node2 (blocked) rsc2 (ocf::pacemaker:Dummy): Started node1 Transition Summary: Executing cluster transition: * Cluster action: clear_failcount for rsc2 on node1 Revised cluster status: Online: [ node1 node2 ] rsc_stonith (stonith:null): Started node1 rsc1 (ocf::pacemaker:Dummy): FAILED node2 (blocked) rsc2 (ocf::pacemaker:Dummy): Started node1 diff --git a/cts/scheduler/failcount-block.summary b/cts/scheduler/failcount-block.summary index 6d957aae4e..a5a1a36094 100644 --- a/cts/scheduler/failcount-block.summary +++ b/cts/scheduler/failcount-block.summary @@ -1,35 +1,36 @@ +0 of 5 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ pcmk-1 ] OFFLINE: [ pcmk-4 ] rsc_pcmk-1 (ocf::heartbeat:IPaddr2): Started pcmk-1 rsc_pcmk-2 (ocf::heartbeat:IPaddr2): FAILED pcmk-1 (blocked) rsc_pcmk-3 (ocf::heartbeat:IPaddr2): Stopped rsc_pcmk-4 (ocf::heartbeat:IPaddr2): Stopped rsc_pcmk-5 (ocf::heartbeat:IPaddr2): Started pcmk-1 Transition Summary: * Start rsc_pcmk-3 ( pcmk-1 ) * Start rsc_pcmk-4 ( pcmk-1 ) Executing cluster transition: * Resource action: rsc_pcmk-1 monitor=5000 on pcmk-1 * Cluster action: clear_failcount for rsc_pcmk-1 on pcmk-1 * Resource action: rsc_pcmk-3 start on pcmk-1 * Cluster action: clear_failcount for rsc_pcmk-3 on pcmk-1 * Resource action: rsc_pcmk-4 start on pcmk-1 * Cluster action: clear_failcount for rsc_pcmk-5 on pcmk-1 * Resource action: rsc_pcmk-3 monitor=5000 on pcmk-1 * Resource action: rsc_pcmk-4 monitor=5000 on pcmk-1 Revised cluster status: Online: [ pcmk-1 ] OFFLINE: [ pcmk-4 ] rsc_pcmk-1 (ocf::heartbeat:IPaddr2): Started pcmk-1 rsc_pcmk-2 (ocf::heartbeat:IPaddr2): FAILED pcmk-1 (blocked) rsc_pcmk-3 (ocf::heartbeat:IPaddr2): Started pcmk-1 rsc_pcmk-4 (ocf::heartbeat:IPaddr2): Started pcmk-1 rsc_pcmk-5 (ocf::heartbeat:IPaddr2): Started pcmk-1 diff --git a/cts/scheduler/group-stop-ordering.summary b/cts/scheduler/group-stop-ordering.summary index 0ec8eb6299..df9e7a3ca9 100644 --- a/cts/scheduler/group-stop-ordering.summary +++ b/cts/scheduler/group-stop-ordering.summary @@ -1,25 +1,26 @@ +0 of 5 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ fastvm-rhel-7-5-73 fastvm-rhel-7-5-74 ] fence-fastvm-rhel-7-5-73 (stonith:fence_xvm): Started fastvm-rhel-7-5-74 fence-fastvm-rhel-7-5-74 (stonith:fence_xvm): Started fastvm-rhel-7-5-73 outside_resource (ocf::pacemaker:Dummy): FAILED fastvm-rhel-7-5-73 (blocked) Resource Group: grp inside_resource_2 (ocf::pacemaker:Dummy): Started fastvm-rhel-7-5-74 inside_resource_3 (ocf::pacemaker:Dummy): Started fastvm-rhel-7-5-74 Transition Summary: Executing cluster transition: Revised cluster status: Online: [ fastvm-rhel-7-5-73 fastvm-rhel-7-5-74 ] fence-fastvm-rhel-7-5-73 (stonith:fence_xvm): Started fastvm-rhel-7-5-74 fence-fastvm-rhel-7-5-74 (stonith:fence_xvm): Started fastvm-rhel-7-5-73 outside_resource (ocf::pacemaker:Dummy): FAILED fastvm-rhel-7-5-73 (blocked) Resource Group: grp inside_resource_2 (ocf::pacemaker:Dummy): Started fastvm-rhel-7-5-74 inside_resource_3 (ocf::pacemaker:Dummy): Started fastvm-rhel-7-5-74 diff --git a/cts/scheduler/group11.summary b/cts/scheduler/group11.summary index 41b5994782..8ebad4cc18 100644 --- a/cts/scheduler/group11.summary +++ b/cts/scheduler/group11.summary @@ -1,29 +1,29 @@ -2 of 3 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 3 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 ] Resource Group: group1 rsc1 (ocf::heartbeat:apache): Started node1 rsc2 (ocf::heartbeat:apache): Started node1 (disabled) rsc3 (ocf::heartbeat:apache): Started node1 Transition Summary: * Stop rsc2 ( node1 ) due to node availability * Stop rsc3 ( node1 ) due to node availability Executing cluster transition: * Pseudo action: group1_stop_0 * Resource action: rsc3 stop on node1 * Resource action: rsc2 stop on node1 * Pseudo action: group1_stopped_0 * Pseudo action: group1_start_0 Revised cluster status: Online: [ node1 ] Resource Group: group1 rsc1 (ocf::heartbeat:apache): Started node1 rsc2 (ocf::heartbeat:apache): Stopped (disabled) rsc3 (ocf::heartbeat:apache): Stopped diff --git a/cts/scheduler/intervals.summary b/cts/scheduler/intervals.summary index 83603dc37f..69daed7181 100644 --- a/cts/scheduler/intervals.summary +++ b/cts/scheduler/intervals.summary @@ -1,48 +1,49 @@ Using the original execution date of: 2018-03-21 23:12:42Z +0 of 7 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ rhel7-1 rhel7-2 rhel7-3 rhel7-4 rhel7-5 ] Fencing (stonith:fence_xvm): Started rhel7-1 rsc1 (ocf::pacemaker:Dummy): Started rhel7-2 rsc2 (ocf::pacemaker:Dummy): Stopped rsc3 (ocf::pacemaker:Dummy): Started rhel7-4 rsc4 (ocf::pacemaker:Dummy): FAILED rhel7-5 (blocked) rsc5 (ocf::pacemaker:Dummy): Started rhel7-1 rsc6 (ocf::pacemaker:Dummy): Started rhel7-2 Transition Summary: * Start rsc2 ( rhel7-3 ) * Move rsc5 ( rhel7-1 -> rhel7-2 ) * Move rsc6 ( rhel7-2 -> rhel7-1 ) Executing cluster transition: * Resource action: rsc2 monitor on rhel7-5 * Resource action: rsc2 monitor on rhel7-4 * Resource action: rsc2 monitor on rhel7-3 * Resource action: rsc2 monitor on rhel7-2 * Resource action: rsc2 monitor on rhel7-1 * Resource action: rsc5 stop on rhel7-1 * Resource action: rsc5 cancel=25000 on rhel7-2 * Resource action: rsc6 stop on rhel7-2 * Resource action: rsc2 start on rhel7-3 * Resource action: rsc5 monitor=25000 on rhel7-1 * Resource action: rsc5 start on rhel7-2 * Resource action: rsc6 start on rhel7-1 * Resource action: rsc2 monitor=90000 on rhel7-3 * Resource action: rsc2 monitor=40000 on rhel7-3 * Resource action: rsc5 monitor=20000 on rhel7-2 * Resource action: rsc6 monitor=28000 on rhel7-1 Using the original execution date of: 2018-03-21 23:12:42Z Revised cluster status: Online: [ rhel7-1 rhel7-2 rhel7-3 rhel7-4 rhel7-5 ] Fencing (stonith:fence_xvm): Started rhel7-1 rsc1 (ocf::pacemaker:Dummy): Started rhel7-2 rsc2 (ocf::pacemaker:Dummy): Started rhel7-3 rsc3 (ocf::pacemaker:Dummy): Started rhel7-4 rsc4 (ocf::pacemaker:Dummy): FAILED rhel7-5 (blocked) rsc5 (ocf::pacemaker:Dummy): Started rhel7-2 rsc6 (ocf::pacemaker:Dummy): Started rhel7-1 diff --git a/cts/scheduler/load-stopped-loop-2.summary b/cts/scheduler/load-stopped-loop-2.summary index 5a26b60a90..09e85cc79b 100644 --- a/cts/scheduler/load-stopped-loop-2.summary +++ b/cts/scheduler/load-stopped-loop-2.summary @@ -1,111 +1,111 @@ -4 of 25 resources DISABLED and 0 BLOCKED from being started due to failures +4 of 25 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ xfc0 xfc1 xfc2 xfc3 ] Clone Set: cl_glusterd [p_glusterd] Started: [ xfc0 xfc1 xfc2 xfc3 ] Clone Set: cl_p_bl_glusterfs [p_bl_glusterfs] Started: [ xfc0 xfc1 xfc2 xfc3 ] xu-test8 (ocf::heartbeat:Xen): Started xfc3 xu-test1 (ocf::heartbeat:Xen): Started xfc3 xu-test10 (ocf::heartbeat:Xen): Started xfc3 xu-test11 (ocf::heartbeat:Xen): Started xfc3 xu-test12 (ocf::heartbeat:Xen): Started xfc2 xu-test13 (ocf::heartbeat:Xen): Stopped xu-test14 (ocf::heartbeat:Xen): Stopped (disabled) xu-test15 (ocf::heartbeat:Xen): Stopped (disabled) xu-test16 (ocf::heartbeat:Xen): Stopped (disabled) xu-test17 (ocf::heartbeat:Xen): Stopped (disabled) xu-test2 (ocf::heartbeat:Xen): Started xfc3 xu-test3 (ocf::heartbeat:Xen): Started xfc1 xu-test4 (ocf::heartbeat:Xen): Started xfc0 xu-test5 (ocf::heartbeat:Xen): Started xfc2 xu-test6 (ocf::heartbeat:Xen): Started xfc3 xu-test7 (ocf::heartbeat:Xen): Started xfc1 xu-test9 (ocf::heartbeat:Xen): Started xfc0 Transition Summary: * Migrate xu-test12 ( xfc2 -> xfc3 ) * Migrate xu-test2 ( xfc3 -> xfc1 ) * Migrate xu-test3 ( xfc1 -> xfc0 ) * Migrate xu-test4 ( xfc0 -> xfc2 ) * Migrate xu-test5 ( xfc2 -> xfc3 ) * Migrate xu-test6 ( xfc3 -> xfc1 ) * Migrate xu-test7 ( xfc1 -> xfc0 ) * Migrate xu-test9 ( xfc0 -> xfc2 ) * Start xu-test13 ( xfc3 ) Executing cluster transition: * Resource action: xu-test4 migrate_to on xfc0 * Resource action: xu-test5 migrate_to on xfc2 * Resource action: xu-test6 migrate_to on xfc3 * Resource action: xu-test7 migrate_to on xfc1 * Resource action: xu-test9 migrate_to on xfc0 * Resource action: xu-test4 migrate_from on xfc2 * Resource action: xu-test4 stop on xfc0 * Resource action: xu-test5 migrate_from on xfc3 * Resource action: xu-test5 stop on xfc2 * Resource action: xu-test6 migrate_from on xfc1 * Resource action: xu-test6 stop on xfc3 * Resource action: xu-test7 migrate_from on xfc0 * Resource action: xu-test7 stop on xfc1 * Resource action: xu-test9 migrate_from on xfc2 * Resource action: xu-test9 stop on xfc0 * Pseudo action: load_stopped_xfc0 * Resource action: xu-test3 migrate_to on xfc1 * Pseudo action: xu-test7_start_0 * Resource action: xu-test3 migrate_from on xfc0 * Resource action: xu-test3 stop on xfc1 * Resource action: xu-test7 monitor=10000 on xfc0 * Pseudo action: load_stopped_xfc1 * Resource action: xu-test2 migrate_to on xfc3 * Pseudo action: xu-test3_start_0 * Pseudo action: xu-test6_start_0 * Resource action: xu-test2 migrate_from on xfc1 * Resource action: xu-test2 stop on xfc3 * Resource action: xu-test3 monitor=10000 on xfc0 * Resource action: xu-test6 monitor=10000 on xfc1 * Pseudo action: load_stopped_xfc3 * Resource action: xu-test12 migrate_to on xfc2 * Pseudo action: xu-test2_start_0 * Pseudo action: xu-test5_start_0 * Resource action: xu-test13 start on xfc3 * Resource action: xu-test12 migrate_from on xfc3 * Resource action: xu-test12 stop on xfc2 * Resource action: xu-test2 monitor=10000 on xfc1 * Resource action: xu-test5 monitor=10000 on xfc3 * Resource action: xu-test13 monitor=10000 on xfc3 * Pseudo action: load_stopped_xfc2 * Pseudo action: xu-test12_start_0 * Pseudo action: xu-test4_start_0 * Pseudo action: xu-test9_start_0 * Resource action: xu-test12 monitor=10000 on xfc3 * Resource action: xu-test4 monitor=10000 on xfc2 * Resource action: xu-test9 monitor=10000 on xfc2 Revised cluster status: Online: [ xfc0 xfc1 xfc2 xfc3 ] Clone Set: cl_glusterd [p_glusterd] Started: [ xfc0 xfc1 xfc2 xfc3 ] Clone Set: cl_p_bl_glusterfs [p_bl_glusterfs] Started: [ xfc0 xfc1 xfc2 xfc3 ] xu-test8 (ocf::heartbeat:Xen): Started xfc3 xu-test1 (ocf::heartbeat:Xen): Started xfc3 xu-test10 (ocf::heartbeat:Xen): Started xfc3 xu-test11 (ocf::heartbeat:Xen): Started xfc3 xu-test12 (ocf::heartbeat:Xen): Started xfc3 xu-test13 (ocf::heartbeat:Xen): Started xfc3 xu-test14 (ocf::heartbeat:Xen): Stopped (disabled) xu-test15 (ocf::heartbeat:Xen): Stopped (disabled) xu-test16 (ocf::heartbeat:Xen): Stopped (disabled) xu-test17 (ocf::heartbeat:Xen): Stopped (disabled) xu-test2 (ocf::heartbeat:Xen): Started xfc1 xu-test3 (ocf::heartbeat:Xen): Started xfc0 xu-test4 (ocf::heartbeat:Xen): Started xfc2 xu-test5 (ocf::heartbeat:Xen): Started xfc3 xu-test6 (ocf::heartbeat:Xen): Started xfc1 xu-test7 (ocf::heartbeat:Xen): Started xfc0 xu-test9 (ocf::heartbeat:Xen): Started xfc2 diff --git a/cts/scheduler/load-stopped-loop.summary b/cts/scheduler/load-stopped-loop.summary index 9106e69a9f..908fbd3ee8 100644 --- a/cts/scheduler/load-stopped-loop.summary +++ b/cts/scheduler/load-stopped-loop.summary @@ -1,334 +1,334 @@ -32 of 308 resources DISABLED and 0 BLOCKED from being started due to failures +32 of 308 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ mgmt01 v03-a v03-b ] stonith-v02-a (stonith:fence_ipmilan): Stopped (disabled) stonith-v02-b (stonith:fence_ipmilan): Stopped (disabled) stonith-v02-c (stonith:fence_ipmilan): Stopped (disabled) stonith-v02-d (stonith:fence_ipmilan): Stopped (disabled) stonith-mgmt01 (stonith:fence_xvm): Started v03-b stonith-mgmt02 (stonith:meatware): Started mgmt01 stonith-v03-c (stonith:fence_ipmilan): Stopped (disabled) stonith-v03-a (stonith:fence_ipmilan): Started v03-b stonith-v03-b (stonith:fence_ipmilan): Started v03-a stonith-v03-d (stonith:fence_ipmilan): Stopped (disabled) Clone Set: cl-clvmd [clvmd] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-dlm [dlm] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-iscsid [iscsid] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-libvirtd [libvirtd] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-multipathd [multipathd] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-node-params [node-params] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan1-if [vlan1-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan101-if [vlan101-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan102-if [vlan102-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan103-if [vlan103-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan104-if [vlan104-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan3-if [vlan3-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan4-if [vlan4-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan5-if [vlan5-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan900-if [vlan900-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan909-if [vlan909-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-libvirt-images-fs [libvirt-images-fs] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-libvirt-install-fs [libvirt-install-fs] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-0-iscsi [vds-ok-pool-0-iscsi] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-0-vg [vds-ok-pool-0-vg] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-1-iscsi [vds-ok-pool-1-iscsi] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-1-vg [vds-ok-pool-1-vg] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-libvirt-images-pool [libvirt-images-pool] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vds-ok-pool-0-pool [vds-ok-pool-0-pool] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vds-ok-pool-1-pool [vds-ok-pool-1-pool] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] git.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd01-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a vd01-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a vd01-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b vd01-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped vd02-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd02-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd02-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd02-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) f13-x64-devel.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b eu2.ca-pages.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) zakaz.transferrus.ru-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) Clone Set: cl-vlan200-if [vlan200-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] lenny-x32-devel-vm (ocf::vds-ok:VirtualDomain): Started v03-a dist.express-consult.org-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) eu1.ca-pages.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) gotin-bbb-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) maxb-c55-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) metae.ru-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) rodovoepomestie.ru-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) ubuntu9.10-gotin-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) c5-x64-devel.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b Clone Set: cl-mcast-test-net [mcast-test-net] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] dist.fly-uni.org-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) ktstudio.net-vm (ocf::vds-ok:VirtualDomain): Started v03-a cloudsrv.credo-dialogue.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b c6-x64-devel.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a lustre01-right.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b lustre02-right.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a lustre03-left.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b lustre03-right.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a lustre04-left.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b lustre04-right.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a Clone Set: cl-mcast-anbriz-net [mcast-anbriz-net] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] gw.anbriz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b license.anbriz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b terminal.anbriz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) lustre01-left.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a lustre02-left.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b test-01.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a Clone Set: cl-libvirt-qpid [libvirt-qpid] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] gw.gleb.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) gw.gotin.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) terminal0.anbriz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a Clone Set: cl-mcast-gleb-net [mcast-gleb-net] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Transition Summary: * Reload vds-ok-pool-0-iscsi:0 ( mgmt01 ) * Reload vds-ok-pool-0-iscsi:1 ( v03-b ) * Reload vds-ok-pool-0-iscsi:2 ( v03-a ) * Reload vds-ok-pool-1-iscsi:0 ( mgmt01 ) * Reload vds-ok-pool-1-iscsi:1 ( v03-b ) * Reload vds-ok-pool-1-iscsi:2 ( v03-a ) * Restart stonith-v03-b ( v03-a ) due to resource definition change * Restart stonith-v03-a ( v03-b ) due to resource definition change * Migrate license.anbriz.vds-ok.com-vm ( v03-b -> v03-a ) * Migrate terminal0.anbriz.vds-ok.com-vm ( v03-a -> v03-b ) * Start vd01-d.cdev.ttc.prague.cz.vds-ok.com-vm ( v03-a ) Executing cluster transition: * Resource action: vds-ok-pool-0-iscsi:1 reload on mgmt01 * Resource action: vds-ok-pool-0-iscsi:1 monitor=30000 on mgmt01 * Resource action: vds-ok-pool-0-iscsi:0 reload on v03-b * Resource action: vds-ok-pool-0-iscsi:0 monitor=30000 on v03-b * Resource action: vds-ok-pool-0-iscsi:2 reload on v03-a * Resource action: vds-ok-pool-0-iscsi:2 monitor=30000 on v03-a * Resource action: vds-ok-pool-1-iscsi:1 reload on mgmt01 * Resource action: vds-ok-pool-1-iscsi:1 monitor=30000 on mgmt01 * Resource action: vds-ok-pool-1-iscsi:0 reload on v03-b * Resource action: vds-ok-pool-1-iscsi:0 monitor=30000 on v03-b * Resource action: vds-ok-pool-1-iscsi:2 reload on v03-a * Resource action: vds-ok-pool-1-iscsi:2 monitor=30000 on v03-a * Resource action: stonith-v03-b stop on v03-a * Resource action: stonith-v03-b start on v03-a * Resource action: stonith-v03-b monitor=60000 on v03-a * Resource action: stonith-v03-a stop on v03-b * Resource action: stonith-v03-a start on v03-b * Resource action: stonith-v03-a monitor=60000 on v03-b * Resource action: terminal0.anbriz.vds-ok.com-vm migrate_to on v03-a * Pseudo action: load_stopped_mgmt01 * Resource action: terminal0.anbriz.vds-ok.com-vm migrate_from on v03-b * Resource action: terminal0.anbriz.vds-ok.com-vm stop on v03-a * Pseudo action: load_stopped_v03-a * Resource action: license.anbriz.vds-ok.com-vm migrate_to on v03-b * Resource action: vd01-d.cdev.ttc.prague.cz.vds-ok.com-vm start on v03-a * Resource action: license.anbriz.vds-ok.com-vm migrate_from on v03-a * Resource action: license.anbriz.vds-ok.com-vm stop on v03-b * Resource action: vd01-d.cdev.ttc.prague.cz.vds-ok.com-vm monitor=10000 on v03-a * Pseudo action: load_stopped_v03-b * Pseudo action: license.anbriz.vds-ok.com-vm_start_0 * Pseudo action: terminal0.anbriz.vds-ok.com-vm_start_0 * Resource action: license.anbriz.vds-ok.com-vm monitor=10000 on v03-a * Resource action: terminal0.anbriz.vds-ok.com-vm monitor=10000 on v03-b Revised cluster status: Online: [ mgmt01 v03-a v03-b ] stonith-v02-a (stonith:fence_ipmilan): Stopped (disabled) stonith-v02-b (stonith:fence_ipmilan): Stopped (disabled) stonith-v02-c (stonith:fence_ipmilan): Stopped (disabled) stonith-v02-d (stonith:fence_ipmilan): Stopped (disabled) stonith-mgmt01 (stonith:fence_xvm): Started v03-b stonith-mgmt02 (stonith:meatware): Started mgmt01 stonith-v03-c (stonith:fence_ipmilan): Stopped (disabled) stonith-v03-a (stonith:fence_ipmilan): Started v03-b stonith-v03-b (stonith:fence_ipmilan): Started v03-a stonith-v03-d (stonith:fence_ipmilan): Stopped (disabled) Clone Set: cl-clvmd [clvmd] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-dlm [dlm] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-iscsid [iscsid] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-libvirtd [libvirtd] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-multipathd [multipathd] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-node-params [node-params] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan1-if [vlan1-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan101-if [vlan101-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan102-if [vlan102-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan103-if [vlan103-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan104-if [vlan104-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan3-if [vlan3-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan4-if [vlan4-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan5-if [vlan5-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan900-if [vlan900-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan909-if [vlan909-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-libvirt-images-fs [libvirt-images-fs] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-libvirt-install-fs [libvirt-install-fs] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-0-iscsi [vds-ok-pool-0-iscsi] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-0-vg [vds-ok-pool-0-vg] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-1-iscsi [vds-ok-pool-1-iscsi] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-1-vg [vds-ok-pool-1-vg] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-libvirt-images-pool [libvirt-images-pool] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vds-ok-pool-0-pool [vds-ok-pool-0-pool] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vds-ok-pool-1-pool [vds-ok-pool-1-pool] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] git.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd01-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a vd01-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a vd01-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b vd01-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a vd02-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd02-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd02-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd02-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) f13-x64-devel.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b eu2.ca-pages.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) zakaz.transferrus.ru-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) Clone Set: cl-vlan200-if [vlan200-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] lenny-x32-devel-vm (ocf::vds-ok:VirtualDomain): Started v03-a dist.express-consult.org-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) eu1.ca-pages.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) gotin-bbb-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) maxb-c55-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) metae.ru-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) rodovoepomestie.ru-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) ubuntu9.10-gotin-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) c5-x64-devel.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b Clone Set: cl-mcast-test-net [mcast-test-net] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] dist.fly-uni.org-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) ktstudio.net-vm (ocf::vds-ok:VirtualDomain): Started v03-a cloudsrv.credo-dialogue.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b c6-x64-devel.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a lustre01-right.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b lustre02-right.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a lustre03-left.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b lustre03-right.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a lustre04-left.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b lustre04-right.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a Clone Set: cl-mcast-anbriz-net [mcast-anbriz-net] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] gw.anbriz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b license.anbriz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a terminal.anbriz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) lustre01-left.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a lustre02-left.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b test-01.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a Clone Set: cl-libvirt-qpid [libvirt-qpid] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] gw.gleb.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) gw.gotin.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) terminal0.anbriz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b Clone Set: cl-mcast-gleb-net [mcast-gleb-net] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] diff --git a/cts/scheduler/master-asymmetrical-order.summary b/cts/scheduler/master-asymmetrical-order.summary index 0fcd24375a..3087bd369b 100644 --- a/cts/scheduler/master-asymmetrical-order.summary +++ b/cts/scheduler/master-asymmetrical-order.summary @@ -1,34 +1,34 @@ -2 of 4 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 4 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] Clone Set: ms1 [rsc1] (promotable) Masters: [ node1 ] Slaves: [ node2 ] Clone Set: ms2 [rsc2] (promotable) Masters: [ node2 ] Slaves: [ node1 ] Transition Summary: * Stop rsc1:0 ( Master node1 ) due to node availability * Stop rsc1:1 ( Slave node2 ) due to node availability Executing cluster transition: * Pseudo action: ms1_demote_0 * Resource action: rsc1:0 demote on node1 * Pseudo action: ms1_demoted_0 * Pseudo action: ms1_stop_0 * Resource action: rsc1:0 stop on node1 * Resource action: rsc1:1 stop on node2 * Pseudo action: ms1_stopped_0 Revised cluster status: Online: [ node1 node2 ] Clone Set: ms1 [rsc1] (promotable) Stopped (disabled): [ node1 node2 ] Clone Set: ms2 [rsc2] (promotable) Masters: [ node2 ] Slaves: [ node1 ] diff --git a/cts/scheduler/master-demote-block.summary b/cts/scheduler/master-demote-block.summary index 3873f9f346..246d55e772 100644 --- a/cts/scheduler/master-demote-block.summary +++ b/cts/scheduler/master-demote-block.summary @@ -1,22 +1,23 @@ +0 of 2 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Node dl380g5c (21c624bd-c426-43dc-9665-bbfb92054bcd): standby Online: [ dl380g5d ] Clone Set: stateful [dummy] (promotable) dummy (ocf::pacemaker:Stateful): FAILED Master dl380g5c (blocked) Slaves: [ dl380g5d ] Transition Summary: Executing cluster transition: * Resource action: dummy:1 monitor=20000 on dl380g5d Revised cluster status: Node dl380g5c (21c624bd-c426-43dc-9665-bbfb92054bcd): standby Online: [ dl380g5d ] Clone Set: stateful [dummy] (promotable) dummy (ocf::pacemaker:Stateful): FAILED Master dl380g5c (blocked) Slaves: [ dl380g5d ] diff --git a/cts/scheduler/master-depend.summary b/cts/scheduler/master-depend.summary index 03fb495e1b..5b43e3e355 100644 --- a/cts/scheduler/master-depend.summary +++ b/cts/scheduler/master-depend.summary @@ -1,59 +1,59 @@ -3 of 10 resources DISABLED and 0 BLOCKED from being started due to failures +3 of 10 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ vbox4 ] OFFLINE: [ vbox3 ] Clone Set: drbd [drbd0] (promotable) Stopped: [ vbox3 vbox4 ] Clone Set: cman_clone [cman] Stopped: [ vbox3 vbox4 ] Clone Set: clvmd_clone [clvmd] Stopped: [ vbox3 vbox4 ] vmnci36 (ocf::heartbeat:vm): Stopped vmnci37 (ocf::heartbeat:vm): Stopped (disabled) vmnci38 (ocf::heartbeat:vm): Stopped (disabled) vmnci55 (ocf::heartbeat:vm): Stopped (disabled) Transition Summary: * Start drbd0:0 ( vbox4 ) * Start cman:0 ( vbox4 ) Executing cluster transition: * Resource action: drbd0:0 monitor on vbox4 * Pseudo action: drbd_pre_notify_start_0 * Resource action: cman:0 monitor on vbox4 * Pseudo action: cman_clone_start_0 * Resource action: clvmd:0 monitor on vbox4 * Resource action: vmnci36 monitor on vbox4 * Resource action: vmnci37 monitor on vbox4 * Resource action: vmnci38 monitor on vbox4 * Resource action: vmnci55 monitor on vbox4 * Pseudo action: drbd_confirmed-pre_notify_start_0 * Pseudo action: drbd_start_0 * Resource action: cman:0 start on vbox4 * Pseudo action: cman_clone_running_0 * Resource action: drbd0:0 start on vbox4 * Pseudo action: drbd_running_0 * Pseudo action: drbd_post_notify_running_0 * Resource action: drbd0:0 notify on vbox4 * Pseudo action: drbd_confirmed-post_notify_running_0 * Resource action: drbd0:0 monitor=60000 on vbox4 Revised cluster status: Online: [ vbox4 ] OFFLINE: [ vbox3 ] Clone Set: drbd [drbd0] (promotable) Slaves: [ vbox4 ] Stopped: [ vbox3 ] Clone Set: cman_clone [cman] Started: [ vbox4 ] Stopped: [ vbox3 ] Clone Set: clvmd_clone [clvmd] Stopped: [ vbox3 vbox4 ] vmnci36 (ocf::heartbeat:vm): Stopped vmnci37 (ocf::heartbeat:vm): Stopped (disabled) vmnci38 (ocf::heartbeat:vm): Stopped (disabled) vmnci55 (ocf::heartbeat:vm): Stopped (disabled) diff --git a/cts/scheduler/master-probed-score.summary b/cts/scheduler/master-probed-score.summary index d91dd71c79..aa1e3a131e 100644 --- a/cts/scheduler/master-probed-score.summary +++ b/cts/scheduler/master-probed-score.summary @@ -1,326 +1,326 @@ -2 of 60 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 60 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] Clone Set: AdminClone [AdminDrbd] (promotable) Stopped: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] CronAmbientTemperature (ocf::heartbeat:symlink): Stopped StonithHypatia (stonith:fence_nut): Stopped StonithOrestes (stonith:fence_nut): Stopped Resource Group: DhcpGroup SymlinkDhcpdConf (ocf::heartbeat:symlink): Stopped SymlinkSysconfigDhcpd (ocf::heartbeat:symlink): Stopped SymlinkDhcpdLeases (ocf::heartbeat:symlink): Stopped Dhcpd (lsb:dhcpd): Stopped (disabled) DhcpIP (ocf::heartbeat:IPaddr2): Stopped Clone Set: CupsClone [CupsGroup] Stopped: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] Clone Set: IPClone [IPGroup] (unique) Resource Group: IPGroup:0 ClusterIP:0 (ocf::heartbeat:IPaddr2): Stopped ClusterIPLocal:0 (ocf::heartbeat:IPaddr2): Stopped ClusterIPSandbox:0 (ocf::heartbeat:IPaddr2): Stopped Resource Group: IPGroup:1 ClusterIP:1 (ocf::heartbeat:IPaddr2): Stopped ClusterIPLocal:1 (ocf::heartbeat:IPaddr2): Stopped ClusterIPSandbox:1 (ocf::heartbeat:IPaddr2): Stopped Clone Set: LibvirtdClone [LibvirtdGroup] Stopped: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] Clone Set: TftpClone [TftpGroup] Stopped: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] Clone Set: ExportsClone [ExportsGroup] Stopped: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] Clone Set: FilesystemClone [FilesystemGroup] Stopped: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] KVM-guest (ocf::heartbeat:VirtualDomain): Stopped Proxy (ocf::heartbeat:VirtualDomain): Stopped Transition Summary: * Promote AdminDrbd:0 ( Stopped -> Master hypatia-corosync.nevis.columbia.edu ) * Promote AdminDrbd:1 ( Stopped -> Master orestes-corosync.nevis.columbia.edu ) * Start CronAmbientTemperature ( hypatia-corosync.nevis.columbia.edu ) * Start StonithHypatia ( orestes-corosync.nevis.columbia.edu ) * Start StonithOrestes ( hypatia-corosync.nevis.columbia.edu ) * Start SymlinkDhcpdConf ( orestes-corosync.nevis.columbia.edu ) * Start SymlinkSysconfigDhcpd ( orestes-corosync.nevis.columbia.edu ) * Start SymlinkDhcpdLeases ( orestes-corosync.nevis.columbia.edu ) * Start SymlinkUsrShareCups:0 ( hypatia-corosync.nevis.columbia.edu ) * Start SymlinkCupsdConf:0 ( hypatia-corosync.nevis.columbia.edu ) * Start Cups:0 ( hypatia-corosync.nevis.columbia.edu ) * Start SymlinkUsrShareCups:1 ( orestes-corosync.nevis.columbia.edu ) * Start SymlinkCupsdConf:1 ( orestes-corosync.nevis.columbia.edu ) * Start Cups:1 ( orestes-corosync.nevis.columbia.edu ) * Start ClusterIP:0 ( hypatia-corosync.nevis.columbia.edu ) * Start ClusterIPLocal:0 ( hypatia-corosync.nevis.columbia.edu ) * Start ClusterIPSandbox:0 ( hypatia-corosync.nevis.columbia.edu ) * Start ClusterIP:1 ( orestes-corosync.nevis.columbia.edu ) * Start ClusterIPLocal:1 ( orestes-corosync.nevis.columbia.edu ) * Start ClusterIPSandbox:1 ( orestes-corosync.nevis.columbia.edu ) * Start SymlinkEtcLibvirt:0 ( hypatia-corosync.nevis.columbia.edu ) * Start Libvirtd:0 ( hypatia-corosync.nevis.columbia.edu ) * Start SymlinkEtcLibvirt:1 ( orestes-corosync.nevis.columbia.edu ) * Start Libvirtd:1 ( orestes-corosync.nevis.columbia.edu ) * Start SymlinkTftp:0 ( hypatia-corosync.nevis.columbia.edu ) * Start Xinetd:0 ( hypatia-corosync.nevis.columbia.edu ) * Start SymlinkTftp:1 ( orestes-corosync.nevis.columbia.edu ) * Start Xinetd:1 ( orestes-corosync.nevis.columbia.edu ) * Start ExportMail:0 ( hypatia-corosync.nevis.columbia.edu ) * Start ExportMailInbox:0 ( hypatia-corosync.nevis.columbia.edu ) * Start ExportMailFolders:0 ( hypatia-corosync.nevis.columbia.edu ) * Start ExportMailForward:0 ( hypatia-corosync.nevis.columbia.edu ) * Start ExportMailProcmailrc:0 ( hypatia-corosync.nevis.columbia.edu ) * Start ExportUsrNevis:0 ( hypatia-corosync.nevis.columbia.edu ) * Start ExportUsrNevisOffsite:0 ( hypatia-corosync.nevis.columbia.edu ) * Start ExportWWW:0 ( hypatia-corosync.nevis.columbia.edu ) * Start ExportMail:1 ( orestes-corosync.nevis.columbia.edu ) * Start ExportMailInbox:1 ( orestes-corosync.nevis.columbia.edu ) * Start ExportMailFolders:1 ( orestes-corosync.nevis.columbia.edu ) * Start ExportMailForward:1 ( orestes-corosync.nevis.columbia.edu ) * Start ExportMailProcmailrc:1 ( orestes-corosync.nevis.columbia.edu ) * Start ExportUsrNevis:1 ( orestes-corosync.nevis.columbia.edu ) * Start ExportUsrNevisOffsite:1 ( orestes-corosync.nevis.columbia.edu ) * Start ExportWWW:1 ( orestes-corosync.nevis.columbia.edu ) * Start AdminLvm:0 ( hypatia-corosync.nevis.columbia.edu ) * Start FSUsrNevis:0 ( hypatia-corosync.nevis.columbia.edu ) * Start FSVarNevis:0 ( hypatia-corosync.nevis.columbia.edu ) * Start FSVirtualMachines:0 ( hypatia-corosync.nevis.columbia.edu ) * Start FSMail:0 ( hypatia-corosync.nevis.columbia.edu ) * Start FSWork:0 ( hypatia-corosync.nevis.columbia.edu ) * Start AdminLvm:1 ( orestes-corosync.nevis.columbia.edu ) * Start FSUsrNevis:1 ( orestes-corosync.nevis.columbia.edu ) * Start FSVarNevis:1 ( orestes-corosync.nevis.columbia.edu ) * Start FSVirtualMachines:1 ( orestes-corosync.nevis.columbia.edu ) * Start FSMail:1 ( orestes-corosync.nevis.columbia.edu ) * Start FSWork:1 ( orestes-corosync.nevis.columbia.edu ) * Start KVM-guest ( hypatia-corosync.nevis.columbia.edu ) * Start Proxy ( orestes-corosync.nevis.columbia.edu ) Executing cluster transition: * Pseudo action: AdminClone_pre_notify_start_0 * Resource action: StonithHypatia start on orestes-corosync.nevis.columbia.edu * Resource action: StonithOrestes start on hypatia-corosync.nevis.columbia.edu * Resource action: SymlinkEtcLibvirt:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: Libvirtd:0 monitor on orestes-corosync.nevis.columbia.edu * Resource action: Libvirtd:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: SymlinkTftp:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: Xinetd:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: SymlinkTftp:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: Xinetd:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: ExportMail:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: ExportMailInbox:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: ExportMailFolders:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: ExportMailForward:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: ExportMailProcmailrc:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: ExportUsrNevis:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: ExportUsrNevisOffsite:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: ExportWWW:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: ExportMail:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: ExportMailInbox:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: ExportMailFolders:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: ExportMailForward:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: ExportMailProcmailrc:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: ExportUsrNevis:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: ExportUsrNevisOffsite:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: ExportWWW:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: AdminLvm:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: FSUsrNevis:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: FSVarNevis:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: FSVirtualMachines:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: FSMail:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: FSWork:0 monitor on hypatia-corosync.nevis.columbia.edu * Resource action: AdminLvm:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: FSUsrNevis:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: FSVarNevis:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: FSVirtualMachines:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: FSMail:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: FSWork:1 monitor on orestes-corosync.nevis.columbia.edu * Resource action: KVM-guest monitor on orestes-corosync.nevis.columbia.edu * Resource action: KVM-guest monitor on hypatia-corosync.nevis.columbia.edu * Resource action: Proxy monitor on orestes-corosync.nevis.columbia.edu * Resource action: Proxy monitor on hypatia-corosync.nevis.columbia.edu * Pseudo action: AdminClone_confirmed-pre_notify_start_0 * Pseudo action: AdminClone_start_0 * Resource action: AdminDrbd:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: AdminDrbd:1 start on orestes-corosync.nevis.columbia.edu * Pseudo action: AdminClone_running_0 * Pseudo action: AdminClone_post_notify_running_0 * Resource action: AdminDrbd:0 notify on hypatia-corosync.nevis.columbia.edu * Resource action: AdminDrbd:1 notify on orestes-corosync.nevis.columbia.edu * Pseudo action: AdminClone_confirmed-post_notify_running_0 * Pseudo action: AdminClone_pre_notify_promote_0 * Resource action: AdminDrbd:0 notify on hypatia-corosync.nevis.columbia.edu * Resource action: AdminDrbd:1 notify on orestes-corosync.nevis.columbia.edu * Pseudo action: AdminClone_confirmed-pre_notify_promote_0 * Pseudo action: AdminClone_promote_0 * Resource action: AdminDrbd:0 promote on hypatia-corosync.nevis.columbia.edu * Resource action: AdminDrbd:1 promote on orestes-corosync.nevis.columbia.edu * Pseudo action: AdminClone_promoted_0 * Pseudo action: AdminClone_post_notify_promoted_0 * Resource action: AdminDrbd:0 notify on hypatia-corosync.nevis.columbia.edu * Resource action: AdminDrbd:1 notify on orestes-corosync.nevis.columbia.edu * Pseudo action: AdminClone_confirmed-post_notify_promoted_0 * Pseudo action: FilesystemClone_start_0 * Resource action: AdminDrbd:0 monitor=59000 on hypatia-corosync.nevis.columbia.edu * Resource action: AdminDrbd:1 monitor=59000 on orestes-corosync.nevis.columbia.edu * Pseudo action: FilesystemGroup:0_start_0 * Resource action: AdminLvm:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: FSUsrNevis:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: FSVarNevis:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: FSVirtualMachines:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: FSMail:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: FSWork:0 start on hypatia-corosync.nevis.columbia.edu * Pseudo action: FilesystemGroup:1_start_0 * Resource action: AdminLvm:1 start on orestes-corosync.nevis.columbia.edu * Resource action: FSUsrNevis:1 start on orestes-corosync.nevis.columbia.edu * Resource action: FSVarNevis:1 start on orestes-corosync.nevis.columbia.edu * Resource action: FSVirtualMachines:1 start on orestes-corosync.nevis.columbia.edu * Resource action: FSMail:1 start on orestes-corosync.nevis.columbia.edu * Resource action: FSWork:1 start on orestes-corosync.nevis.columbia.edu * Pseudo action: FilesystemGroup:0_running_0 * Resource action: AdminLvm:0 monitor=30000 on hypatia-corosync.nevis.columbia.edu * Resource action: FSUsrNevis:0 monitor=20000 on hypatia-corosync.nevis.columbia.edu * Resource action: FSVarNevis:0 monitor=20000 on hypatia-corosync.nevis.columbia.edu * Resource action: FSVirtualMachines:0 monitor=20000 on hypatia-corosync.nevis.columbia.edu * Resource action: FSMail:0 monitor=20000 on hypatia-corosync.nevis.columbia.edu * Resource action: FSWork:0 monitor=20000 on hypatia-corosync.nevis.columbia.edu * Pseudo action: FilesystemGroup:1_running_0 * Resource action: AdminLvm:1 monitor=30000 on orestes-corosync.nevis.columbia.edu * Resource action: FSUsrNevis:1 monitor=20000 on orestes-corosync.nevis.columbia.edu * Resource action: FSVarNevis:1 monitor=20000 on orestes-corosync.nevis.columbia.edu * Resource action: FSVirtualMachines:1 monitor=20000 on orestes-corosync.nevis.columbia.edu * Resource action: FSMail:1 monitor=20000 on orestes-corosync.nevis.columbia.edu * Resource action: FSWork:1 monitor=20000 on orestes-corosync.nevis.columbia.edu * Pseudo action: FilesystemClone_running_0 * Resource action: CronAmbientTemperature start on hypatia-corosync.nevis.columbia.edu * Pseudo action: DhcpGroup_start_0 * Resource action: SymlinkDhcpdConf start on orestes-corosync.nevis.columbia.edu * Resource action: SymlinkSysconfigDhcpd start on orestes-corosync.nevis.columbia.edu * Resource action: SymlinkDhcpdLeases start on orestes-corosync.nevis.columbia.edu * Pseudo action: CupsClone_start_0 * Pseudo action: IPClone_start_0 * Pseudo action: LibvirtdClone_start_0 * Pseudo action: TftpClone_start_0 * Pseudo action: ExportsClone_start_0 * Resource action: CronAmbientTemperature monitor=60000 on hypatia-corosync.nevis.columbia.edu * Resource action: SymlinkDhcpdConf monitor=60000 on orestes-corosync.nevis.columbia.edu * Resource action: SymlinkSysconfigDhcpd monitor=60000 on orestes-corosync.nevis.columbia.edu * Resource action: SymlinkDhcpdLeases monitor=60000 on orestes-corosync.nevis.columbia.edu * Pseudo action: CupsGroup:0_start_0 * Resource action: SymlinkUsrShareCups:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: SymlinkCupsdConf:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: Cups:0 start on hypatia-corosync.nevis.columbia.edu * Pseudo action: CupsGroup:1_start_0 * Resource action: SymlinkUsrShareCups:1 start on orestes-corosync.nevis.columbia.edu * Resource action: SymlinkCupsdConf:1 start on orestes-corosync.nevis.columbia.edu * Resource action: Cups:1 start on orestes-corosync.nevis.columbia.edu * Pseudo action: IPGroup:0_start_0 * Resource action: ClusterIP:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: ClusterIPLocal:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: ClusterIPSandbox:0 start on hypatia-corosync.nevis.columbia.edu * Pseudo action: IPGroup:1_start_0 * Resource action: ClusterIP:1 start on orestes-corosync.nevis.columbia.edu * Resource action: ClusterIPLocal:1 start on orestes-corosync.nevis.columbia.edu * Resource action: ClusterIPSandbox:1 start on orestes-corosync.nevis.columbia.edu * Pseudo action: LibvirtdGroup:0_start_0 * Resource action: SymlinkEtcLibvirt:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: Libvirtd:0 start on hypatia-corosync.nevis.columbia.edu * Pseudo action: LibvirtdGroup:1_start_0 * Resource action: SymlinkEtcLibvirt:1 start on orestes-corosync.nevis.columbia.edu * Resource action: Libvirtd:1 start on orestes-corosync.nevis.columbia.edu * Pseudo action: TftpGroup:0_start_0 * Resource action: SymlinkTftp:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: Xinetd:0 start on hypatia-corosync.nevis.columbia.edu * Pseudo action: TftpGroup:1_start_0 * Resource action: SymlinkTftp:1 start on orestes-corosync.nevis.columbia.edu * Resource action: Xinetd:1 start on orestes-corosync.nevis.columbia.edu * Pseudo action: ExportsGroup:0_start_0 * Resource action: ExportMail:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: ExportMailInbox:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: ExportMailFolders:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: ExportMailForward:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: ExportMailProcmailrc:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: ExportUsrNevis:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: ExportUsrNevisOffsite:0 start on hypatia-corosync.nevis.columbia.edu * Resource action: ExportWWW:0 start on hypatia-corosync.nevis.columbia.edu * Pseudo action: ExportsGroup:1_start_0 * Resource action: ExportMail:1 start on orestes-corosync.nevis.columbia.edu * Resource action: ExportMailInbox:1 start on orestes-corosync.nevis.columbia.edu * Resource action: ExportMailFolders:1 start on orestes-corosync.nevis.columbia.edu * Resource action: ExportMailForward:1 start on orestes-corosync.nevis.columbia.edu * Resource action: ExportMailProcmailrc:1 start on orestes-corosync.nevis.columbia.edu * Resource action: ExportUsrNevis:1 start on orestes-corosync.nevis.columbia.edu * Resource action: ExportUsrNevisOffsite:1 start on orestes-corosync.nevis.columbia.edu * Resource action: ExportWWW:1 start on orestes-corosync.nevis.columbia.edu * Pseudo action: CupsGroup:0_running_0 * Resource action: SymlinkUsrShareCups:0 monitor=60000 on hypatia-corosync.nevis.columbia.edu * Resource action: SymlinkCupsdConf:0 monitor=60000 on hypatia-corosync.nevis.columbia.edu * Resource action: Cups:0 monitor=30000 on hypatia-corosync.nevis.columbia.edu * Pseudo action: CupsGroup:1_running_0 * Resource action: SymlinkUsrShareCups:1 monitor=60000 on orestes-corosync.nevis.columbia.edu * Resource action: SymlinkCupsdConf:1 monitor=60000 on orestes-corosync.nevis.columbia.edu * Resource action: Cups:1 monitor=30000 on orestes-corosync.nevis.columbia.edu * Pseudo action: CupsClone_running_0 * Pseudo action: IPGroup:0_running_0 * Resource action: ClusterIP:0 monitor=30000 on hypatia-corosync.nevis.columbia.edu * Resource action: ClusterIPLocal:0 monitor=31000 on hypatia-corosync.nevis.columbia.edu * Resource action: ClusterIPSandbox:0 monitor=32000 on hypatia-corosync.nevis.columbia.edu * Pseudo action: IPGroup:1_running_0 * Resource action: ClusterIP:1 monitor=30000 on orestes-corosync.nevis.columbia.edu * Resource action: ClusterIPLocal:1 monitor=31000 on orestes-corosync.nevis.columbia.edu * Resource action: ClusterIPSandbox:1 monitor=32000 on orestes-corosync.nevis.columbia.edu * Pseudo action: IPClone_running_0 * Pseudo action: LibvirtdGroup:0_running_0 * Resource action: SymlinkEtcLibvirt:0 monitor=60000 on hypatia-corosync.nevis.columbia.edu * Resource action: Libvirtd:0 monitor=30000 on hypatia-corosync.nevis.columbia.edu * Pseudo action: LibvirtdGroup:1_running_0 * Resource action: SymlinkEtcLibvirt:1 monitor=60000 on orestes-corosync.nevis.columbia.edu * Resource action: Libvirtd:1 monitor=30000 on orestes-corosync.nevis.columbia.edu * Pseudo action: LibvirtdClone_running_0 * Pseudo action: TftpGroup:0_running_0 * Resource action: SymlinkTftp:0 monitor=60000 on hypatia-corosync.nevis.columbia.edu * Pseudo action: TftpGroup:1_running_0 * Resource action: SymlinkTftp:1 monitor=60000 on orestes-corosync.nevis.columbia.edu * Pseudo action: TftpClone_running_0 * Pseudo action: ExportsGroup:0_running_0 * Pseudo action: ExportsGroup:1_running_0 * Pseudo action: ExportsClone_running_0 * Resource action: KVM-guest start on hypatia-corosync.nevis.columbia.edu * Resource action: Proxy start on orestes-corosync.nevis.columbia.edu Revised cluster status: Online: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] Clone Set: AdminClone [AdminDrbd] (promotable) Masters: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] CronAmbientTemperature (ocf::heartbeat:symlink): Started hypatia-corosync.nevis.columbia.edu StonithHypatia (stonith:fence_nut): Started orestes-corosync.nevis.columbia.edu StonithOrestes (stonith:fence_nut): Started hypatia-corosync.nevis.columbia.edu Resource Group: DhcpGroup SymlinkDhcpdConf (ocf::heartbeat:symlink): Started orestes-corosync.nevis.columbia.edu SymlinkSysconfigDhcpd (ocf::heartbeat:symlink): Started orestes-corosync.nevis.columbia.edu SymlinkDhcpdLeases (ocf::heartbeat:symlink): Started orestes-corosync.nevis.columbia.edu Dhcpd (lsb:dhcpd): Stopped (disabled) DhcpIP (ocf::heartbeat:IPaddr2): Stopped Clone Set: CupsClone [CupsGroup] Started: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] Clone Set: IPClone [IPGroup] (unique) Resource Group: IPGroup:0 ClusterIP:0 (ocf::heartbeat:IPaddr2): Started hypatia-corosync.nevis.columbia.edu ClusterIPLocal:0 (ocf::heartbeat:IPaddr2): Started hypatia-corosync.nevis.columbia.edu ClusterIPSandbox:0 (ocf::heartbeat:IPaddr2): Started hypatia-corosync.nevis.columbia.edu Resource Group: IPGroup:1 ClusterIP:1 (ocf::heartbeat:IPaddr2): Started orestes-corosync.nevis.columbia.edu ClusterIPLocal:1 (ocf::heartbeat:IPaddr2): Started orestes-corosync.nevis.columbia.edu ClusterIPSandbox:1 (ocf::heartbeat:IPaddr2): Started orestes-corosync.nevis.columbia.edu Clone Set: LibvirtdClone [LibvirtdGroup] Started: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] Clone Set: TftpClone [TftpGroup] Started: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] Clone Set: ExportsClone [ExportsGroup] Started: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] Clone Set: FilesystemClone [FilesystemGroup] Started: [ hypatia-corosync.nevis.columbia.edu orestes-corosync.nevis.columbia.edu ] KVM-guest (ocf::heartbeat:VirtualDomain): Started hypatia-corosync.nevis.columbia.edu Proxy (ocf::heartbeat:VirtualDomain): Started orestes-corosync.nevis.columbia.edu diff --git a/cts/scheduler/master-promotion-constraint.summary b/cts/scheduler/master-promotion-constraint.summary index 49a8b2366f..888fd3a387 100644 --- a/cts/scheduler/master-promotion-constraint.summary +++ b/cts/scheduler/master-promotion-constraint.summary @@ -1,33 +1,33 @@ -4 of 5 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 5 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ hex-13 hex-14 ] fencing-sbd (stonith:external/sbd): Started hex-13 Resource Group: g0 d0 (ocf::pacemaker:Dummy): Stopped (disabled) d1 (ocf::pacemaker:Dummy): Stopped (disabled) Clone Set: ms0 [s0] (promotable) Masters: [ hex-14 ] Slaves: [ hex-13 ] Transition Summary: * Demote s0:0 ( Master -> Slave hex-14 ) Executing cluster transition: * Resource action: s0:1 cancel=20000 on hex-14 * Pseudo action: ms0_demote_0 * Resource action: s0:1 demote on hex-14 * Pseudo action: ms0_demoted_0 * Resource action: s0:1 monitor=21000 on hex-14 Revised cluster status: Online: [ hex-13 hex-14 ] fencing-sbd (stonith:external/sbd): Started hex-13 Resource Group: g0 d0 (ocf::pacemaker:Dummy): Stopped (disabled) d1 (ocf::pacemaker:Dummy): Stopped (disabled) Clone Set: ms0 [s0] (promotable) Slaves: [ hex-13 hex-14 ] diff --git a/cts/scheduler/multiple-active-block-group.summary b/cts/scheduler/multiple-active-block-group.summary index 0515613bee..12321ce566 100644 --- a/cts/scheduler/multiple-active-block-group.summary +++ b/cts/scheduler/multiple-active-block-group.summary @@ -1,23 +1,24 @@ +0 of 4 resource instances DISABLED and 3 BLOCKED from further action due to failure Current cluster status: Online: [ node2 node3 ] st-sbd (stonith:external/sbd): Started node2 Resource Group: dgroup dummy (ocf::heartbeat:DummyTimeout): FAILED (blocked)[ node2 node3 ] dummy2 (ocf::heartbeat:Dummy): Started node2 (blocked) dummy3 (ocf::heartbeat:Dummy): Started node2 (blocked) Transition Summary: Executing cluster transition: Revised cluster status: Online: [ node2 node3 ] st-sbd (stonith:external/sbd): Started node2 Resource Group: dgroup dummy (ocf::heartbeat:DummyTimeout): FAILED (blocked)[ node2 node3 ] dummy2 (ocf::heartbeat:Dummy): Started node2 (blocked) dummy3 (ocf::heartbeat:Dummy): Started node2 (blocked) diff --git a/cts/scheduler/not-reschedule-unneeded-monitor.summary b/cts/scheduler/not-reschedule-unneeded-monitor.summary index 626791f8aa..770f9d41ba 100644 --- a/cts/scheduler/not-reschedule-unneeded-monitor.summary +++ b/cts/scheduler/not-reschedule-unneeded-monitor.summary @@ -1,36 +1,36 @@ -1 of 11 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 11 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ castor kimball ] sbd (stonith:external/sbd): Started kimball Clone Set: base-clone [dlm] Started: [ castor kimball ] Clone Set: c-vm-fs [vm1] Started: [ castor kimball ] xen-f (ocf::heartbeat:VirtualDomain): Stopped (disabled) sle12-kvm (ocf::heartbeat:VirtualDomain): FAILED castor Clone Set: cl_sgdisk [sgdisk] Started: [ castor kimball ] Transition Summary: * Recover sle12-kvm ( castor -> kimball ) Executing cluster transition: * Resource action: sle12-kvm stop on castor * Resource action: sle12-kvm start on kimball * Resource action: sle12-kvm monitor=10000 on kimball Revised cluster status: Online: [ castor kimball ] sbd (stonith:external/sbd): Started kimball Clone Set: base-clone [dlm] Started: [ castor kimball ] Clone Set: c-vm-fs [vm1] Started: [ castor kimball ] xen-f (ocf::heartbeat:VirtualDomain): Stopped (disabled) sle12-kvm (ocf::heartbeat:VirtualDomain): Started kimball Clone Set: cl_sgdisk [sgdisk] Started: [ castor kimball ] diff --git a/cts/scheduler/novell-251689.summary b/cts/scheduler/novell-251689.summary index 8e0e622297..a4841beb93 100644 --- a/cts/scheduler/novell-251689.summary +++ b/cts/scheduler/novell-251689.summary @@ -1,46 +1,46 @@ -1 of 11 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 11 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] Clone Set: stonithcloneset [stonithclone] Started: [ node1 node2 ] Clone Set: evmsdcloneset [evmsdclone] Started: [ node1 node2 ] Clone Set: evmscloneset [evmsclone] Started: [ node1 node2 ] Clone Set: imagestorecloneset [imagestoreclone] Started: [ node1 node2 ] Clone Set: configstorecloneset [configstoreclone] Started: [ node1 node2 ] sles10 (ocf::heartbeat:Xen): Started node2 (disabled) Transition Summary: * Stop sles10 ( node2 ) due to node availability Executing cluster transition: * Resource action: stonithclone:0 monitor=5000 on node2 * Resource action: stonithclone:1 monitor=5000 on node1 * Resource action: evmsdclone:0 monitor=5000 on node2 * Resource action: evmsdclone:1 monitor=5000 on node1 * Resource action: imagestoreclone:0 monitor=20000 on node2 * Resource action: imagestoreclone:1 monitor=20000 on node1 * Resource action: configstoreclone:0 monitor=20000 on node2 * Resource action: configstoreclone:1 monitor=20000 on node1 * Resource action: sles10 stop on node2 Revised cluster status: Online: [ node1 node2 ] Clone Set: stonithcloneset [stonithclone] Started: [ node1 node2 ] Clone Set: evmsdcloneset [evmsdclone] Started: [ node1 node2 ] Clone Set: evmscloneset [evmsclone] Started: [ node1 node2 ] Clone Set: imagestorecloneset [imagestoreclone] Started: [ node1 node2 ] Clone Set: configstorecloneset [configstoreclone] Started: [ node1 node2 ] sles10 (ocf::heartbeat:Xen): Stopped (disabled) diff --git a/cts/scheduler/one-or-more-1.summary b/cts/scheduler/one-or-more-1.summary index 3db1ba0683..c1c75f0cf5 100644 --- a/cts/scheduler/one-or-more-1.summary +++ b/cts/scheduler/one-or-more-1.summary @@ -1,31 +1,31 @@ -1 of 4 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 4 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Stopped (disabled) B (ocf::pacemaker:Dummy): Stopped C (ocf::pacemaker:Dummy): Stopped D (ocf::pacemaker:Dummy): Stopped Transition Summary: * Start B ( fc16-builder ) due to unrunnable A start (blocked) * Start C ( fc16-builder ) due to unrunnable A start (blocked) * Start D ( fc16-builder ) due to unrunnable one-or-more:require-all-set-1 (blocked) Executing cluster transition: * Resource action: A monitor on fc16-builder * Resource action: B monitor on fc16-builder * Resource action: C monitor on fc16-builder * Resource action: D monitor on fc16-builder Revised cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Stopped (disabled) B (ocf::pacemaker:Dummy): Stopped C (ocf::pacemaker:Dummy): Stopped D (ocf::pacemaker:Dummy): Stopped diff --git a/cts/scheduler/one-or-more-2.summary b/cts/scheduler/one-or-more-2.summary index 0345f21239..591f968cc1 100644 --- a/cts/scheduler/one-or-more-2.summary +++ b/cts/scheduler/one-or-more-2.summary @@ -1,35 +1,35 @@ -1 of 4 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 4 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Stopped B (ocf::pacemaker:Dummy): Stopped (disabled) C (ocf::pacemaker:Dummy): Stopped D (ocf::pacemaker:Dummy): Stopped Transition Summary: * Start A ( fc16-builder ) * Start C ( fc16-builder ) * Start D ( fc16-builder ) Executing cluster transition: * Resource action: A monitor on fc16-builder * Resource action: B monitor on fc16-builder * Resource action: C monitor on fc16-builder * Resource action: D monitor on fc16-builder * Resource action: A start on fc16-builder * Resource action: C start on fc16-builder * Pseudo action: one-or-more:require-all-set-1 * Resource action: D start on fc16-builder Revised cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Started fc16-builder B (ocf::pacemaker:Dummy): Stopped (disabled) C (ocf::pacemaker:Dummy): Started fc16-builder D (ocf::pacemaker:Dummy): Started fc16-builder diff --git a/cts/scheduler/one-or-more-3.summary b/cts/scheduler/one-or-more-3.summary index 6b07ca8c9a..73ea57bc67 100644 --- a/cts/scheduler/one-or-more-3.summary +++ b/cts/scheduler/one-or-more-3.summary @@ -1,31 +1,31 @@ -2 of 4 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 4 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Stopped B (ocf::pacemaker:Dummy): Stopped (disabled) C (ocf::pacemaker:Dummy): Stopped (disabled) D (ocf::pacemaker:Dummy): Stopped Transition Summary: * Start A ( fc16-builder ) * Start D ( fc16-builder ) due to unrunnable one-or-more:require-all-set-1 (blocked) Executing cluster transition: * Resource action: A monitor on fc16-builder * Resource action: B monitor on fc16-builder * Resource action: C monitor on fc16-builder * Resource action: D monitor on fc16-builder * Resource action: A start on fc16-builder Revised cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Started fc16-builder B (ocf::pacemaker:Dummy): Stopped (disabled) C (ocf::pacemaker:Dummy): Stopped (disabled) D (ocf::pacemaker:Dummy): Stopped diff --git a/cts/scheduler/one-or-more-4.summary b/cts/scheduler/one-or-more-4.summary index e60331e0f7..ab1632b4d2 100644 --- a/cts/scheduler/one-or-more-4.summary +++ b/cts/scheduler/one-or-more-4.summary @@ -1,35 +1,35 @@ -1 of 4 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 4 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Stopped B (ocf::pacemaker:Dummy): Stopped C (ocf::pacemaker:Dummy): Stopped D (ocf::pacemaker:Dummy): Stopped (disabled) Transition Summary: * Start A ( fc16-builder ) * Start B ( fc16-builder ) * Start C ( fc16-builder ) Executing cluster transition: * Resource action: A monitor on fc16-builder * Resource action: B monitor on fc16-builder * Resource action: C monitor on fc16-builder * Resource action: D monitor on fc16-builder * Resource action: A start on fc16-builder * Resource action: B start on fc16-builder * Resource action: C start on fc16-builder * Pseudo action: one-or-more:require-all-set-1 Revised cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Started fc16-builder B (ocf::pacemaker:Dummy): Started fc16-builder C (ocf::pacemaker:Dummy): Started fc16-builder D (ocf::pacemaker:Dummy): Stopped (disabled) diff --git a/cts/scheduler/one-or-more-5.summary b/cts/scheduler/one-or-more-5.summary index 7ca1cd733d..83b36768e2 100644 --- a/cts/scheduler/one-or-more-5.summary +++ b/cts/scheduler/one-or-more-5.summary @@ -1,44 +1,44 @@ -2 of 6 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 6 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Stopped B (ocf::pacemaker:Dummy): Stopped C (ocf::pacemaker:Dummy): Stopped (disabled) D (ocf::pacemaker:Dummy): Stopped (disabled) E (ocf::pacemaker:Dummy): Stopped F (ocf::pacemaker:Dummy): Stopped Transition Summary: * Start A ( fc16-builder ) * Start B ( fc16-builder ) * Start E ( fc16-builder ) * Start F ( fc16-builder ) Executing cluster transition: * Resource action: A monitor on fc16-builder * Resource action: B monitor on fc16-builder * Resource action: C monitor on fc16-builder * Resource action: D monitor on fc16-builder * Resource action: E monitor on fc16-builder * Resource action: F monitor on fc16-builder * Resource action: B start on fc16-builder * Pseudo action: one-or-more:require-all-set-1 * Resource action: A start on fc16-builder * Resource action: E start on fc16-builder * Pseudo action: one-or-more:require-all-set-3 * Resource action: F start on fc16-builder Revised cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Started fc16-builder B (ocf::pacemaker:Dummy): Started fc16-builder C (ocf::pacemaker:Dummy): Stopped (disabled) D (ocf::pacemaker:Dummy): Stopped (disabled) E (ocf::pacemaker:Dummy): Started fc16-builder F (ocf::pacemaker:Dummy): Started fc16-builder diff --git a/cts/scheduler/one-or-more-6.summary b/cts/scheduler/one-or-more-6.summary index 07240e5805..970a7b53d5 100644 --- a/cts/scheduler/one-or-more-6.summary +++ b/cts/scheduler/one-or-more-6.summary @@ -1,24 +1,24 @@ -1 of 3 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 3 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Started fc16-builder B (ocf::pacemaker:Dummy): Started fc16-builder (disabled) C (ocf::pacemaker:Dummy): Started fc16-builder Transition Summary: * Stop B ( fc16-builder ) due to node availability Executing cluster transition: * Resource action: B stop on fc16-builder Revised cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Started fc16-builder B (ocf::pacemaker:Dummy): Stopped (disabled) C (ocf::pacemaker:Dummy): Started fc16-builder diff --git a/cts/scheduler/one-or-more-7.summary b/cts/scheduler/one-or-more-7.summary index 9d469c8556..cad1afe0e1 100644 --- a/cts/scheduler/one-or-more-7.summary +++ b/cts/scheduler/one-or-more-7.summary @@ -1,24 +1,24 @@ -1 of 3 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 3 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Started fc16-builder B (ocf::pacemaker:Dummy): Started fc16-builder C (ocf::pacemaker:Dummy): Started fc16-builder (disabled) Transition Summary: * Stop C ( fc16-builder ) due to node availability Executing cluster transition: * Resource action: C stop on fc16-builder Revised cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Started fc16-builder B (ocf::pacemaker:Dummy): Started fc16-builder C (ocf::pacemaker:Dummy): Stopped (disabled) diff --git a/cts/scheduler/order-clone.summary b/cts/scheduler/order-clone.summary index 29da82016a..c2b153d9f7 100644 --- a/cts/scheduler/order-clone.summary +++ b/cts/scheduler/order-clone.summary @@ -1,42 +1,42 @@ -4 of 25 resources DISABLED and 0 BLOCKED from being started due to failures +4 of 25 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ hex-0 hex-7 hex-8 hex-9 ] fencing-sbd (stonith:external/sbd): Stopped Clone Set: o2cb-clone [o2cb] Stopped: [ hex-0 hex-7 hex-8 hex-9 ] Clone Set: vg1-clone [vg1] Stopped: [ hex-0 hex-7 hex-8 hex-9 ] Clone Set: fs2-clone [ocfs2-2] Stopped: [ hex-0 hex-7 hex-8 hex-9 ] Clone Set: fs1-clone [ocfs2-1] Stopped: [ hex-0 hex-7 hex-8 hex-9 ] Clone Set: dlm-clone [dlm] Stopped (disabled): [ hex-0 hex-7 hex-8 hex-9 ] Clone Set: clvm-clone [clvm] Stopped: [ hex-0 hex-7 hex-8 hex-9 ] Transition Summary: * Start fencing-sbd ( hex-0 ) Executing cluster transition: * Resource action: fencing-sbd start on hex-0 Revised cluster status: Online: [ hex-0 hex-7 hex-8 hex-9 ] fencing-sbd (stonith:external/sbd): Started hex-0 Clone Set: o2cb-clone [o2cb] Stopped: [ hex-0 hex-7 hex-8 hex-9 ] Clone Set: vg1-clone [vg1] Stopped: [ hex-0 hex-7 hex-8 hex-9 ] Clone Set: fs2-clone [ocfs2-2] Stopped: [ hex-0 hex-7 hex-8 hex-9 ] Clone Set: fs1-clone [ocfs2-1] Stopped: [ hex-0 hex-7 hex-8 hex-9 ] Clone Set: dlm-clone [dlm] Stopped (disabled): [ hex-0 hex-7 hex-8 hex-9 ] Clone Set: clvm-clone [clvm] Stopped: [ hex-0 hex-7 hex-8 hex-9 ] diff --git a/cts/scheduler/order7.summary b/cts/scheduler/order7.summary index 7e54014a4e..0e8e91e002 100644 --- a/cts/scheduler/order7.summary +++ b/cts/scheduler/order7.summary @@ -1,36 +1,37 @@ +0 of 6 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ node1 ] rsc1 (ocf::heartbeat:apache): Started node1 rsc2 (ocf::heartbeat:apache): Stopped rsc3 (ocf::heartbeat:apache): Stopped rscA (ocf::heartbeat:apache): FAILED node1 (blocked) rscB (ocf::heartbeat:apache): Stopped rscC (ocf::heartbeat:apache): Stopped Transition Summary: * Start rsc2 ( node1 ) * Start rsc3 ( node1 ) * Start rscB ( node1 ) * Start rscC ( node1 ) due to unrunnable rscA start (blocked) Executing cluster transition: * Resource action: rsc2 monitor on node1 * Resource action: rsc3 monitor on node1 * Resource action: rscB monitor on node1 * Resource action: rscC monitor on node1 * Resource action: rsc2 start on node1 * Resource action: rsc3 start on node1 * Resource action: rscB start on node1 Revised cluster status: Online: [ node1 ] rsc1 (ocf::heartbeat:apache): Started node1 rsc2 (ocf::heartbeat:apache): Started node1 rsc3 (ocf::heartbeat:apache): Started node1 rscA (ocf::heartbeat:apache): FAILED node1 (blocked) rscB (ocf::heartbeat:apache): Started node1 rscC (ocf::heartbeat:apache): Stopped diff --git a/cts/scheduler/order_constraint_stops_master.summary b/cts/scheduler/order_constraint_stops_master.summary index 94bfbcce22..83654629c8 100644 --- a/cts/scheduler/order_constraint_stops_master.summary +++ b/cts/scheduler/order_constraint_stops_master.summary @@ -1,41 +1,41 @@ -1 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 2 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder fc16-builder2 ] Clone Set: MASTER_RSC_A [NATIVE_RSC_A] (promotable) Masters: [ fc16-builder ] NATIVE_RSC_B (ocf::pacemaker:Dummy): Started fc16-builder2 (disabled) Transition Summary: * Stop NATIVE_RSC_A:0 ( Master fc16-builder ) due to required NATIVE_RSC_B start * Stop NATIVE_RSC_B ( fc16-builder2 ) due to node availability Executing cluster transition: * Pseudo action: MASTER_RSC_A_pre_notify_demote_0 * Resource action: NATIVE_RSC_A:0 notify on fc16-builder * Pseudo action: MASTER_RSC_A_confirmed-pre_notify_demote_0 * Pseudo action: MASTER_RSC_A_demote_0 * Resource action: NATIVE_RSC_A:0 demote on fc16-builder * Pseudo action: MASTER_RSC_A_demoted_0 * Pseudo action: MASTER_RSC_A_post_notify_demoted_0 * Resource action: NATIVE_RSC_A:0 notify on fc16-builder * Pseudo action: MASTER_RSC_A_confirmed-post_notify_demoted_0 * Pseudo action: MASTER_RSC_A_pre_notify_stop_0 * Resource action: NATIVE_RSC_A:0 notify on fc16-builder * Pseudo action: MASTER_RSC_A_confirmed-pre_notify_stop_0 * Pseudo action: MASTER_RSC_A_stop_0 * Resource action: NATIVE_RSC_A:0 stop on fc16-builder * Resource action: NATIVE_RSC_A:0 delete on fc16-builder2 * Pseudo action: MASTER_RSC_A_stopped_0 * Pseudo action: MASTER_RSC_A_post_notify_stopped_0 * Pseudo action: MASTER_RSC_A_confirmed-post_notify_stopped_0 * Resource action: NATIVE_RSC_B stop on fc16-builder2 Revised cluster status: Online: [ fc16-builder fc16-builder2 ] Clone Set: MASTER_RSC_A [NATIVE_RSC_A] (promotable) Stopped: [ fc16-builder fc16-builder2 ] NATIVE_RSC_B (ocf::pacemaker:Dummy): Stopped (disabled) diff --git a/cts/scheduler/order_constraint_stops_slave.summary b/cts/scheduler/order_constraint_stops_slave.summary index 33eae2d7d0..032261571e 100644 --- a/cts/scheduler/order_constraint_stops_slave.summary +++ b/cts/scheduler/order_constraint_stops_slave.summary @@ -1,33 +1,33 @@ -1 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 2 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] Clone Set: MASTER_RSC_A [NATIVE_RSC_A] (promotable) Slaves: [ fc16-builder ] NATIVE_RSC_B (ocf::pacemaker:Dummy): Started fc16-builder (disabled) Transition Summary: * Stop NATIVE_RSC_A:0 ( Slave fc16-builder ) due to required NATIVE_RSC_B start * Stop NATIVE_RSC_B ( fc16-builder ) due to node availability Executing cluster transition: * Pseudo action: MASTER_RSC_A_pre_notify_stop_0 * Resource action: NATIVE_RSC_A:0 notify on fc16-builder * Pseudo action: MASTER_RSC_A_confirmed-pre_notify_stop_0 * Pseudo action: MASTER_RSC_A_stop_0 * Resource action: NATIVE_RSC_A:0 stop on fc16-builder * Pseudo action: MASTER_RSC_A_stopped_0 * Pseudo action: MASTER_RSC_A_post_notify_stopped_0 * Pseudo action: MASTER_RSC_A_confirmed-post_notify_stopped_0 * Resource action: NATIVE_RSC_B stop on fc16-builder Revised cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] Clone Set: MASTER_RSC_A [NATIVE_RSC_A] (promotable) Stopped: [ fc16-builder fc16-builder2 ] NATIVE_RSC_B (ocf::pacemaker:Dummy): Stopped (disabled) diff --git a/cts/scheduler/ordered-set-basic-startup.summary b/cts/scheduler/ordered-set-basic-startup.summary index 49a1a9a8cf..691ca6e53c 100644 --- a/cts/scheduler/ordered-set-basic-startup.summary +++ b/cts/scheduler/ordered-set-basic-startup.summary @@ -1,39 +1,39 @@ -2 of 6 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 6 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Stopped B (ocf::pacemaker:Dummy): Stopped C (ocf::pacemaker:Dummy): Stopped (disabled) D (ocf::pacemaker:Dummy): Stopped (disabled) E (ocf::pacemaker:Dummy): Stopped F (ocf::pacemaker:Dummy): Stopped Transition Summary: * Start A ( fc16-builder ) due to unrunnable C start (blocked) * Start B ( fc16-builder ) * Start E ( fc16-builder ) due to unrunnable A start (blocked) * Start F ( fc16-builder ) due to unrunnable D start (blocked) Executing cluster transition: * Resource action: A monitor on fc16-builder * Resource action: B monitor on fc16-builder * Resource action: C monitor on fc16-builder * Resource action: D monitor on fc16-builder * Resource action: E monitor on fc16-builder * Resource action: F monitor on fc16-builder * Resource action: B start on fc16-builder Revised cluster status: Online: [ fc16-builder ] OFFLINE: [ fc16-builder2 ] A (ocf::pacemaker:Dummy): Stopped B (ocf::pacemaker:Dummy): Started fc16-builder C (ocf::pacemaker:Dummy): Stopped (disabled) D (ocf::pacemaker:Dummy): Stopped (disabled) E (ocf::pacemaker:Dummy): Stopped F (ocf::pacemaker:Dummy): Stopped diff --git a/cts/scheduler/ordered-set-natural.summary b/cts/scheduler/ordered-set-natural.summary index e051a4a0f8..ba82e97d45 100644 --- a/cts/scheduler/ordered-set-natural.summary +++ b/cts/scheduler/ordered-set-natural.summary @@ -1,52 +1,52 @@ -4 of 15 resources DISABLED and 0 BLOCKED from being started due to failures +3 of 15 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] Resource Group: rgroup dummy1-1 (ocf::heartbeat:Dummy): Stopped dummy1-2 (ocf::heartbeat:Dummy): Stopped dummy1-3 (ocf::heartbeat:Dummy): Stopped (disabled) dummy1-4 (ocf::heartbeat:Dummy): Stopped dummy1-5 (ocf::heartbeat:Dummy): Stopped dummy2-1 (ocf::heartbeat:Dummy): Stopped dummy2-2 (ocf::heartbeat:Dummy): Stopped dummy2-3 (ocf::heartbeat:Dummy): Stopped (disabled) dummy3-1 (ocf::heartbeat:Dummy): Stopped dummy3-2 (ocf::heartbeat:Dummy): Stopped dummy3-3 (ocf::heartbeat:Dummy): Stopped (disabled) dummy3-4 (ocf::heartbeat:Dummy): Stopped dummy3-5 (ocf::heartbeat:Dummy): Stopped dummy2-4 (ocf::heartbeat:Dummy): Stopped dummy2-5 (ocf::heartbeat:Dummy): Stopped Transition Summary: * Start dummy1-1 ( node1 ) due to no quorum (blocked) * Start dummy1-2 ( node1 ) due to no quorum (blocked) * Start dummy2-1 ( node2 ) due to no quorum (blocked) * Start dummy2-2 ( node2 ) due to no quorum (blocked) * Start dummy3-4 ( node1 ) due to no quorum (blocked) * Start dummy3-5 ( node1 ) due to no quorum (blocked) Executing cluster transition: Revised cluster status: Online: [ node1 node2 ] Resource Group: rgroup dummy1-1 (ocf::heartbeat:Dummy): Stopped dummy1-2 (ocf::heartbeat:Dummy): Stopped dummy1-3 (ocf::heartbeat:Dummy): Stopped (disabled) dummy1-4 (ocf::heartbeat:Dummy): Stopped dummy1-5 (ocf::heartbeat:Dummy): Stopped dummy2-1 (ocf::heartbeat:Dummy): Stopped dummy2-2 (ocf::heartbeat:Dummy): Stopped dummy2-3 (ocf::heartbeat:Dummy): Stopped (disabled) dummy3-1 (ocf::heartbeat:Dummy): Stopped dummy3-2 (ocf::heartbeat:Dummy): Stopped dummy3-3 (ocf::heartbeat:Dummy): Stopped (disabled) dummy3-4 (ocf::heartbeat:Dummy): Stopped dummy3-5 (ocf::heartbeat:Dummy): Stopped dummy2-4 (ocf::heartbeat:Dummy): Stopped dummy2-5 (ocf::heartbeat:Dummy): Stopped diff --git a/cts/scheduler/params-6.summary b/cts/scheduler/params-6.summary index 14ca154e86..ab634c1627 100644 --- a/cts/scheduler/params-6.summary +++ b/cts/scheduler/params-6.summary @@ -1,376 +1,376 @@ -90 of 337 resources DISABLED and 0 BLOCKED from being started due to failures +90 of 337 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ mgmt01 v03-a v03-b ] stonith-v02-a (stonith:fence_ipmilan): Stopped (disabled) stonith-v02-b (stonith:fence_ipmilan): Stopped (disabled) stonith-v02-c (stonith:fence_ipmilan): Stopped (disabled) stonith-v02-d (stonith:fence_ipmilan): Stopped (disabled) stonith-mgmt01 (stonith:fence_xvm): Started v03-a stonith-mgmt02 (stonith:meatware): Started v03-a stonith-v03-c (stonith:fence_ipmilan): Stopped (disabled) stonith-v03-a (stonith:fence_ipmilan): Started v03-b stonith-v03-b (stonith:fence_ipmilan): Started mgmt01 stonith-v03-d (stonith:fence_ipmilan): Stopped (disabled) Clone Set: cl-clvmd [clvmd] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-dlm [dlm] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-iscsid [iscsid] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-libvirtd [libvirtd] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-multipathd [multipathd] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-node-params [node-params] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan1-if [vlan1-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan101-if [vlan101-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan102-if [vlan102-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan103-if [vlan103-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan104-if [vlan104-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan3-if [vlan3-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan4-if [vlan4-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan5-if [vlan5-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan900-if [vlan900-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan909-if [vlan909-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-libvirt-images-fs [libvirt-images-fs] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-libvirt-install-fs [libvirt-install-fs] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-0-iscsi [vds-ok-pool-0-iscsi] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-0-vg [vds-ok-pool-0-vg] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-1-iscsi [vds-ok-pool-1-iscsi] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-1-vg [vds-ok-pool-1-vg] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-libvirt-images-pool [libvirt-images-pool] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vds-ok-pool-0-pool [vds-ok-pool-0-pool] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vds-ok-pool-1-pool [vds-ok-pool-1-pool] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] git.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) Clone Set: cl-libvirt-qpid [libvirt-qpid] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] vd01-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b vd01-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b vd01-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a vd01-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b vd02-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd02-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd02-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd02-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) f13-x64-devel.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a eu2.ca-pages.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b zakaz.transferrus.ru-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) Clone Set: cl-vlan200-if [vlan200-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] anbriz-gw-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) anbriz-work-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) lenny-x32-devel-vm (ocf::vds-ok:VirtualDomain): Started v03-a vptest1.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest2.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest3.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest4.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest5.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest6.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest7.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest8.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest9.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest10.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest11.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest12.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest13.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest14.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest15.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest16.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest17.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest18.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest19.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest20.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest21.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest22.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest23.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest24.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest25.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest26.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest27.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest28.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest29.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest30.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest31.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest32.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest33.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest34.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest35.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest36.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest37.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest38.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest39.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest40.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest41.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest42.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest43.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest44.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest45.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest46.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest47.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest48.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest49.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest50.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest51.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest52.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest53.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest54.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest55.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest56.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest57.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest58.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest59.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest60.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) sl6-x64-devel.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b dist.express-consult.org-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) eu1.ca-pages.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) gotin-bbb-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) maxb-c55-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) metae.ru-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) rodovoepomestie.ru-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) ubuntu9.10-gotin-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) c5-x64-devel.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a Clone Set: cl-mcast-test-net [mcast-test-net] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] dist.fly-uni.org-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) Transition Summary: * Reload c5-x64-devel.vds-ok.com-vm ( v03-a ) Executing cluster transition: * Resource action: vd01-b.cdev.ttc.prague.cz.vds-ok.com-vm monitor=10000 on v03-b * Resource action: vd01-d.cdev.ttc.prague.cz.vds-ok.com-vm monitor=10000 on v03-b * Resource action: c5-x64-devel.vds-ok.com-vm reload on v03-a * Resource action: c5-x64-devel.vds-ok.com-vm monitor=10000 on v03-a * Pseudo action: load_stopped_v03-b * Pseudo action: load_stopped_v03-a * Pseudo action: load_stopped_mgmt01 Revised cluster status: Online: [ mgmt01 v03-a v03-b ] stonith-v02-a (stonith:fence_ipmilan): Stopped (disabled) stonith-v02-b (stonith:fence_ipmilan): Stopped (disabled) stonith-v02-c (stonith:fence_ipmilan): Stopped (disabled) stonith-v02-d (stonith:fence_ipmilan): Stopped (disabled) stonith-mgmt01 (stonith:fence_xvm): Started v03-a stonith-mgmt02 (stonith:meatware): Started v03-a stonith-v03-c (stonith:fence_ipmilan): Stopped (disabled) stonith-v03-a (stonith:fence_ipmilan): Started v03-b stonith-v03-b (stonith:fence_ipmilan): Started mgmt01 stonith-v03-d (stonith:fence_ipmilan): Stopped (disabled) Clone Set: cl-clvmd [clvmd] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-dlm [dlm] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-iscsid [iscsid] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-libvirtd [libvirtd] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-multipathd [multipathd] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-node-params [node-params] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan1-if [vlan1-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan101-if [vlan101-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan102-if [vlan102-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan103-if [vlan103-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan104-if [vlan104-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan3-if [vlan3-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan4-if [vlan4-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan5-if [vlan5-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan900-if [vlan900-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vlan909-if [vlan909-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-libvirt-images-fs [libvirt-images-fs] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-libvirt-install-fs [libvirt-install-fs] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-0-iscsi [vds-ok-pool-0-iscsi] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-0-vg [vds-ok-pool-0-vg] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-1-iscsi [vds-ok-pool-1-iscsi] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-vds-ok-pool-1-vg [vds-ok-pool-1-vg] Started: [ mgmt01 v03-a v03-b ] Clone Set: cl-libvirt-images-pool [libvirt-images-pool] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vds-ok-pool-0-pool [vds-ok-pool-0-pool] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] Clone Set: cl-vds-ok-pool-1-pool [vds-ok-pool-1-pool] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] git.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) Clone Set: cl-libvirt-qpid [libvirt-qpid] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] vd01-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b vd01-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b vd01-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a vd01-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b vd02-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd02-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd02-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd02-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd03-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-a.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-b.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-c.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vd04-d.cdev.ttc.prague.cz.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) f13-x64-devel.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a eu2.ca-pages.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b zakaz.transferrus.ru-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) Clone Set: cl-vlan200-if [vlan200-if] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] anbriz-gw-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) anbriz-work-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) lenny-x32-devel-vm (ocf::vds-ok:VirtualDomain): Started v03-a vptest1.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest2.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest3.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest4.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest5.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest6.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest7.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest8.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest9.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest10.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest11.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest12.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest13.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest14.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest15.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest16.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest17.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest18.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest19.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest20.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest21.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest22.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest23.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest24.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest25.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest26.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest27.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest28.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest29.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest30.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest31.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest32.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest33.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest34.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest35.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest36.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest37.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest38.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest39.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest40.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest41.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest42.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest43.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest44.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest45.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest46.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest47.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest48.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest49.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest50.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest51.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest52.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest53.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest54.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest55.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest56.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest57.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest58.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest59.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) vptest60.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) sl6-x64-devel.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-b dist.express-consult.org-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) eu1.ca-pages.com-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) gotin-bbb-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) maxb-c55-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) metae.ru-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) rodovoepomestie.ru-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) ubuntu9.10-gotin-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) c5-x64-devel.vds-ok.com-vm (ocf::vds-ok:VirtualDomain): Started v03-a Clone Set: cl-mcast-test-net [mcast-test-net] Started: [ v03-a v03-b ] Stopped: [ mgmt01 ] dist.fly-uni.org-vm (ocf::vds-ok:VirtualDomain): Stopped (disabled) diff --git a/cts/scheduler/rec-rsc-4.summary b/cts/scheduler/rec-rsc-4.summary index 5e37bc1f6a..289c82b447 100644 --- a/cts/scheduler/rec-rsc-4.summary +++ b/cts/scheduler/rec-rsc-4.summary @@ -1,16 +1,17 @@ +0 of 1 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::heartbeat:apache): FAILED node2 (blocked) Transition Summary: Executing cluster transition: * Resource action: rsc1 monitor on node1 Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::heartbeat:apache): FAILED node2 (blocked) diff --git a/cts/scheduler/rec-rsc-8.summary b/cts/scheduler/rec-rsc-8.summary index 2968557a08..65f2f8dbd1 100644 --- a/cts/scheduler/rec-rsc-8.summary +++ b/cts/scheduler/rec-rsc-8.summary @@ -1,15 +1,16 @@ +0 of 1 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::heartbeat:apache): Started (blocked)[ node1 node2 ] Transition Summary: Executing cluster transition: Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::heartbeat:apache): Started (blocked)[ node1 node2 ] diff --git a/cts/scheduler/remote-disable.summary b/cts/scheduler/remote-disable.summary index e036225bc3..7923180fed 100644 --- a/cts/scheduler/remote-disable.summary +++ b/cts/scheduler/remote-disable.summary @@ -1,32 +1,32 @@ -2 of 6 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 6 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ 18builder 18node1 18node2 ] RemoteOnline: [ remote1 ] shooter (stonith:fence_xvm): Started 18node1 remote1 (ocf::pacemaker:remote): Started 18builder (disabled) FAKE1 (ocf::heartbeat:Dummy): Started 18node2 FAKE2 (ocf::heartbeat:Dummy): Started remote1 FAKE3 (ocf::heartbeat:Dummy): Started 18builder FAKE4 (ocf::heartbeat:Dummy): Started 18node1 Transition Summary: * Stop remote1 ( 18builder ) due to node availability * Stop FAKE2 ( remote1 ) due to node availability Executing cluster transition: * Resource action: FAKE2 stop on remote1 * Resource action: remote1 stop on 18builder Revised cluster status: Online: [ 18builder 18node1 18node2 ] RemoteOFFLINE: [ remote1 ] shooter (stonith:fence_xvm): Started 18node1 remote1 (ocf::pacemaker:remote): Stopped (disabled) FAKE1 (ocf::heartbeat:Dummy): Started 18node2 FAKE2 (ocf::heartbeat:Dummy): Stopped FAKE3 (ocf::heartbeat:Dummy): Started 18builder FAKE4 (ocf::heartbeat:Dummy): Started 18node1 diff --git a/cts/scheduler/remote-probe-disable.summary b/cts/scheduler/remote-probe-disable.summary index e3b77808fc..fe5a27d693 100644 --- a/cts/scheduler/remote-probe-disable.summary +++ b/cts/scheduler/remote-probe-disable.summary @@ -1,34 +1,34 @@ -2 of 6 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 6 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ 18builder 18node1 18node2 ] RemoteOnline: [ remote1 ] shooter (stonith:fence_xvm): Started 18node1 remote1 (ocf::pacemaker:remote): Started 18builder (disabled) FAKE1 (ocf::heartbeat:Dummy): Started 18node2 FAKE2 (ocf::heartbeat:Dummy): Stopped FAKE3 (ocf::heartbeat:Dummy): Started 18builder FAKE4 (ocf::heartbeat:Dummy): Started 18node1 Transition Summary: * Stop remote1 ( 18builder ) due to node availability Executing cluster transition: * Resource action: FAKE1 monitor on remote1 * Resource action: FAKE2 monitor on remote1 * Resource action: FAKE3 monitor on remote1 * Resource action: FAKE4 monitor on remote1 * Resource action: remote1 stop on 18builder Revised cluster status: Online: [ 18builder 18node1 18node2 ] RemoteOFFLINE: [ remote1 ] shooter (stonith:fence_xvm): Started 18node1 remote1 (ocf::pacemaker:remote): Stopped (disabled) FAKE1 (ocf::heartbeat:Dummy): Started 18node2 FAKE2 (ocf::heartbeat:Dummy): Stopped FAKE3 (ocf::heartbeat:Dummy): Started 18builder FAKE4 (ocf::heartbeat:Dummy): Started 18node1 diff --git a/cts/scheduler/rsc-maintenance.summary b/cts/scheduler/rsc-maintenance.summary index 4a43217822..0db255a690 100644 --- a/cts/scheduler/rsc-maintenance.summary +++ b/cts/scheduler/rsc-maintenance.summary @@ -1,28 +1,28 @@ -4 of 4 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 4 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] Resource Group: group1 rsc1 (ocf::pacemaker:Dummy): Started node1 (disabled, unmanaged) rsc2 (ocf::pacemaker:Dummy): Started node1 (disabled, unmanaged) Resource Group: group2 rsc3 (ocf::pacemaker:Dummy): Started node2 rsc4 (ocf::pacemaker:Dummy): Started node2 Transition Summary: Executing cluster transition: * Resource action: rsc1 cancel=10000 on node1 * Resource action: rsc2 cancel=10000 on node1 Revised cluster status: Online: [ node1 node2 ] Resource Group: group1 rsc1 (ocf::pacemaker:Dummy): Started node1 (disabled, unmanaged) rsc2 (ocf::pacemaker:Dummy): Started node1 (disabled, unmanaged) Resource Group: group2 rsc3 (ocf::pacemaker:Dummy): Started node2 rsc4 (ocf::pacemaker:Dummy): Started node2 diff --git a/cts/scheduler/rsc-sets-clone-1.summary b/cts/scheduler/rsc-sets-clone-1.summary index 2732dfabb5..a80e4f1fbc 100644 --- a/cts/scheduler/rsc-sets-clone-1.summary +++ b/cts/scheduler/rsc-sets-clone-1.summary @@ -1,83 +1,83 @@ -5 of 24 resources DISABLED and 0 BLOCKED from being started due to failures +5 of 24 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ sys2 sys3 ] vm1 (ocf::heartbeat:Xen): Started sys2 vm2 (ocf::heartbeat:Xen): Stopped (disabled) vm3 (ocf::heartbeat:Xen): Stopped (disabled) vm4 (ocf::heartbeat:Xen): Stopped (disabled) stonithsys2 (stonith:external/ipmi): Stopped stonithsys3 (stonith:external/ipmi): Started sys2 Clone Set: baseclone [basegrp] Started: [ sys2 ] Stopped: [ sys3 ] Clone Set: fs1 [nfs1] Stopped (disabled): [ sys2 sys3 ] Transition Summary: * Restart stonithsys3 ( sys2 ) due to resource definition change * Start controld:1 ( sys3 ) * Start clvmd:1 ( sys3 ) * Start o2cb:1 ( sys3 ) * Start iscsi1:1 ( sys3 ) * Start iscsi2:1 ( sys3 ) * Start vg1:1 ( sys3 ) * Start vg2:1 ( sys3 ) * Start fs2:1 ( sys3 ) * Start stonithsys2 ( sys3 ) Executing cluster transition: * Resource action: vm1 monitor on sys3 * Resource action: vm2 monitor on sys3 * Resource action: vm3 monitor on sys3 * Resource action: vm4 monitor on sys3 * Resource action: stonithsys3 stop on sys2 * Resource action: stonithsys3 monitor on sys3 * Resource action: stonithsys3 start on sys2 * Resource action: stonithsys3 monitor=15000 on sys2 * Resource action: controld:1 monitor on sys3 * Resource action: clvmd:1 monitor on sys3 * Resource action: o2cb:1 monitor on sys3 * Resource action: iscsi1:1 monitor on sys3 * Resource action: iscsi2:1 monitor on sys3 * Resource action: vg1:1 monitor on sys3 * Resource action: vg2:1 monitor on sys3 * Resource action: fs2:1 monitor on sys3 * Pseudo action: baseclone_start_0 * Resource action: nfs1:0 monitor on sys3 * Resource action: stonithsys2 monitor on sys3 * Pseudo action: load_stopped_sys3 * Pseudo action: load_stopped_sys2 * Pseudo action: basegrp:1_start_0 * Resource action: controld:1 start on sys3 * Resource action: clvmd:1 start on sys3 * Resource action: o2cb:1 start on sys3 * Resource action: iscsi1:1 start on sys3 * Resource action: iscsi2:1 start on sys3 * Resource action: vg1:1 start on sys3 * Resource action: vg2:1 start on sys3 * Resource action: fs2:1 start on sys3 * Resource action: stonithsys2 start on sys3 * Pseudo action: basegrp:1_running_0 * Resource action: controld:1 monitor=10000 on sys3 * Resource action: iscsi1:1 monitor=120000 on sys3 * Resource action: iscsi2:1 monitor=120000 on sys3 * Resource action: fs2:1 monitor=20000 on sys3 * Pseudo action: baseclone_running_0 * Resource action: stonithsys2 monitor=15000 on sys3 Revised cluster status: Online: [ sys2 sys3 ] vm1 (ocf::heartbeat:Xen): Started sys2 vm2 (ocf::heartbeat:Xen): Stopped (disabled) vm3 (ocf::heartbeat:Xen): Stopped (disabled) vm4 (ocf::heartbeat:Xen): Stopped (disabled) stonithsys2 (stonith:external/ipmi): Started sys3 stonithsys3 (stonith:external/ipmi): Started sys2 Clone Set: baseclone [basegrp] Started: [ sys2 sys3 ] Clone Set: fs1 [nfs1] Stopped (disabled): [ sys2 sys3 ] diff --git a/cts/scheduler/stop-failure-no-fencing.summary b/cts/scheduler/stop-failure-no-fencing.summary index b51d0182f4..ae46d999fa 100644 --- a/cts/scheduler/stop-failure-no-fencing.summary +++ b/cts/scheduler/stop-failure-no-fencing.summary @@ -1,29 +1,30 @@ +0 of 9 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Node pcmk-3 (103): UNCLEAN (offline) Node pcmk-4 (104): UNCLEAN (offline) Online: [ pcmk-1 pcmk-2 ] Clone Set: dlm-clone [dlm] Stopped: [ pcmk-1 pcmk-2 pcmk-3 pcmk-4 ] Clone Set: clvm-clone [clvm] clvm (lsb:clvmd): FAILED pcmk-3 (UNCLEAN, blocked) Stopped: [ pcmk-1 pcmk-2 pcmk-3 pcmk-4 ] ClusterIP (ocf::heartbeat:IPaddr2): Stopped Transition Summary: Executing cluster transition: Revised cluster status: Node pcmk-3 (103): UNCLEAN (offline) Node pcmk-4 (104): UNCLEAN (offline) Online: [ pcmk-1 pcmk-2 ] Clone Set: dlm-clone [dlm] Stopped: [ pcmk-1 pcmk-2 pcmk-3 pcmk-4 ] Clone Set: clvm-clone [clvm] clvm (lsb:clvmd): FAILED pcmk-3 (UNCLEAN, blocked) Stopped: [ pcmk-1 pcmk-2 pcmk-3 pcmk-4 ] ClusterIP (ocf::heartbeat:IPaddr2): Stopped diff --git a/cts/scheduler/stop-failure-no-quorum.summary b/cts/scheduler/stop-failure-no-quorum.summary index ac7a788f0c..b492f75a25 100644 --- a/cts/scheduler/stop-failure-no-quorum.summary +++ b/cts/scheduler/stop-failure-no-quorum.summary @@ -1,44 +1,45 @@ +0 of 10 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Node pcmk-2 (102): UNCLEAN (online) Node pcmk-3 (103): UNCLEAN (offline) Node pcmk-4 (104): UNCLEAN (offline) Online: [ pcmk-1 ] Clone Set: dlm-clone [dlm] Stopped: [ pcmk-1 pcmk-2 pcmk-3 pcmk-4 ] Clone Set: clvm-clone [clvm] clvm (lsb:clvmd): FAILED pcmk-2 clvm (lsb:clvmd): FAILED pcmk-3 (UNCLEAN, blocked) Stopped: [ pcmk-1 pcmk-3 pcmk-4 ] ClusterIP (ocf::heartbeat:IPaddr2): Stopped Fencing (stonith:fence_xvm): Stopped Transition Summary: * Fence (reboot) pcmk-2 'clvm:0 failed there' * Start dlm:0 ( pcmk-1 ) due to no quorum (blocked) * Stop clvm:0 ( pcmk-2 ) due to node availability * Start clvm:2 ( pcmk-1 ) due to no quorum (blocked) * Start ClusterIP ( pcmk-1 ) due to no quorum (blocked) * Start Fencing ( pcmk-1 ) due to no quorum (blocked) Executing cluster transition: * Fencing pcmk-2 (reboot) * Pseudo action: clvm-clone_stop_0 * Pseudo action: clvm_stop_0 * Pseudo action: clvm-clone_stopped_0 Revised cluster status: Node pcmk-3 (103): UNCLEAN (offline) Node pcmk-4 (104): UNCLEAN (offline) Online: [ pcmk-1 ] OFFLINE: [ pcmk-2 ] Clone Set: dlm-clone [dlm] Stopped: [ pcmk-1 pcmk-2 pcmk-3 pcmk-4 ] Clone Set: clvm-clone [clvm] clvm (lsb:clvmd): FAILED pcmk-3 (UNCLEAN, blocked) Stopped: [ pcmk-1 pcmk-2 pcmk-3 pcmk-4 ] ClusterIP (ocf::heartbeat:IPaddr2): Stopped Fencing (stonith:fence_xvm): Stopped diff --git a/cts/scheduler/stopped-monitor-03.summary b/cts/scheduler/stopped-monitor-03.summary index c7cc3cd126..e3700707d1 100644 --- a/cts/scheduler/stopped-monitor-03.summary +++ b/cts/scheduler/stopped-monitor-03.summary @@ -1,19 +1,19 @@ -1 of 1 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 1 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): Started node1 (disabled) Transition Summary: * Stop rsc1 ( node1 ) due to node availability Executing cluster transition: * Resource action: rsc1 stop on node1 * Resource action: rsc1 monitor=20000 on node1 Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): Stopped (disabled) diff --git a/cts/scheduler/stopped-monitor-04.summary b/cts/scheduler/stopped-monitor-04.summary index 08a4225e45..ed0ba619f1 100644 --- a/cts/scheduler/stopped-monitor-04.summary +++ b/cts/scheduler/stopped-monitor-04.summary @@ -1,16 +1,16 @@ -1 of 1 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 1 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED node1 (disabled, blocked) Transition Summary: Executing cluster transition: Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED node1 (disabled, blocked) diff --git a/cts/scheduler/stopped-monitor-05.summary b/cts/scheduler/stopped-monitor-05.summary index b340756baa..3860bfb404 100644 --- a/cts/scheduler/stopped-monitor-05.summary +++ b/cts/scheduler/stopped-monitor-05.summary @@ -1,15 +1,16 @@ +0 of 1 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED node1 (blocked) Transition Summary: Executing cluster transition: Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED node1 (blocked) diff --git a/cts/scheduler/stopped-monitor-06.summary b/cts/scheduler/stopped-monitor-06.summary index 0d6afc25ea..e18cb69f2f 100644 --- a/cts/scheduler/stopped-monitor-06.summary +++ b/cts/scheduler/stopped-monitor-06.summary @@ -1,16 +1,16 @@ -1 of 1 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 1 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED (disabled, blocked)[ node1 node2 ] Transition Summary: Executing cluster transition: Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED (disabled, blocked)[ node1 node2 ] diff --git a/cts/scheduler/stopped-monitor-07.summary b/cts/scheduler/stopped-monitor-07.summary index 2d4a102c3b..766d4e9865 100644 --- a/cts/scheduler/stopped-monitor-07.summary +++ b/cts/scheduler/stopped-monitor-07.summary @@ -1,15 +1,16 @@ +0 of 1 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED (blocked)[ node1 node2 ] Transition Summary: Executing cluster transition: Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED (blocked)[ node1 node2 ] diff --git a/cts/scheduler/stopped-monitor-11.summary b/cts/scheduler/stopped-monitor-11.summary index 39cda798ba..33df7988a2 100644 --- a/cts/scheduler/stopped-monitor-11.summary +++ b/cts/scheduler/stopped-monitor-11.summary @@ -1,16 +1,16 @@ -1 of 1 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 1 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): Started node1 (disabled, unmanaged) Transition Summary: Executing cluster transition: Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): Started node1 (disabled, unmanaged) diff --git a/cts/scheduler/stopped-monitor-12.summary b/cts/scheduler/stopped-monitor-12.summary index 1508d77977..8903c463fb 100644 --- a/cts/scheduler/stopped-monitor-12.summary +++ b/cts/scheduler/stopped-monitor-12.summary @@ -1,16 +1,16 @@ -1 of 1 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 1 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED (disabled, unmanaged)[ node1 node2 ] Transition Summary: Executing cluster transition: Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED (disabled, unmanaged)[ node1 node2 ] diff --git a/cts/scheduler/stopped-monitor-20.summary b/cts/scheduler/stopped-monitor-20.summary index 517ca86ce7..4871c36fdf 100644 --- a/cts/scheduler/stopped-monitor-20.summary +++ b/cts/scheduler/stopped-monitor-20.summary @@ -1,20 +1,20 @@ -1 of 1 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 1 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): Stopped (disabled) Transition Summary: Executing cluster transition: * Resource action: rsc1 monitor on node2 * Resource action: rsc1 monitor on node1 * Resource action: rsc1 monitor=20000 on node2 * Resource action: rsc1 monitor=20000 on node1 Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): Stopped (disabled) diff --git a/cts/scheduler/stopped-monitor-21.summary b/cts/scheduler/stopped-monitor-21.summary index ced4678c4f..bb2a193bb2 100644 --- a/cts/scheduler/stopped-monitor-21.summary +++ b/cts/scheduler/stopped-monitor-21.summary @@ -1,19 +1,19 @@ -1 of 1 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 1 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED node1 (disabled) Transition Summary: * Stop rsc1 ( node1 ) due to node availability Executing cluster transition: * Resource action: rsc1 stop on node1 * Resource action: rsc1 monitor=20000 on node1 Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): Stopped (disabled) diff --git a/cts/scheduler/stopped-monitor-22.summary b/cts/scheduler/stopped-monitor-22.summary index a3348e47b7..94a2cf1bb9 100644 --- a/cts/scheduler/stopped-monitor-22.summary +++ b/cts/scheduler/stopped-monitor-22.summary @@ -1,22 +1,22 @@ -1 of 1 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 1 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED (disabled)[ node1 node2 ] Transition Summary: * Stop rsc1 ( node1 ) due to node availability * Stop rsc1 ( node2 ) due to node availability Executing cluster transition: * Resource action: rsc1 stop on node2 * Resource action: rsc1 monitor=20000 on node2 * Resource action: rsc1 stop on node1 * Resource action: rsc1 monitor=20000 on node1 Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): Stopped (disabled) diff --git a/cts/scheduler/stopped-monitor-24.summary b/cts/scheduler/stopped-monitor-24.summary index 4634a6b8c9..469db307c0 100644 --- a/cts/scheduler/stopped-monitor-24.summary +++ b/cts/scheduler/stopped-monitor-24.summary @@ -1,16 +1,16 @@ -1 of 1 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 1 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): Stopped (disabled, unmanaged) Transition Summary: Executing cluster transition: Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): Stopped (disabled, unmanaged) diff --git a/cts/scheduler/stopped-monitor-25.summary b/cts/scheduler/stopped-monitor-25.summary index ee4b9fda1b..17ce035194 100644 --- a/cts/scheduler/stopped-monitor-25.summary +++ b/cts/scheduler/stopped-monitor-25.summary @@ -1,18 +1,18 @@ -1 of 1 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 1 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED (disabled, unmanaged)[ node1 node2 ] Transition Summary: Executing cluster transition: * Resource action: rsc1 monitor=10000 on node1 * Resource action: rsc1 cancel=20000 on node1 Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): FAILED (disabled, unmanaged)[ node1 node2 ] diff --git a/cts/scheduler/stopped-monitor-31.summary b/cts/scheduler/stopped-monitor-31.summary index ebda3d79f1..1d9974c3fd 100644 --- a/cts/scheduler/stopped-monitor-31.summary +++ b/cts/scheduler/stopped-monitor-31.summary @@ -1,18 +1,18 @@ -1 of 1 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 1 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 node3 ] rsc1 (ocf::pacemaker:Dummy): Stopped (disabled) Transition Summary: Executing cluster transition: * Resource action: rsc1 monitor on node3 * Resource action: rsc1 monitor=20000 on node3 Revised cluster status: Online: [ node1 node2 node3 ] rsc1 (ocf::pacemaker:Dummy): Stopped (disabled) diff --git a/cts/scheduler/target-1.summary b/cts/scheduler/target-1.summary index 9fffecd887..c481ef114b 100644 --- a/cts/scheduler/target-1.summary +++ b/cts/scheduler/target-1.summary @@ -1,40 +1,40 @@ -1 of 5 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 5 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ c001n01 c001n02 c001n03 c001n08 ] DcIPaddr (ocf::heartbeat:IPaddr): Started c001n02 rsc_c001n08 (ocf::heartbeat:IPaddr): Started c001n08 (disabled) rsc_c001n02 (ocf::heartbeat:IPaddr): Started c001n02 Clone Set: promoteme [rsc_c001n03] (promotable) Slaves: [ c001n03 ] rsc_c001n01 (ocf::heartbeat:IPaddr): Started c001n01 Transition Summary: * Stop rsc_c001n08 ( c001n08 ) due to node availability Executing cluster transition: * Resource action: DcIPaddr monitor on c001n08 * Resource action: DcIPaddr monitor on c001n03 * Resource action: DcIPaddr monitor on c001n01 * Resource action: rsc_c001n08 stop on c001n08 * Resource action: rsc_c001n08 monitor on c001n03 * Resource action: rsc_c001n08 monitor on c001n02 * Resource action: rsc_c001n08 monitor on c001n01 * Resource action: rsc_c001n02 monitor on c001n08 * Resource action: rsc_c001n02 monitor on c001n03 * Resource action: rsc_c001n02 monitor on c001n01 * Resource action: rsc_c001n01 monitor on c001n08 * Resource action: rsc_c001n01 monitor on c001n03 * Resource action: rsc_c001n01 monitor on c001n02 Revised cluster status: Online: [ c001n01 c001n02 c001n03 c001n08 ] DcIPaddr (ocf::heartbeat:IPaddr): Started c001n02 rsc_c001n08 (ocf::heartbeat:IPaddr): Stopped (disabled) rsc_c001n02 (ocf::heartbeat:IPaddr): Started c001n02 Clone Set: promoteme [rsc_c001n03] (promotable) Slaves: [ c001n03 ] rsc_c001n01 (ocf::heartbeat:IPaddr): Started c001n01 diff --git a/cts/scheduler/target-2.summary b/cts/scheduler/target-2.summary index 13e98ca06a..8d3d29b5aa 100644 --- a/cts/scheduler/target-2.summary +++ b/cts/scheduler/target-2.summary @@ -1,41 +1,41 @@ -1 of 5 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 5 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ c001n01 c001n02 c001n03 c001n08 ] DcIPaddr (ocf::heartbeat:IPaddr): Started c001n02 rsc_c001n08 (ocf::heartbeat:IPaddr): Started c001n08 (disabled) rsc_c001n02 (ocf::heartbeat:IPaddr): Started c001n02 rsc_c001n03 (ocf::heartbeat:IPaddr): Started c001n03 rsc_c001n01 (ocf::heartbeat:IPaddr): Started c001n01 Transition Summary: * Stop rsc_c001n08 ( c001n08 ) due to node availability Executing cluster transition: * Resource action: DcIPaddr monitor on c001n08 * Resource action: DcIPaddr monitor on c001n03 * Resource action: DcIPaddr monitor on c001n01 * Resource action: rsc_c001n08 stop on c001n08 * Resource action: rsc_c001n08 monitor on c001n03 * Resource action: rsc_c001n08 monitor on c001n02 * Resource action: rsc_c001n08 monitor on c001n01 * Resource action: rsc_c001n02 monitor on c001n08 * Resource action: rsc_c001n02 monitor on c001n03 * Resource action: rsc_c001n02 monitor on c001n01 * Resource action: rsc_c001n03 monitor on c001n08 * Resource action: rsc_c001n03 monitor on c001n02 * Resource action: rsc_c001n03 monitor on c001n01 * Resource action: rsc_c001n01 monitor on c001n08 * Resource action: rsc_c001n01 monitor on c001n03 * Resource action: rsc_c001n01 monitor on c001n02 Revised cluster status: Online: [ c001n01 c001n02 c001n03 c001n08 ] DcIPaddr (ocf::heartbeat:IPaddr): Started c001n02 rsc_c001n08 (ocf::heartbeat:IPaddr): Stopped (disabled) rsc_c001n02 (ocf::heartbeat:IPaddr): Started c001n02 rsc_c001n03 (ocf::heartbeat:IPaddr): Started c001n03 rsc_c001n01 (ocf::heartbeat:IPaddr): Started c001n01 diff --git a/cts/scheduler/template-1.summary b/cts/scheduler/template-1.summary index 7c6129daff..ecad222fbb 100644 --- a/cts/scheduler/template-1.summary +++ b/cts/scheduler/template-1.summary @@ -1,27 +1,27 @@ -1 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 2 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): Stopped (disabled) rsc2 (ocf::pacemaker:Dummy): Stopped Transition Summary: * Start rsc2 ( node1 ) Executing cluster transition: * Resource action: rsc1 monitor on node2 * Resource action: rsc1 monitor on node1 * Resource action: rsc2 monitor on node2 * Resource action: rsc2 monitor on node1 * Pseudo action: load_stopped_node2 * Pseudo action: load_stopped_node1 * Resource action: rsc2 start on node1 * Resource action: rsc2 monitor=10000 on node1 Revised cluster status: Online: [ node1 node2 ] rsc1 (ocf::pacemaker:Dummy): Stopped (disabled) rsc2 (ocf::pacemaker:Dummy): Started node1 diff --git a/cts/scheduler/unmanaged-block-restart.summary b/cts/scheduler/unmanaged-block-restart.summary index 101b7fd7b2..0410a8ed29 100644 --- a/cts/scheduler/unmanaged-block-restart.summary +++ b/cts/scheduler/unmanaged-block-restart.summary @@ -1,30 +1,31 @@ +0 of 4 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ yingying.site ] Resource Group: group1 rsc1 (ocf::pacemaker:Dummy): Stopped rsc2 (ocf::pacemaker:Dummy): Started yingying.site rsc3 (ocf::pacemaker:Dummy): Started yingying.site rsc4 (ocf::pacemaker:Dummy): FAILED yingying.site (blocked) Transition Summary: * Start rsc1 ( yingying.site ) * Stop rsc2 ( yingying.site ) due to unrunnable rsc3 stop (blocked) * Stop rsc3 ( yingying.site ) due to required rsc2 stop (blocked) Executing cluster transition: * Pseudo action: group1_stop_0 * Pseudo action: group1_start_0 * Resource action: rsc1 start on yingying.site * Resource action: rsc1 monitor=10000 on yingying.site Revised cluster status: Online: [ yingying.site ] Resource Group: group1 rsc1 (ocf::pacemaker:Dummy): Started yingying.site rsc2 (ocf::pacemaker:Dummy): Started yingying.site rsc3 (ocf::pacemaker:Dummy): Started yingying.site rsc4 (ocf::pacemaker:Dummy): FAILED yingying.site (blocked) diff --git a/cts/scheduler/unmanaged-stop-1.summary b/cts/scheduler/unmanaged-stop-1.summary index 8706e70901..32df41134b 100644 --- a/cts/scheduler/unmanaged-stop-1.summary +++ b/cts/scheduler/unmanaged-stop-1.summary @@ -1,19 +1,19 @@ -1 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 2 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ yingying.site ] rsc1 (ocf::pacemaker:Dummy): Started yingying.site (disabled) rsc2 (ocf::pacemaker:Dummy): FAILED yingying.site (blocked) Transition Summary: * Stop rsc1 ( yingying.site ) due to node availability (blocked) Executing cluster transition: Revised cluster status: Online: [ yingying.site ] rsc1 (ocf::pacemaker:Dummy): Started yingying.site (disabled) rsc2 (ocf::pacemaker:Dummy): FAILED yingying.site (blocked) diff --git a/cts/scheduler/unmanaged-stop-2.summary b/cts/scheduler/unmanaged-stop-2.summary index 8706e70901..32df41134b 100644 --- a/cts/scheduler/unmanaged-stop-2.summary +++ b/cts/scheduler/unmanaged-stop-2.summary @@ -1,19 +1,19 @@ -1 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 2 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ yingying.site ] rsc1 (ocf::pacemaker:Dummy): Started yingying.site (disabled) rsc2 (ocf::pacemaker:Dummy): FAILED yingying.site (blocked) Transition Summary: * Stop rsc1 ( yingying.site ) due to node availability (blocked) Executing cluster transition: Revised cluster status: Online: [ yingying.site ] rsc1 (ocf::pacemaker:Dummy): Started yingying.site (disabled) rsc2 (ocf::pacemaker:Dummy): FAILED yingying.site (blocked) diff --git a/cts/scheduler/unmanaged-stop-3.summary b/cts/scheduler/unmanaged-stop-3.summary index 8539f20908..599da31f81 100644 --- a/cts/scheduler/unmanaged-stop-3.summary +++ b/cts/scheduler/unmanaged-stop-3.summary @@ -1,22 +1,22 @@ -4 of 2 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 2 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ yingying.site ] Resource Group: group1 rsc1 (ocf::pacemaker:Dummy): Started yingying.site (disabled) rsc2 (ocf::pacemaker:Dummy): FAILED yingying.site (disabled, blocked) Transition Summary: * Stop rsc1 ( yingying.site ) due to node availability (blocked) Executing cluster transition: * Pseudo action: group1_stop_0 Revised cluster status: Online: [ yingying.site ] Resource Group: group1 rsc1 (ocf::pacemaker:Dummy): Started yingying.site (disabled) rsc2 (ocf::pacemaker:Dummy): FAILED yingying.site (disabled, blocked) diff --git a/cts/scheduler/unmanaged-stop-4.summary b/cts/scheduler/unmanaged-stop-4.summary index a2c2f14a77..c3e16217e1 100644 --- a/cts/scheduler/unmanaged-stop-4.summary +++ b/cts/scheduler/unmanaged-stop-4.summary @@ -1,24 +1,24 @@ -6 of 3 resources DISABLED and 0 BLOCKED from being started due to failures +3 of 3 resource instances DISABLED and 1 BLOCKED from further action due to failure Current cluster status: Online: [ yingying.site ] Resource Group: group1 rsc1 (ocf::pacemaker:Dummy): Started yingying.site (disabled) rsc2 (ocf::pacemaker:Dummy): FAILED yingying.site (disabled, blocked) rsc3 (ocf::heartbeat:Dummy): Stopped (disabled) Transition Summary: * Stop rsc1 ( yingying.site ) due to node availability (blocked) Executing cluster transition: * Pseudo action: group1_stop_0 Revised cluster status: Online: [ yingying.site ] Resource Group: group1 rsc1 (ocf::pacemaker:Dummy): Started yingying.site (disabled) rsc2 (ocf::pacemaker:Dummy): FAILED yingying.site (disabled, blocked) rsc3 (ocf::heartbeat:Dummy): Stopped (disabled) diff --git a/cts/scheduler/unrunnable-2.summary b/cts/scheduler/unrunnable-2.summary index 06b7596f24..dfd7b99f54 100644 --- a/cts/scheduler/unrunnable-2.summary +++ b/cts/scheduler/unrunnable-2.summary @@ -1,175 +1,175 @@ -6 of 117 resources DISABLED and 0 BLOCKED from being started due to failures +6 of 117 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] ip-192.0.2.12 (ocf::heartbeat:IPaddr2): Started overcloud-controller-0 Clone Set: haproxy-clone [haproxy] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: galera-master [galera] (promotable) Masters: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: memcached-clone [memcached] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: rabbitmq-clone [rabbitmq] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-core-clone [openstack-core] Stopped (disabled): [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: redis-master [redis] (promotable) Masters: [ overcloud-controller-1 ] Slaves: [ overcloud-controller-0 overcloud-controller-2 ] ip-192.0.2.11 (ocf::heartbeat:IPaddr2): Started overcloud-controller-1 Clone Set: mongod-clone [mongod] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-aodh-evaluator-clone [openstack-aodh-evaluator] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-nova-scheduler-clone [openstack-nova-scheduler] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-l3-agent-clone [neutron-l3-agent] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-netns-cleanup-clone [neutron-netns-cleanup] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-ovs-cleanup-clone [neutron-ovs-cleanup] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] openstack-cinder-volume (systemd:openstack-cinder-volume): Stopped Clone Set: openstack-heat-engine-clone [openstack-heat-engine] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-ceilometer-api-clone [openstack-ceilometer-api] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-aodh-listener-clone [openstack-aodh-listener] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-metadata-agent-clone [neutron-metadata-agent] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-aodh-notifier-clone [openstack-aodh-notifier] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-heat-api-clone [openstack-heat-api] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-ceilometer-collector-clone [openstack-ceilometer-collector] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-glance-api-clone [openstack-glance-api] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-cinder-scheduler-clone [openstack-cinder-scheduler] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-nova-api-clone [openstack-nova-api] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-nova-consoleauth-clone [openstack-nova-consoleauth] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-sahara-api-clone [openstack-sahara-api] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-heat-api-cloudwatch-clone [openstack-heat-api-cloudwatch] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-sahara-engine-clone [openstack-sahara-engine] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-glance-registry-clone [openstack-glance-registry] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-ceilometer-notification-clone [openstack-ceilometer-notification] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-cinder-api-clone [openstack-cinder-api] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-dhcp-agent-clone [neutron-dhcp-agent] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-openvswitch-agent-clone [neutron-openvswitch-agent] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-nova-novncproxy-clone [openstack-nova-novncproxy] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: delay-clone [delay] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-server-clone [neutron-server] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-ceilometer-central-clone [openstack-ceilometer-central] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: httpd-clone [httpd] Stopped (disabled): [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-heat-api-cfn-clone [openstack-heat-api-cfn] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-nova-conductor-clone [openstack-nova-conductor] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Transition Summary: * Start openstack-cinder-volume ( overcloud-controller-2 ) due to unrunnable openstack-cinder-scheduler-clone running (blocked) Executing cluster transition: Revised cluster status: Online: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] ip-192.0.2.12 (ocf::heartbeat:IPaddr2): Started overcloud-controller-0 Clone Set: haproxy-clone [haproxy] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: galera-master [galera] (promotable) Masters: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: memcached-clone [memcached] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: rabbitmq-clone [rabbitmq] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-core-clone [openstack-core] Stopped (disabled): [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: redis-master [redis] (promotable) Masters: [ overcloud-controller-1 ] Slaves: [ overcloud-controller-0 overcloud-controller-2 ] ip-192.0.2.11 (ocf::heartbeat:IPaddr2): Started overcloud-controller-1 Clone Set: mongod-clone [mongod] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-aodh-evaluator-clone [openstack-aodh-evaluator] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-nova-scheduler-clone [openstack-nova-scheduler] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-l3-agent-clone [neutron-l3-agent] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-netns-cleanup-clone [neutron-netns-cleanup] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-ovs-cleanup-clone [neutron-ovs-cleanup] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] openstack-cinder-volume (systemd:openstack-cinder-volume): Stopped Clone Set: openstack-heat-engine-clone [openstack-heat-engine] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-ceilometer-api-clone [openstack-ceilometer-api] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-aodh-listener-clone [openstack-aodh-listener] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-metadata-agent-clone [neutron-metadata-agent] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-aodh-notifier-clone [openstack-aodh-notifier] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-heat-api-clone [openstack-heat-api] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-ceilometer-collector-clone [openstack-ceilometer-collector] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-glance-api-clone [openstack-glance-api] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-cinder-scheduler-clone [openstack-cinder-scheduler] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-nova-api-clone [openstack-nova-api] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-nova-consoleauth-clone [openstack-nova-consoleauth] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-sahara-api-clone [openstack-sahara-api] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-heat-api-cloudwatch-clone [openstack-heat-api-cloudwatch] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-sahara-engine-clone [openstack-sahara-engine] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-glance-registry-clone [openstack-glance-registry] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-ceilometer-notification-clone [openstack-ceilometer-notification] Started: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-cinder-api-clone [openstack-cinder-api] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-dhcp-agent-clone [neutron-dhcp-agent] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-openvswitch-agent-clone [neutron-openvswitch-agent] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-nova-novncproxy-clone [openstack-nova-novncproxy] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: delay-clone [delay] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: neutron-server-clone [neutron-server] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-ceilometer-central-clone [openstack-ceilometer-central] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: httpd-clone [httpd] Stopped (disabled): [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-heat-api-cfn-clone [openstack-heat-api-cfn] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] Clone Set: openstack-nova-conductor-clone [openstack-nova-conductor] Stopped: [ overcloud-controller-0 overcloud-controller-1 overcloud-controller-2 ] diff --git a/cts/scheduler/use-after-free-merge.summary b/cts/scheduler/use-after-free-merge.summary index b47e80c4d4..b7196d8ccf 100644 --- a/cts/scheduler/use-after-free-merge.summary +++ b/cts/scheduler/use-after-free-merge.summary @@ -1,42 +1,42 @@ -4 of 5 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 5 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ hex-13 hex-14 ] fencing-sbd (stonith:external/sbd): Stopped Resource Group: g0 d0 (ocf::heartbeat:Dummy): Stopped (disabled) d1 (ocf::heartbeat:Dummy): Stopped (disabled) Clone Set: ms0 [s0] (promotable) Stopped: [ hex-13 hex-14 ] Transition Summary: * Start fencing-sbd ( hex-14 ) * Start s0:0 ( hex-13 ) * Start s0:1 ( hex-14 ) Executing cluster transition: * Resource action: fencing-sbd monitor on hex-14 * Resource action: fencing-sbd monitor on hex-13 * Resource action: d0 monitor on hex-14 * Resource action: d0 monitor on hex-13 * Resource action: d1 monitor on hex-14 * Resource action: d1 monitor on hex-13 * Resource action: s0:0 monitor on hex-13 * Resource action: s0:1 monitor on hex-14 * Pseudo action: ms0_start_0 * Resource action: fencing-sbd start on hex-14 * Resource action: s0:0 start on hex-13 * Resource action: s0:1 start on hex-14 * Pseudo action: ms0_running_0 Revised cluster status: Online: [ hex-13 hex-14 ] fencing-sbd (stonith:external/sbd): Started hex-14 Resource Group: g0 d0 (ocf::heartbeat:Dummy): Stopped (disabled) d1 (ocf::heartbeat:Dummy): Stopped (disabled) Clone Set: ms0 [s0] (promotable) Slaves: [ hex-13 hex-14 ] diff --git a/cts/scheduler/utilization-order4.summary b/cts/scheduler/utilization-order4.summary index c794588cfb..4097318b1e 100644 --- a/cts/scheduler/utilization-order4.summary +++ b/cts/scheduler/utilization-order4.summary @@ -1,60 +1,60 @@ -2 of 13 resources DISABLED and 0 BLOCKED from being started due to failures +2 of 13 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Node deglxen002: standby Online: [ deglxen001 ] degllx62-vm (ocf::heartbeat:Xen): Started deglxen002 degllx63-vm (ocf::heartbeat:Xen): Stopped (disabled) degllx61-vm (ocf::heartbeat:Xen): Started deglxen001 degllx64-vm (ocf::heartbeat:Xen): Stopped (disabled) stonith_sbd (stonith:external/sbd): Started deglxen001 Clone Set: clone-nfs [grp-nfs] Started: [ deglxen001 deglxen002 ] Clone Set: clone-ping [prim-ping] Started: [ deglxen001 deglxen002 ] Transition Summary: * Migrate degllx62-vm ( deglxen002 -> deglxen001 ) * Stop degllx61-vm ( deglxen001 ) due to node availability * Stop nfs-xen_config:1 ( deglxen002 ) due to node availability * Stop nfs-xen_swapfiles:1 ( deglxen002 ) due to node availability * Stop nfs-xen_images:1 ( deglxen002 ) due to node availability * Stop prim-ping:1 ( deglxen002 ) due to node availability Executing cluster transition: * Resource action: degllx61-vm stop on deglxen001 * Pseudo action: load_stopped_deglxen001 * Resource action: degllx62-vm migrate_to on deglxen002 * Resource action: degllx62-vm migrate_from on deglxen001 * Resource action: degllx62-vm stop on deglxen002 * Pseudo action: clone-nfs_stop_0 * Pseudo action: load_stopped_deglxen002 * Pseudo action: degllx62-vm_start_0 * Pseudo action: grp-nfs:1_stop_0 * Resource action: nfs-xen_images:1 stop on deglxen002 * Resource action: degllx62-vm monitor=30000 on deglxen001 * Resource action: nfs-xen_swapfiles:1 stop on deglxen002 * Resource action: nfs-xen_config:1 stop on deglxen002 * Pseudo action: grp-nfs:1_stopped_0 * Pseudo action: clone-nfs_stopped_0 * Pseudo action: clone-ping_stop_0 * Resource action: prim-ping:0 stop on deglxen002 * Pseudo action: clone-ping_stopped_0 Revised cluster status: Node deglxen002: standby Online: [ deglxen001 ] degllx62-vm (ocf::heartbeat:Xen): Started deglxen001 degllx63-vm (ocf::heartbeat:Xen): Stopped (disabled) degllx61-vm (ocf::heartbeat:Xen): Stopped deglxen002 degllx64-vm (ocf::heartbeat:Xen): Stopped (disabled) stonith_sbd (stonith:external/sbd): Started deglxen001 Clone Set: clone-nfs [grp-nfs] Started: [ deglxen001 ] Stopped: [ deglxen002 ] Clone Set: clone-ping [prim-ping] Started: [ deglxen001 ] Stopped: [ deglxen002 ] diff --git a/cts/scheduler/whitebox-asymmetric.summary b/cts/scheduler/whitebox-asymmetric.summary index 872bb13d90..2548ee4770 100644 --- a/cts/scheduler/whitebox-asymmetric.summary +++ b/cts/scheduler/whitebox-asymmetric.summary @@ -1,39 +1,39 @@ -2 of 7 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 7 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ 18builder ] fence_false (stonith:fence_false): Stopped container2 (ocf::pacemaker:Dummy): Started 18builder webserver (ocf::pacemaker:Dummy): Stopped nfs_mount (ocf::pacemaker:Dummy): Stopped Resource Group: mygroup vg_tags (ocf::heartbeat:LVM): Stopped (disabled) vg_tags_dup (ocf::heartbeat:LVM): Stopped Transition Summary: * Start nfs_mount ( 18node2 ) * Start 18node2 ( 18builder ) Executing cluster transition: * Resource action: 18node2 start on 18builder * Resource action: webserver monitor on 18node2 * Resource action: nfs_mount monitor on 18node2 * Resource action: vg_tags monitor on 18node2 * Resource action: vg_tags_dup monitor on 18node2 * Resource action: 18node2 monitor=30000 on 18builder * Resource action: nfs_mount start on 18node2 * Resource action: nfs_mount monitor=10000 on 18node2 Revised cluster status: Online: [ 18builder ] GuestOnline: [ 18node2:container2 ] fence_false (stonith:fence_false): Stopped container2 (ocf::pacemaker:Dummy): Started 18builder webserver (ocf::pacemaker:Dummy): Stopped nfs_mount (ocf::pacemaker:Dummy): Started 18node2 Resource Group: mygroup vg_tags (ocf::heartbeat:LVM): Stopped (disabled) vg_tags_dup (ocf::heartbeat:LVM): Stopped diff --git a/cts/scheduler/whitebox-stop.summary b/cts/scheduler/whitebox-stop.summary index e9238d5830..cb1a0143b7 100644 --- a/cts/scheduler/whitebox-stop.summary +++ b/cts/scheduler/whitebox-stop.summary @@ -1,50 +1,50 @@ -1 of 14 resources DISABLED and 0 BLOCKED from being started due to failures +1 of 14 resource instances DISABLED and 0 BLOCKED from further action due to failure Current cluster status: Online: [ 18node1 18node2 18node3 ] GuestOnline: [ lxc1:container1 lxc2:container2 ] container1 (ocf::heartbeat:VirtualDomain): Started 18node2 (disabled) container2 (ocf::heartbeat:VirtualDomain): Started 18node2 shoot1 (stonith:fence_xvm): Started 18node3 Clone Set: M-clone [M] Started: [ 18node1 18node2 18node3 lxc1 lxc2 ] A (ocf::pacemaker:Dummy): Started 18node1 B (ocf::pacemaker:Dummy): Started lxc1 C (ocf::pacemaker:Dummy): Started lxc2 D (ocf::pacemaker:Dummy): Started 18node1 Transition Summary: * Stop container1 ( 18node2 ) due to node availability * Stop M:4 ( lxc1 ) due to node availability * Move B ( lxc1 -> lxc2 ) * Stop lxc1 ( 18node2 ) due to node availability Executing cluster transition: * Pseudo action: M-clone_stop_0 * Resource action: A monitor on lxc2 * Resource action: B stop on lxc1 * Resource action: B monitor on lxc2 * Resource action: D monitor on lxc2 * Resource action: M stop on lxc1 * Pseudo action: M-clone_stopped_0 * Resource action: B start on lxc2 * Resource action: lxc1 stop on 18node2 * Resource action: container1 stop on 18node2 * Resource action: B monitor=10000 on lxc2 Revised cluster status: Online: [ 18node1 18node2 18node3 ] GuestOnline: [ lxc2:container2 ] container1 (ocf::heartbeat:VirtualDomain): Stopped (disabled) container2 (ocf::heartbeat:VirtualDomain): Started 18node2 shoot1 (stonith:fence_xvm): Started 18node3 Clone Set: M-clone [M] Started: [ 18node1 18node2 18node3 lxc2 ] Stopped: [ lxc1 ] A (ocf::pacemaker:Dummy): Started 18node1 B (ocf::pacemaker:Dummy): Started lxc2 C (ocf::pacemaker:Dummy): Started lxc2 D (ocf::pacemaker:Dummy): Started 18node1 diff --git a/daemons/execd/cts-exec-helper.c b/daemons/execd/cts-exec-helper.c index 90e5622849..eb0d910042 100644 --- a/daemons/execd/cts-exec-helper.c +++ b/daemons/execd/cts-exec-helper.c @@ -1,620 +1,623 @@ /* - * Copyright 2012-2019 David Vossel + * Copyright 2012-2019 the Pacemaker project contributors + * + * The version control history for this file may have further details. * * This source code is licensed under the GNU Lesser General Public License * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include /* *INDENT-OFF* */ static struct crm_option long_options[] = { {"help", 0, 0, '?'}, {"verbose", 0, 0, 'V', "\t\tPrint out logs and events to screen"}, {"quiet", 0, 0, 'Q', "\t\tSuppress all output to screen"}, {"tls", 0, 0, 'S', "\t\tUse tls backend for local connection"}, {"listen", 1, 0, 'l', "\tListen for a specific event string"}, {"api-call", 1, 0, 'c', "\tDirectly relates to executor API functions"}, {"no-wait", 0, 0, 'w', "\tMake api call and do not wait for result."}, {"is-running", 0, 0, 'R', "\tDetermine if a resource is registered and running."}, {"notify-orig", 0, 0, 'n', "\tOnly notify this client the results of an api action."}, {"notify-changes", 0, 0, 'o', "\tOnly notify client changes to recurring operations."}, {"-spacer-", 1, 0, '-', "\nParameters for api-call option"}, {"action", 1, 0, 'a'}, {"rsc-id", 1, 0, 'r'}, {"cancel-call-id", 1, 0, 'x'}, {"provider", 1, 0, 'P'}, {"class", 1, 0, 'C'}, {"type", 1, 0, 'T'}, {"interval", 1, 0, 'i'}, {"timeout", 1, 0, 't'}, {"start-delay", 1, 0, 's'}, {"param-key", 1, 0, 'k'}, {"param-val", 1, 0, 'v'}, {"-spacer-", 1, 0, '-'}, {0, 0, 0, 0} }; /* *INDENT-ON* */ static cib_t *cib_conn = NULL; static int exec_call_id = 0; static int exec_call_opts = 0; static gboolean start_test(gpointer user_data); static void try_connect(void); static struct { int verbose; int quiet; guint interval_ms; int timeout; int start_delay; int cancel_call_id; int no_wait; int is_running; int no_connect; const char *api_call; const char *rsc_id; const char *provider; const char *class; const char *type; const char *action; const char *listen; lrmd_key_value_t *params; } options; static GMainLoop *mainloop = NULL; static lrmd_t *lrmd_conn = NULL; static char event_buf_v0[1024]; static void test_exit(crm_exit_t exit_code) { lrmd_api_delete(lrmd_conn); crm_exit(exit_code); } #define print_result(result) \ if (!options.quiet) { \ result; \ } \ #define report_event(event) \ snprintf(event_buf_v0, sizeof(event_buf_v0), "NEW_EVENT event_type:%s rsc_id:%s action:%s rc:%s op_status:%s", \ lrmd_event_type2str(event->type), \ event->rsc_id, \ event->op_type ? event->op_type : "none", \ services_ocf_exitcode_str(event->rc), \ services_lrm_status_str(event->op_status)); \ crm_info("%s", event_buf_v0); static void test_shutdown(int nsig) { lrmd_api_delete(lrmd_conn); lrmd_conn = NULL; } static void read_events(lrmd_event_data_t * event) { report_event(event); if (options.listen) { if (safe_str_eq(options.listen, event_buf_v0)) { print_result(printf("LISTEN EVENT SUCCESSFUL\n")); test_exit(CRM_EX_OK); } } if (exec_call_id && (event->call_id == exec_call_id)) { if (event->op_status == 0 && event->rc == 0) { print_result(printf("API-CALL SUCCESSFUL for 'exec'\n")); } else { print_result(printf("API-CALL FAILURE for 'exec', rc:%d lrmd_op_status:%s\n", event->rc, services_lrm_status_str(event->op_status))); test_exit(CRM_EX_ERROR); } if (!options.listen) { test_exit(CRM_EX_OK); } } } static gboolean timeout_err(gpointer data) { print_result(printf("LISTEN EVENT FAILURE - timeout occurred, never found.\n")); test_exit(CRM_EX_TIMEOUT); return FALSE; } static void connection_events(lrmd_event_data_t * event) { int rc = event->connection_rc; if (event->type != lrmd_event_connect) { /* ignore */ return; } if (!rc) { crm_info("Executor client connection established"); start_test(NULL); return; } else { sleep(1); try_connect(); crm_notice("Executor client connection failed"); } } static void try_connect(void) { int tries = 10; static int num_tries = 0; int rc = 0; lrmd_conn->cmds->set_callback(lrmd_conn, connection_events); for (; num_tries < tries; num_tries++) { rc = lrmd_conn->cmds->connect_async(lrmd_conn, "pacemaker-execd", 3000); if (!rc) { return; /* we'll hear back in async callback */ } sleep(1); } print_result(printf("API CONNECTION FAILURE\n")); test_exit(CRM_EX_ERROR); } static gboolean start_test(gpointer user_data) { int rc = 0; if (!options.no_connect) { if (!lrmd_conn->cmds->is_connected(lrmd_conn)) { try_connect(); /* async connect -- this function will get called back into */ return 0; } } lrmd_conn->cmds->set_callback(lrmd_conn, read_events); if (options.timeout) { g_timeout_add(options.timeout, timeout_err, NULL); } if (!options.api_call) { return 0; } if (safe_str_eq(options.api_call, "exec")) { rc = lrmd_conn->cmds->exec(lrmd_conn, options.rsc_id, options.action, NULL, options.interval_ms, options.timeout, options.start_delay, exec_call_opts, options.params); if (rc > 0) { exec_call_id = rc; print_result(printf("API-CALL 'exec' action pending, waiting on response\n")); } } else if (safe_str_eq(options.api_call, "register_rsc")) { rc = lrmd_conn->cmds->register_rsc(lrmd_conn, options.rsc_id, options.class, options.provider, options.type, 0); } else if (safe_str_eq(options.api_call, "get_rsc_info")) { lrmd_rsc_info_t *rsc_info; rsc_info = lrmd_conn->cmds->get_rsc_info(lrmd_conn, options.rsc_id, 0); if (rsc_info) { print_result(printf("RSC_INFO: id:%s class:%s provider:%s type:%s\n", rsc_info->id, rsc_info->standard, rsc_info->provider ? rsc_info->provider : "", rsc_info->type)); lrmd_free_rsc_info(rsc_info); rc = pcmk_ok; } else { rc = -1; } } else if (safe_str_eq(options.api_call, "unregister_rsc")) { rc = lrmd_conn->cmds->unregister_rsc(lrmd_conn, options.rsc_id, 0); } else if (safe_str_eq(options.api_call, "cancel")) { rc = lrmd_conn->cmds->cancel(lrmd_conn, options.rsc_id, options.action, options.interval_ms); } else if (safe_str_eq(options.api_call, "metadata")) { char *output = NULL; rc = lrmd_conn->cmds->get_metadata(lrmd_conn, options.class, options.provider, options.type, &output, 0); if (rc == pcmk_ok) { print_result(printf("%s", output)); free(output); } } else if (safe_str_eq(options.api_call, "list_agents")) { lrmd_list_t *list = NULL; lrmd_list_t *iter = NULL; rc = lrmd_conn->cmds->list_agents(lrmd_conn, &list, options.class, options.provider); if (rc > 0) { print_result(printf("%d agents found\n", rc)); for (iter = list; iter != NULL; iter = iter->next) { print_result(printf("%s\n", iter->val)); } lrmd_list_freeall(list); rc = 0; } else { print_result(printf("API_CALL FAILURE - no agents found\n")); rc = -1; } } else if (safe_str_eq(options.api_call, "list_ocf_providers")) { lrmd_list_t *list = NULL; lrmd_list_t *iter = NULL; rc = lrmd_conn->cmds->list_ocf_providers(lrmd_conn, options.type, &list); if (rc > 0) { print_result(printf("%d providers found\n", rc)); for (iter = list; iter != NULL; iter = iter->next) { print_result(printf("%s\n", iter->val)); } lrmd_list_freeall(list); rc = 0; } else { print_result(printf("API_CALL FAILURE - no providers found\n")); rc = -1; } } else if (safe_str_eq(options.api_call, "list_standards")) { lrmd_list_t *list = NULL; lrmd_list_t *iter = NULL; rc = lrmd_conn->cmds->list_standards(lrmd_conn, &list); if (rc > 0) { print_result(printf("%d standards found\n", rc)); for (iter = list; iter != NULL; iter = iter->next) { print_result(printf("%s\n", iter->val)); } lrmd_list_freeall(list); rc = 0; } else { print_result(printf("API_CALL FAILURE - no providers found\n")); rc = -1; } } else if (safe_str_eq(options.api_call, "get_recurring_ops")) { GList *op_list = NULL; GList *op_item = NULL; rc = lrmd_conn->cmds->get_recurring_ops(lrmd_conn, options.rsc_id, 0, 0, &op_list); for (op_item = op_list; op_item != NULL; op_item = op_item->next) { lrmd_op_info_t *op_info = op_item->data; print_result(printf("RECURRING_OP: %s_%s_%s timeout=%sms\n", op_info->rsc_id, op_info->action, op_info->interval_ms_s, op_info->timeout_ms_s)); lrmd_free_op_info(op_info); } g_list_free(op_list); } else if (options.api_call) { print_result(printf("API-CALL FAILURE unknown action '%s'\n", options.action)); test_exit(CRM_EX_ERROR); } if (rc < 0) { print_result(printf("API-CALL FAILURE for '%s' api_rc:%d\n", options.api_call, rc)); test_exit(CRM_EX_ERROR); } if (options.api_call && rc == pcmk_ok) { print_result(printf("API-CALL SUCCESSFUL for '%s'\n", options.api_call)); if (!options.listen) { test_exit(CRM_EX_OK); } } if (options.no_wait) { /* just make the call and exit regardless of anything else. */ test_exit(CRM_EX_OK); } return 0; } static int generate_params(void) { int rc = 0; pe_working_set_t *data_set = NULL; xmlNode *cib_xml_copy = NULL; resource_t *rsc = NULL; GHashTable *params = NULL; GHashTable *meta = NULL; GHashTableIter iter; if (options.params) { return 0; } data_set = pe_new_working_set(); if (data_set == NULL) { crm_crit("Could not allocate working set"); return -ENOMEM; } + set_bit(data_set->flags, pe_flag_no_counts); cib_conn = cib_new(); rc = cib_conn->cmds->signon(cib_conn, "cts-exec-helper", cib_query); if (rc != pcmk_ok) { crm_err("Could not connect to the CIB manager: %s", pcmk_strerror(rc)); rc = -1; goto param_gen_bail; } rc = cib_conn->cmds->query(cib_conn, NULL, &cib_xml_copy, cib_scope_local | cib_sync_call); if (rc != pcmk_ok) { crm_err("Error retrieving cib copy: %s (%d)", pcmk_strerror(rc), rc); goto param_gen_bail; } else if (cib_xml_copy == NULL) { rc = -ENODATA; crm_err("Error retrieving cib copy: %s (%d)", pcmk_strerror(rc), rc); goto param_gen_bail; } if (cli_config_update(&cib_xml_copy, NULL, FALSE) == FALSE) { crm_err("Error updating cib configuration"); rc = -1; goto param_gen_bail; } data_set->input = cib_xml_copy; data_set->now = crm_time_new(NULL); cluster_status(data_set); if (options.rsc_id) { rsc = pe_find_resource_with_flags(data_set->resources, options.rsc_id, pe_find_renamed|pe_find_any); } if (!rsc) { crm_err("Resource does not exist in config"); rc = -1; goto param_gen_bail; } params = crm_str_table_new(); meta = crm_str_table_new(); get_rsc_attributes(params, rsc, NULL, data_set); get_meta_attributes(meta, rsc, NULL, data_set); if (params) { char *key = NULL; char *value = NULL; g_hash_table_iter_init(&iter, params); while (g_hash_table_iter_next(&iter, (gpointer *) & key, (gpointer *) & value)) { options.params = lrmd_key_value_add(options.params, key, value); } g_hash_table_destroy(params); } if (meta) { char *key = NULL; char *value = NULL; g_hash_table_iter_init(&iter, meta); while (g_hash_table_iter_next(&iter, (gpointer *) & key, (gpointer *) & value)) { char *crm_name = crm_meta_name(key); options.params = lrmd_key_value_add(options.params, crm_name, value); free(crm_name); } g_hash_table_destroy(meta); } param_gen_bail: pe_free_working_set(data_set); return rc; } int main(int argc, char **argv) { int option_index = 0; int argerr = 0; int flag; char *key = NULL; char *val = NULL; gboolean use_tls = FALSE; crm_trigger_t *trig; crm_log_cli_init("cts-exec-helper"); crm_set_options(NULL, "mode [options]", long_options, "Inject commands into the executor, and watch for events\n"); while (1) { flag = crm_get_option(argc, argv, &option_index); if (flag == -1) break; switch (flag) { case '?': crm_help(flag, CRM_EX_OK); break; case 'V': ++options.verbose; crm_bump_log_level(argc, argv); break; case 'Q': options.quiet = 1; options.verbose = 0; break; case 'l': options.listen = optarg; break; case 'w': options.no_wait = 1; break; case 'R': options.is_running = 1; break; case 'n': exec_call_opts = lrmd_opt_notify_orig_only; break; case 'o': exec_call_opts = lrmd_opt_notify_changes_only; break; case 'c': options.api_call = optarg; break; case 'a': options.action = optarg; break; case 'r': options.rsc_id = optarg; break; case 'x': if(optarg) { options.cancel_call_id = atoi(optarg); } break; case 'P': options.provider = optarg; break; case 'C': options.class = optarg; break; case 'T': options.type = optarg; break; case 'i': if(optarg) { options.interval_ms = crm_parse_interval_spec(optarg); } break; case 't': if(optarg) { options.timeout = atoi(optarg); } break; case 's': if(optarg) { options.start_delay = atoi(optarg); } break; case 'k': key = optarg; if (key && val) { options.params = lrmd_key_value_add(options.params, key, val); key = val = NULL; } break; case 'v': val = optarg; if (key && val) { options.params = lrmd_key_value_add(options.params, key, val); key = val = NULL; } break; case 'S': use_tls = TRUE; break; default: ++argerr; break; } } if (argerr) { crm_help('?', CRM_EX_USAGE); } if (optind > argc) { ++argerr; } if (!options.listen && (safe_str_eq(options.api_call, "metadata") || safe_str_eq(options.api_call, "list_agents") || safe_str_eq(options.api_call, "list_standards") || safe_str_eq(options.api_call, "list_ocf_providers"))) { options.no_connect = 1; } crm_log_init(NULL, LOG_INFO, TRUE, (options.verbose? TRUE : FALSE), argc, argv, FALSE); if (options.is_running) { if (!options.timeout) { options.timeout = 30000; } options.interval_ms = 0; if (!options.rsc_id) { crm_err("rsc-id must be given when is-running is used"); test_exit(CRM_EX_ERROR); } if (generate_params()) { print_result(printf ("Failed to retrieve rsc parameters from cib, can not determine if rsc is running.\n")); test_exit(CRM_EX_ERROR); } options.api_call = "exec"; options.action = "monitor"; exec_call_opts = lrmd_opt_notify_orig_only; } /* if we can't perform an api_call or listen for events, * there is nothing to do */ if (!options.api_call && !options.listen) { crm_err("Nothing to be done. Please specify 'api-call' and/or 'listen'"); return CRM_EX_OK; } if (use_tls) { lrmd_conn = lrmd_remote_api_new(NULL, "localhost", 0); } else { lrmd_conn = lrmd_api_new(); } trig = mainloop_add_trigger(G_PRIORITY_HIGH, start_test, NULL); mainloop_set_trigger(trig); mainloop_add_signal(SIGTERM, test_shutdown); crm_info("Starting"); mainloop = g_main_loop_new(NULL, FALSE); g_main_loop_run(mainloop); if (cib_conn != NULL) { cib_conn->cmds->signoff(cib_conn); cib_delete(cib_conn); } test_exit(CRM_EX_OK); return CRM_EX_OK; } diff --git a/daemons/fenced/pacemaker-fenced.c b/daemons/fenced/pacemaker-fenced.c index 7a87f93f7c..7ef4ac0821 100644 --- a/daemons/fenced/pacemaker-fenced.c +++ b/daemons/fenced/pacemaker-fenced.c @@ -1,1521 +1,1521 @@ /* * Copyright 2009-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include /* U32T ~ PRIu32, X32T ~ PRIx32 */ #include #include #include #include #include #include #include #include #include #include #include #include #include char *stonith_our_uname = NULL; char *stonith_our_uuid = NULL; long stonith_watchdog_timeout_ms = 0; static GMainLoop *mainloop = NULL; gboolean stand_alone = FALSE; static gboolean no_cib_connect = FALSE; static gboolean stonith_shutdown_flag = FALSE; static qb_ipcs_service_t *ipcs = NULL; static xmlNode *local_cib = NULL; static pe_working_set_t *fenced_data_set = NULL; static cib_t *cib_api = NULL; static void *cib_library = NULL; static void stonith_shutdown(int nsig); static void stonith_cleanup(void); static int32_t st_ipc_accept(qb_ipcs_connection_t * c, uid_t uid, gid_t gid) { if (stonith_shutdown_flag) { crm_info("Ignoring new client [%d] during shutdown", crm_ipcs_client_pid(c)); return -EPERM; } if (crm_client_new(c, uid, gid) == NULL) { return -EIO; } return 0; } static void st_ipc_created(qb_ipcs_connection_t * c) { crm_trace("Connection created for %p", c); } /* Exit code means? */ static int32_t st_ipc_dispatch(qb_ipcs_connection_t * qbc, void *data, size_t size) { uint32_t id = 0; uint32_t flags = 0; int call_options = 0; xmlNode *request = NULL; crm_client_t *c = crm_client_get(qbc); const char *op = NULL; if (c == NULL) { crm_info("Invalid client: %p", qbc); return 0; } request = crm_ipcs_recv(c, data, size, &id, &flags); if (request == NULL) { crm_ipcs_send_ack(c, id, flags, "nack", __FUNCTION__, __LINE__); return 0; } op = crm_element_value(request, F_CRM_TASK); if(safe_str_eq(op, CRM_OP_RM_NODE_CACHE)) { crm_xml_add(request, F_TYPE, T_STONITH_NG); crm_xml_add(request, F_STONITH_OPERATION, op); crm_xml_add(request, F_STONITH_CLIENTID, c->id); crm_xml_add(request, F_STONITH_CLIENTNAME, crm_client_name(c)); crm_xml_add(request, F_STONITH_CLIENTNODE, stonith_our_uname); send_cluster_message(NULL, crm_msg_stonith_ng, request, FALSE); free_xml(request); return 0; } if (c->name == NULL) { const char *value = crm_element_value(request, F_STONITH_CLIENTNAME); if (value == NULL) { value = "unknown"; } c->name = crm_strdup_printf("%s.%u", value, c->pid); } crm_element_value_int(request, F_STONITH_CALLOPTS, &call_options); crm_trace("Flags %" X32T "/%u for command %" U32T " from %s", flags, call_options, id, crm_client_name(c)); if (is_set(call_options, st_opt_sync_call)) { CRM_ASSERT(flags & crm_ipc_client_response); CRM_LOG_ASSERT(c->request_id == 0); /* This means the client has two synchronous events in-flight */ c->request_id = id; /* Reply only to the last one */ } crm_xml_add(request, F_STONITH_CLIENTID, c->id); crm_xml_add(request, F_STONITH_CLIENTNAME, crm_client_name(c)); crm_xml_add(request, F_STONITH_CLIENTNODE, stonith_our_uname); crm_log_xml_trace(request, "Client[inbound]"); stonith_command(c, id, flags, request, NULL); free_xml(request); return 0; } /* Error code means? */ static int32_t st_ipc_closed(qb_ipcs_connection_t * c) { crm_client_t *client = crm_client_get(c); if (client == NULL) { return 0; } crm_trace("Connection %p closed", c); crm_client_destroy(client); /* 0 means: yes, go ahead and destroy the connection */ return 0; } static void st_ipc_destroy(qb_ipcs_connection_t * c) { crm_trace("Connection %p destroyed", c); st_ipc_closed(c); } static void stonith_peer_callback(xmlNode * msg, void *private_data) { const char *remote_peer = crm_element_value(msg, F_ORIG); const char *op = crm_element_value(msg, F_STONITH_OPERATION); if (crm_str_eq(op, "poke", TRUE)) { return; } crm_log_xml_trace(msg, "Peer[inbound]"); stonith_command(NULL, 0, 0, msg, remote_peer); } #if SUPPORT_COROSYNC static void stonith_peer_ais_callback(cpg_handle_t handle, const struct cpg_name *groupName, uint32_t nodeid, uint32_t pid, void *msg, size_t msg_len) { uint32_t kind = 0; xmlNode *xml = NULL; const char *from = NULL; char *data = pcmk_message_common_cs(handle, nodeid, pid, msg, &kind, &from); if(data == NULL) { return; } if (kind == crm_class_cluster) { xml = string2xml(data); if (xml == NULL) { crm_err("Invalid XML: '%.120s'", data); free(data); return; } crm_xml_add(xml, F_ORIG, from); /* crm_xml_add_int(xml, F_SEQ, wrapper->id); */ stonith_peer_callback(xml, NULL); } free_xml(xml); free(data); return; } static void stonith_peer_cs_destroy(gpointer user_data) { crm_crit("Lost connection to cluster layer, shutting down"); stonith_shutdown(0); } #endif void do_local_reply(xmlNode * notify_src, const char *client_id, gboolean sync_reply, gboolean from_peer) { /* send callback to originating child */ crm_client_t *client_obj = NULL; int local_rc = pcmk_ok; crm_trace("Sending response"); client_obj = crm_client_get_by_id(client_id); crm_trace("Sending callback to request originator"); if (client_obj == NULL) { local_rc = -1; crm_trace("No client to sent the response to. F_STONITH_CLIENTID not set."); } else { int rid = 0; if (sync_reply) { CRM_LOG_ASSERT(client_obj->request_id); rid = client_obj->request_id; client_obj->request_id = 0; crm_trace("Sending response %d to %s %s", rid, client_obj->name, from_peer ? "(originator of delegated request)" : ""); } else { crm_trace("Sending an event to %s %s", client_obj->name, from_peer ? "(originator of delegated request)" : ""); } local_rc = crm_ipcs_send(client_obj, rid, notify_src, sync_reply?crm_ipc_flags_none:crm_ipc_server_event); } if (local_rc < pcmk_ok && client_obj != NULL) { crm_warn("%sSync reply to %s failed: %s", sync_reply ? "" : "A-", client_obj ? client_obj->name : "", pcmk_strerror(local_rc)); } } long long get_stonith_flag(const char *name) { if (safe_str_eq(name, T_STONITH_NOTIFY_FENCE)) { return st_callback_notify_fence; } else if (safe_str_eq(name, STONITH_OP_DEVICE_ADD)) { return st_callback_device_add; } else if (safe_str_eq(name, STONITH_OP_DEVICE_DEL)) { return st_callback_device_del; } else if (safe_str_eq(name, T_STONITH_NOTIFY_HISTORY)) { return st_callback_notify_history; } else if (safe_str_eq(name, T_STONITH_NOTIFY_HISTORY_SYNCED)) { return st_callback_notify_history_synced; } return st_callback_unknown; } static void stonith_notify_client(gpointer key, gpointer value, gpointer user_data) { xmlNode *update_msg = user_data; crm_client_t *client = value; const char *type = NULL; CRM_CHECK(client != NULL, return); CRM_CHECK(update_msg != NULL, return); type = crm_element_value(update_msg, F_SUBTYPE); CRM_CHECK(type != NULL, crm_log_xml_err(update_msg, "notify"); return); if (client->ipcs == NULL) { crm_trace("Skipping client with NULL channel"); return; } if (client->options & get_stonith_flag(type)) { int rc = crm_ipcs_send(client, 0, update_msg, crm_ipc_server_event | crm_ipc_server_error); if (rc <= 0) { crm_warn("%s notification of client %s.%.6s failed: %s (%d)", type, crm_client_name(client), client->id, pcmk_strerror(rc), rc); } else { crm_trace("Sent %s notification to client %s.%.6s", type, crm_client_name(client), client->id); } } } void do_stonith_async_timeout_update(const char *client_id, const char *call_id, int timeout) { crm_client_t *client = NULL; xmlNode *notify_data = NULL; if (!timeout || !call_id || !client_id) { return; } client = crm_client_get_by_id(client_id); if (!client) { return; } notify_data = create_xml_node(NULL, T_STONITH_TIMEOUT_VALUE); crm_xml_add(notify_data, F_TYPE, T_STONITH_TIMEOUT_VALUE); crm_xml_add(notify_data, F_STONITH_CALLID, call_id); crm_xml_add_int(notify_data, F_STONITH_TIMEOUT, timeout); crm_trace("timeout update is %d for client %s and call id %s", timeout, client_id, call_id); if (client) { crm_ipcs_send(client, 0, notify_data, crm_ipc_server_event); } free_xml(notify_data); } void do_stonith_notify(int options, const char *type, int result, xmlNode * data) { /* TODO: Standardize the contents of data */ xmlNode *update_msg = create_xml_node(NULL, "notify"); CRM_CHECK(type != NULL,;); crm_xml_add(update_msg, F_TYPE, T_STONITH_NOTIFY); crm_xml_add(update_msg, F_SUBTYPE, type); crm_xml_add(update_msg, F_STONITH_OPERATION, type); crm_xml_add_int(update_msg, F_STONITH_RC, result); if (data != NULL) { add_message_xml(update_msg, F_STONITH_CALLDATA, data); } crm_trace("Notifying clients"); g_hash_table_foreach(client_connections, stonith_notify_client, update_msg); free_xml(update_msg); crm_trace("Notify complete"); } static void do_stonith_notify_config(int options, const char *op, int rc, const char *desc, int active) { xmlNode *notify_data = create_xml_node(NULL, op); CRM_CHECK(notify_data != NULL, return); crm_xml_add(notify_data, F_STONITH_DEVICE, desc); crm_xml_add_int(notify_data, F_STONITH_ACTIVE, active); do_stonith_notify(options, op, rc, notify_data); free_xml(notify_data); } void do_stonith_notify_device(int options, const char *op, int rc, const char *desc) { do_stonith_notify_config(options, op, rc, desc, g_hash_table_size(device_list)); } void do_stonith_notify_level(int options, const char *op, int rc, const char *desc) { do_stonith_notify_config(options, op, rc, desc, g_hash_table_size(topology)); } static void topology_remove_helper(const char *node, int level) { int rc; char *desc = NULL; xmlNode *data = create_xml_node(NULL, XML_TAG_FENCING_LEVEL); crm_xml_add(data, F_STONITH_ORIGIN, __FUNCTION__); crm_xml_add_int(data, XML_ATTR_STONITH_INDEX, level); crm_xml_add(data, XML_ATTR_STONITH_TARGET, node); rc = stonith_level_remove(data, &desc); do_stonith_notify_level(0, STONITH_OP_LEVEL_DEL, rc, desc); free_xml(data); free(desc); } static void remove_cib_device(xmlXPathObjectPtr xpathObj) { int max = numXpathResults(xpathObj), lpc = 0; for (lpc = 0; lpc < max; lpc++) { const char *rsc_id = NULL; const char *standard = NULL; xmlNode *match = getXpathResult(xpathObj, lpc); CRM_LOG_ASSERT(match != NULL); if(match != NULL) { standard = crm_element_value(match, XML_AGENT_ATTR_CLASS); } if (safe_str_neq(standard, PCMK_RESOURCE_CLASS_STONITH)) { continue; } rsc_id = crm_element_value(match, XML_ATTR_ID); stonith_device_remove(rsc_id, TRUE); } } static void handle_topology_change(xmlNode *match, bool remove) { int rc; char *desc = NULL; CRM_CHECK(match != NULL, return); crm_trace("Updating %s", ID(match)); if(remove) { int index = 0; char *key = stonith_level_key(match, -1); crm_element_value_int(match, XML_ATTR_STONITH_INDEX, &index); topology_remove_helper(key, index); free(key); } rc = stonith_level_register(match, &desc); do_stonith_notify_level(0, STONITH_OP_LEVEL_ADD, rc, desc); free(desc); } static void remove_fencing_topology(xmlXPathObjectPtr xpathObj) { int max = numXpathResults(xpathObj), lpc = 0; for (lpc = 0; lpc < max; lpc++) { xmlNode *match = getXpathResult(xpathObj, lpc); CRM_LOG_ASSERT(match != NULL); if (match && crm_element_value(match, XML_DIFF_MARKER)) { /* Deletion */ int index = 0; char *target = stonith_level_key(match, -1); crm_element_value_int(match, XML_ATTR_STONITH_INDEX, &index); if (target == NULL) { crm_err("Invalid fencing target in element %s", ID(match)); } else if (index <= 0) { crm_err("Invalid level for %s in element %s", target, ID(match)); } else { topology_remove_helper(target, index); } /* } else { Deal with modifications during the 'addition' stage */ } } } static void register_fencing_topology(xmlXPathObjectPtr xpathObj) { int max = numXpathResults(xpathObj), lpc = 0; for (lpc = 0; lpc < max; lpc++) { xmlNode *match = getXpathResult(xpathObj, lpc); handle_topology_change(match, TRUE); } } /* Fencing */ static void fencing_topology_init() { xmlXPathObjectPtr xpathObj = NULL; const char *xpath = "//" XML_TAG_FENCING_LEVEL; crm_trace("Full topology refresh"); free_topology_list(); init_topology_list(); /* Grab everything */ xpathObj = xpath_search(local_cib, xpath); register_fencing_topology(xpathObj); freeXpathObject(xpathObj); } #define rsc_name(x) x->clone_name?x->clone_name:x->id /*! * \internal * \brief Check whether our uname is in a resource's allowed node list * * \param[in] rsc Resource to check * * \return Pointer to node object if found, NULL otherwise */ static node_t * our_node_allowed_for(resource_t *rsc) { GHashTableIter iter; node_t *node = NULL; if (rsc && stonith_our_uname) { g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&node)) { if (node && strcmp(node->details->uname, stonith_our_uname) == 0) { break; } node = NULL; } } return node; } /*! * \internal * \brief If a resource or any of its children are STONITH devices, update their * definitions given a cluster working set. * * \param[in] rsc Resource to check * \param[in] data_set Cluster working set with device information */ static void cib_device_update(resource_t *rsc, pe_working_set_t *data_set) { node_t *node = NULL; const char *value = NULL; const char *rclass = NULL; node_t *parent = NULL; gboolean remove = TRUE; /* If this is a complex resource, check children rather than this resource itself. * TODO: Mark each installed device and remove if untouched when this process finishes. */ if(rsc->children) { GListPtr gIter = NULL; for (gIter = rsc->children; gIter != NULL; gIter = gIter->next) { cib_device_update(gIter->data, data_set); if(pe_rsc_is_clone(rsc)) { crm_trace("Only processing one copy of the clone %s", rsc->id); break; } } return; } /* We only care about STONITH resources. */ rclass = crm_element_value(rsc->xml, XML_AGENT_ATTR_CLASS); if (safe_str_neq(rclass, PCMK_RESOURCE_CLASS_STONITH)) { return; } /* If this STONITH resource is disabled, just remove it. */ - value = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_TARGET_ROLE); - if (safe_str_eq(value, RSC_STOPPED)) { + if (pe__resource_is_disabled(rsc)) { crm_info("Device %s has been disabled", rsc->id); goto update_done; } /* Check whether our node is allowed for this resource (and its parent if in a group) */ node = our_node_allowed_for(rsc); if (rsc->parent && (rsc->parent->variant == pe_group)) { parent = our_node_allowed_for(rsc->parent); } if(node == NULL) { /* Our node is disallowed, so remove the device */ GHashTableIter iter; crm_info("Device %s has been disabled on %s: unknown", rsc->id, stonith_our_uname); g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&node)) { crm_trace("Available: %s = %d", node->details->uname, node->weight); } goto update_done; } else if(node->weight < 0 || (parent && parent->weight < 0)) { /* Our node (or its group) is disallowed by score, so remove the device */ char *score = score2char((node->weight < 0) ? node->weight : parent->weight); crm_info("Device %s has been disabled on %s: score=%s", rsc->id, stonith_our_uname, score); free(score); goto update_done; } else { /* Our node is allowed, so update the device information */ int rc; xmlNode *data; GHashTableIter gIter; stonith_key_value_t *params = NULL; const char *name = NULL; const char *agent = crm_element_value(rsc->xml, XML_EXPR_ATTR_TYPE); const char *rsc_provides = NULL; crm_debug("Device %s is allowed on %s: score=%d", rsc->id, stonith_our_uname, node->weight); get_rsc_attributes(rsc->parameters, rsc, node, data_set); get_meta_attributes(rsc->meta, rsc, node, data_set); rsc_provides = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_PROVIDES); g_hash_table_iter_init(&gIter, rsc->parameters); while (g_hash_table_iter_next(&gIter, (gpointer *) & name, (gpointer *) & value)) { if (!name || !value) { continue; } params = stonith_key_value_add(params, name, value); crm_trace(" %s=%s", name, value); } remove = FALSE; data = create_device_registration_xml(rsc_name(rsc), st_namespace_any, agent, params, rsc_provides); stonith_key_value_freeall(params, 1, 1); rc = stonith_device_register(data, NULL, TRUE); CRM_ASSERT(rc == pcmk_ok); free_xml(data); } update_done: if(remove && g_hash_table_lookup(device_list, rsc_name(rsc))) { stonith_device_remove(rsc_name(rsc), TRUE); } } /*! * \internal * \brief Update all STONITH device definitions based on current CIB */ static void cib_devices_update(void) { GListPtr gIter = NULL; crm_info("Updating devices to version %s.%s.%s", crm_element_value(local_cib, XML_ATTR_GENERATION_ADMIN), crm_element_value(local_cib, XML_ATTR_GENERATION), crm_element_value(local_cib, XML_ATTR_NUMUPDATES)); CRM_ASSERT(fenced_data_set != NULL); fenced_data_set->input = local_cib; fenced_data_set->now = crm_time_new(NULL); fenced_data_set->flags |= pe_flag_quick_location; fenced_data_set->localhost = stonith_our_uname; cluster_status(fenced_data_set); pcmk__schedule_actions(fenced_data_set, NULL, NULL); for (gIter = fenced_data_set->resources; gIter != NULL; gIter = gIter->next) { cib_device_update(gIter->data, fenced_data_set); } fenced_data_set->input = NULL; // Wasn't a copy, so don't let API free it pe_reset_working_set(fenced_data_set); } static void update_cib_stonith_devices_v2(const char *event, xmlNode * msg) { xmlNode *change = NULL; char *reason = NULL; bool needs_update = FALSE; xmlNode *patchset = get_message_xml(msg, F_CIB_UPDATE_RESULT); for (change = __xml_first_child(patchset); change != NULL; change = __xml_next(change)) { const char *op = crm_element_value(change, XML_DIFF_OP); const char *xpath = crm_element_value(change, XML_DIFF_PATH); const char *shortpath = NULL; if ((op == NULL) || (strcmp(op, "move") == 0) || strstr(xpath, "/"XML_CIB_TAG_STATUS)) { continue; } else if (safe_str_eq(op, "delete") && strstr(xpath, "/"XML_CIB_TAG_RESOURCE)) { const char *rsc_id = NULL; char *search = NULL; char *mutable = NULL; if (strstr(xpath, XML_TAG_ATTR_SETS) || strstr(xpath, XML_TAG_META_SETS)) { needs_update = TRUE; reason = strdup("(meta) attribute deleted from resource"); break; } mutable = strdup(xpath); rsc_id = strstr(mutable, "primitive[@id=\'"); if (rsc_id != NULL) { rsc_id += strlen("primitive[@id=\'"); search = strchr(rsc_id, '\''); } if (search != NULL) { *search = 0; stonith_device_remove(rsc_id, TRUE); } else { crm_warn("Ignoring malformed CIB update (resource deletion)"); } free(mutable); } else if (strstr(xpath, "/"XML_CIB_TAG_RESOURCES) || strstr(xpath, "/"XML_CIB_TAG_CONSTRAINTS) || strstr(xpath, "/"XML_CIB_TAG_RSCCONFIG)) { shortpath = strrchr(xpath, '/'); CRM_ASSERT(shortpath); reason = crm_strdup_printf("%s %s", op, shortpath+1); needs_update = TRUE; break; } } if(needs_update) { crm_info("Updating device list from the cib: %s", reason); cib_devices_update(); } else { crm_trace("No updates for device list found in cib"); } free(reason); } static void update_cib_stonith_devices_v1(const char *event, xmlNode * msg) { const char *reason = "none"; gboolean needs_update = FALSE; xmlXPathObjectPtr xpath_obj = NULL; /* process new constraints */ xpath_obj = xpath_search(msg, "//" F_CIB_UPDATE_RESULT "//" XML_CONS_TAG_RSC_LOCATION); if (numXpathResults(xpath_obj) > 0) { int max = numXpathResults(xpath_obj), lpc = 0; /* Safest and simplest to always recompute */ needs_update = TRUE; reason = "new location constraint"; for (lpc = 0; lpc < max; lpc++) { xmlNode *match = getXpathResult(xpath_obj, lpc); crm_log_xml_trace(match, "new constraint"); } } freeXpathObject(xpath_obj); /* process deletions */ xpath_obj = xpath_search(msg, "//" F_CIB_UPDATE_RESULT "//" XML_TAG_DIFF_REMOVED "//" XML_CIB_TAG_RESOURCE); if (numXpathResults(xpath_obj) > 0) { remove_cib_device(xpath_obj); } freeXpathObject(xpath_obj); /* process additions */ xpath_obj = xpath_search(msg, "//" F_CIB_UPDATE_RESULT "//" XML_TAG_DIFF_ADDED "//" XML_CIB_TAG_RESOURCE); if (numXpathResults(xpath_obj) > 0) { int max = numXpathResults(xpath_obj), lpc = 0; for (lpc = 0; lpc < max; lpc++) { const char *rsc_id = NULL; const char *standard = NULL; xmlNode *match = getXpathResult(xpath_obj, lpc); rsc_id = crm_element_value(match, XML_ATTR_ID); standard = crm_element_value(match, XML_AGENT_ATTR_CLASS); if (safe_str_neq(standard, PCMK_RESOURCE_CLASS_STONITH)) { continue; } crm_trace("Fencing resource %s was added or modified", rsc_id); reason = "new resource"; needs_update = TRUE; } } freeXpathObject(xpath_obj); if(needs_update) { crm_info("Updating device list from the cib: %s", reason); cib_devices_update(); } } static void update_cib_stonith_devices(const char *event, xmlNode * msg) { int format = 1; xmlNode *patchset = get_message_xml(msg, F_CIB_UPDATE_RESULT); CRM_ASSERT(patchset); crm_element_value_int(patchset, "format", &format); switch(format) { case 1: update_cib_stonith_devices_v1(event, msg); break; case 2: update_cib_stonith_devices_v2(event, msg); break; default: crm_warn("Unknown patch format: %d", format); } } /* Needs to hold node name + attribute name + attribute value + 75 */ #define XPATH_MAX 512 /*! * \internal * \brief Check whether a node has a specific attribute name/value * * \param[in] node Name of node to check * \param[in] name Name of an attribute to look for * \param[in] value The value the named attribute needs to be set to in order to be considered a match * * \return TRUE if the locally cached CIB has the specified node attribute */ gboolean node_has_attr(const char *node, const char *name, const char *value) { char xpath[XPATH_MAX]; xmlNode *match; int n; CRM_CHECK(local_cib != NULL, return FALSE); /* Search for the node's attributes in the CIB. While the schema allows * multiple sets of instance attributes, and allows instance attributes to * use id-ref to reference values elsewhere, that is intended for resources, * so we ignore that here. */ n = snprintf(xpath, XPATH_MAX, "//" XML_CIB_TAG_NODES "/" XML_CIB_TAG_NODE "[@uname='%s']/" XML_TAG_ATTR_SETS "/" XML_CIB_TAG_NVPAIR "[@name='%s' and @value='%s']", node, name, value); match = get_xpath_object(xpath, local_cib, LOG_TRACE); CRM_CHECK(n < XPATH_MAX, return FALSE); return (match != NULL); } static void update_fencing_topology(const char *event, xmlNode * msg) { int format = 1; const char *xpath; xmlXPathObjectPtr xpathObj = NULL; xmlNode *patchset = get_message_xml(msg, F_CIB_UPDATE_RESULT); CRM_ASSERT(patchset); crm_element_value_int(patchset, "format", &format); if(format == 1) { /* Process deletions (only) */ xpath = "//" F_CIB_UPDATE_RESULT "//" XML_TAG_DIFF_REMOVED "//" XML_TAG_FENCING_LEVEL; xpathObj = xpath_search(msg, xpath); remove_fencing_topology(xpathObj); freeXpathObject(xpathObj); /* Process additions and changes */ xpath = "//" F_CIB_UPDATE_RESULT "//" XML_TAG_DIFF_ADDED "//" XML_TAG_FENCING_LEVEL; xpathObj = xpath_search(msg, xpath); register_fencing_topology(xpathObj); freeXpathObject(xpathObj); } else if(format == 2) { xmlNode *change = NULL; int add[] = { 0, 0, 0 }; int del[] = { 0, 0, 0 }; xml_patch_versions(patchset, add, del); for (change = __xml_first_child(patchset); change != NULL; change = __xml_next(change)) { const char *op = crm_element_value(change, XML_DIFF_OP); const char *xpath = crm_element_value(change, XML_DIFF_PATH); if(op == NULL) { continue; } else if(strstr(xpath, "/" XML_TAG_FENCING_LEVEL) != NULL) { /* Change to a specific entry */ crm_trace("Handling %s operation %d.%d.%d for %s", op, add[0], add[1], add[2], xpath); if(strcmp(op, "move") == 0) { continue; } else if(strcmp(op, "create") == 0) { handle_topology_change(change->children, FALSE); } else if(strcmp(op, "modify") == 0) { xmlNode *match = first_named_child(change, XML_DIFF_RESULT); if(match) { handle_topology_change(match->children, TRUE); } } else if(strcmp(op, "delete") == 0) { /* Nuclear option, all we have is the path and an id... not enough to remove a specific entry */ crm_info("Re-initializing fencing topology after %s operation %d.%d.%d for %s", op, add[0], add[1], add[2], xpath); fencing_topology_init(); return; } } else if (strstr(xpath, "/" XML_TAG_FENCING_TOPOLOGY) != NULL) { /* Change to the topology in general */ crm_info("Re-initializing fencing topology after top-level %s operation %d.%d.%d for %s", op, add[0], add[1], add[2], xpath); fencing_topology_init(); return; } else if (strstr(xpath, "/" XML_CIB_TAG_CONFIGURATION)) { /* Changes to the whole config section, possibly including the topology as a whild */ if(first_named_child(change, XML_TAG_FENCING_TOPOLOGY) == NULL) { crm_trace("Nothing for us in %s operation %d.%d.%d for %s.", op, add[0], add[1], add[2], xpath); } else if(strcmp(op, "delete") == 0 || strcmp(op, "create") == 0) { crm_info("Re-initializing fencing topology after top-level %s operation %d.%d.%d for %s.", op, add[0], add[1], add[2], xpath); fencing_topology_init(); return; } } else { crm_trace("Nothing for us in %s operation %d.%d.%d for %s", op, add[0], add[1], add[2], xpath); } } } else { crm_warn("Unknown patch format: %d", format); } } static bool have_cib_devices = FALSE; static void update_cib_cache_cb(const char *event, xmlNode * msg) { int rc = pcmk_ok; xmlNode *stonith_enabled_xml = NULL; xmlNode *stonith_watchdog_xml = NULL; const char *stonith_enabled_s = NULL; static gboolean stonith_enabled_saved = TRUE; if(!have_cib_devices) { crm_trace("Skipping updates until we get a full dump"); return; } else if(msg == NULL) { crm_trace("Missing %s update", event); return; } /* Maintain a local copy of the CIB so that we have full access * to device definitions, location constraints, and node attributes */ if (local_cib != NULL) { int rc = pcmk_ok; xmlNode *patchset = NULL; crm_element_value_int(msg, F_CIB_RC, &rc); if (rc != pcmk_ok) { return; } patchset = get_message_xml(msg, F_CIB_UPDATE_RESULT); xml_log_patchset(LOG_TRACE, "Config update", patchset); rc = xml_apply_patchset(local_cib, patchset, TRUE); switch (rc) { case pcmk_ok: case -pcmk_err_old_data: break; case -pcmk_err_diff_resync: case -pcmk_err_diff_failed: crm_notice("[%s] Patch aborted: %s (%d)", event, pcmk_strerror(rc), rc); free_xml(local_cib); local_cib = NULL; break; default: crm_warn("[%s] ABORTED: %s (%d)", event, pcmk_strerror(rc), rc); free_xml(local_cib); local_cib = NULL; } } if (local_cib == NULL) { crm_trace("Re-requesting the full cib"); rc = cib_api->cmds->query(cib_api, NULL, &local_cib, cib_scope_local | cib_sync_call); if(rc != pcmk_ok) { crm_err("Couldn't retrieve the CIB: %s (%d)", pcmk_strerror(rc), rc); return; } CRM_ASSERT(local_cib != NULL); stonith_enabled_saved = FALSE; /* Trigger a full refresh below */ } crm_peer_caches_refresh(local_cib); stonith_enabled_xml = get_xpath_object("//nvpair[@name='stonith-enabled']", local_cib, LOG_TRACE); if (stonith_enabled_xml) { stonith_enabled_s = crm_element_value(stonith_enabled_xml, XML_NVPAIR_ATTR_VALUE); } if (stonith_enabled_s == NULL || crm_is_true(stonith_enabled_s)) { long timeout_ms = 0; const char *value = NULL; stonith_watchdog_xml = get_xpath_object("//nvpair[@name='stonith-watchdog-timeout']", local_cib, LOG_TRACE); if (stonith_watchdog_xml) { value = crm_element_value(stonith_watchdog_xml, XML_NVPAIR_ATTR_VALUE); } if(value) { timeout_ms = crm_get_msec(value); } if (timeout_ms < 0) { timeout_ms = crm_auto_watchdog_timeout(); } if(timeout_ms != stonith_watchdog_timeout_ms) { crm_notice("New watchdog timeout %lds (was %lds)", timeout_ms/1000, stonith_watchdog_timeout_ms/1000); stonith_watchdog_timeout_ms = timeout_ms; } } else { stonith_watchdog_timeout_ms = 0; } if (stonith_enabled_s && crm_is_true(stonith_enabled_s) == FALSE) { crm_trace("Ignoring cib updates while stonith is disabled"); stonith_enabled_saved = FALSE; return; } else if (stonith_enabled_saved == FALSE) { crm_info("Updating stonith device and topology lists now that stonith is enabled"); stonith_enabled_saved = TRUE; fencing_topology_init(); cib_devices_update(); } else { update_fencing_topology(event, msg); update_cib_stonith_devices(event, msg); } } static void init_cib_cache_cb(xmlNode * msg, int call_id, int rc, xmlNode * output, void *user_data) { crm_info("Updating device list from the cib: init"); have_cib_devices = TRUE; local_cib = copy_xml(output); crm_peer_caches_refresh(local_cib); fencing_topology_init(); cib_devices_update(); } static void stonith_shutdown(int nsig) { stonith_shutdown_flag = TRUE; crm_info("Terminating with %d clients", crm_hash_table_size(client_connections)); if (mainloop != NULL && g_main_loop_is_running(mainloop)) { g_main_loop_quit(mainloop); } else { stonith_cleanup(); crm_exit(CRM_EX_OK); } } static void cib_connection_destroy(gpointer user_data) { if (stonith_shutdown_flag) { crm_info("Connection to the CIB manager closed"); return; } else { crm_crit("Lost connection to the CIB manager, shutting down"); } if (cib_api) { cib_api->cmds->signoff(cib_api); } stonith_shutdown(0); } static void stonith_cleanup(void) { if (cib_api) { cib_api->cmds->signoff(cib_api); } if (ipcs) { qb_ipcs_destroy(ipcs); } crm_peer_destroy(); crm_client_cleanup(); free_stonith_remote_op_list(); free_topology_list(); free_device_list(); free_metadata_cache(); free(stonith_our_uname); stonith_our_uname = NULL; free_xml(local_cib); local_cib = NULL; } /* *INDENT-OFF* */ static struct crm_option long_options[] = { {"stand-alone", 0, 0, 's'}, {"stand-alone-w-cpg", 0, 0, 'c'}, {"logfile", 1, 0, 'l'}, {"verbose", 0, 0, 'V'}, {"version", 0, 0, '$'}, {"help", 0, 0, '?'}, {0, 0, 0, 0} }; /* *INDENT-ON* */ static void setup_cib(void) { int rc, retries = 0; static cib_t *(*cib_new_fn) (void) = NULL; if (cib_new_fn == NULL) { cib_new_fn = find_library_function(&cib_library, CIB_LIBRARY, "cib_new", TRUE); } if (cib_new_fn != NULL) { cib_api = (*cib_new_fn) (); } if (cib_api == NULL) { crm_err("No connection to the CIB manager"); return; } do { sleep(retries); rc = cib_api->cmds->signon(cib_api, CRM_SYSTEM_STONITHD, cib_command); } while (rc == -ENOTCONN && ++retries < 5); if (rc != pcmk_ok) { crm_err("Could not connect to the CIB manager: %s (%d)", pcmk_strerror(rc), rc); } else if (pcmk_ok != cib_api->cmds->add_notify_callback(cib_api, T_CIB_DIFF_NOTIFY, update_cib_cache_cb)) { crm_err("Could not set CIB notification callback"); } else { rc = cib_api->cmds->query(cib_api, NULL, NULL, cib_scope_local); cib_api->cmds->register_callback(cib_api, rc, 120, FALSE, NULL, "init_cib_cache_cb", init_cib_cache_cb); cib_api->cmds->set_connection_dnotify(cib_api, cib_connection_destroy); crm_info("Watching for stonith topology changes"); } } struct qb_ipcs_service_handlers ipc_callbacks = { .connection_accept = st_ipc_accept, .connection_created = st_ipc_created, .msg_process = st_ipc_dispatch, .connection_closed = st_ipc_closed, .connection_destroyed = st_ipc_destroy }; /*! * \internal * \brief Callback for peer status changes * * \param[in] type What changed * \param[in] node What peer had the change * \param[in] data Previous value of what changed */ static void st_peer_update_callback(enum crm_status_type type, crm_node_t * node, const void *data) { if ((type != crm_status_processes) && !is_set(node->flags, crm_remote_node)) { /* * This is a hack until we can send to a nodeid and/or we fix node name lookups * These messages are ignored in stonith_peer_callback() */ xmlNode *query = create_xml_node(NULL, "stonith_command"); crm_xml_add(query, F_XML_TAGNAME, "stonith_command"); crm_xml_add(query, F_TYPE, T_STONITH_NG); crm_xml_add(query, F_STONITH_OPERATION, "poke"); crm_debug("Broadcasting our uname because of node %u", node->id); send_cluster_message(NULL, crm_msg_stonith_ng, query, FALSE); free_xml(query); } } int main(int argc, char **argv) { int flag; int lpc = 0; int argerr = 0; int option_index = 0; crm_cluster_t cluster; const char *actions[] = { "reboot", "off", "on", "list", "monitor", "status" }; crm_ipc_t *old_instance = NULL; crm_log_preinit(NULL, argc, argv); crm_set_options(NULL, "mode [options]", long_options, "Provides a summary of cluster's current state." "\n\nOutputs varying levels of detail in a number of different formats.\n"); while (1) { flag = crm_get_option(argc, argv, &option_index); if (flag == -1) { break; } switch (flag) { case 'V': crm_bump_log_level(argc, argv); break; case 'l': crm_add_logfile(optarg); break; case 's': stand_alone = TRUE; break; case 'c': stand_alone = FALSE; no_cib_connect = TRUE; break; case '$': case '?': crm_help(flag, CRM_EX_OK); break; default: ++argerr; break; } } if (argc - optind == 1 && safe_str_eq("metadata", argv[optind])) { printf("\n"); printf("\n"); printf(" 1.0\n"); printf(" Instance attributes available for all \"stonith\"-class resources" " and used by Pacemaker's fence daemon, formerly known as stonithd\n"); printf(" Instance attributes available for all \"stonith\"-class resources\n"); printf(" \n"); #if 0 // priority is not implemented yet printf(" \n"); printf(" Devices that are not in a topology " "are tried in order of highest to lowest integer priority\n"); printf(" \n"); printf(" \n"); #endif printf(" \n", STONITH_ATTR_HOSTARG); printf (" Advanced use only: An alternate parameter to supply instead of 'port'\n"); printf (" Some devices do not support the standard 'port' parameter or may provide additional ones.\n" "Use this to specify an alternate, device-specific, parameter that should indicate the machine to be fenced.\n" "A value of 'none' can be used to tell the cluster not to supply any additional parameters.\n" " \n"); printf(" \n"); printf(" \n"); printf(" \n", STONITH_ATTR_HOSTMAP); printf (" A mapping of host names to ports numbers for devices that do not support host names.\n"); printf (" Eg. node1:1;node2:2,3 would tell the cluster to use port 1 for node1 and ports 2 and 3 for node2\n"); printf(" \n"); printf(" \n"); printf(" \n", STONITH_ATTR_HOSTLIST); printf (" A list of machines controlled by this device (Optional unless %s=static-list).\n", STONITH_ATTR_HOSTCHECK); printf(" \n"); printf(" \n"); printf(" \n", STONITH_ATTR_HOSTCHECK); printf (" How to determine which machines are controlled by the device.\n"); printf(" Allowed values: dynamic-list " "(query the device via the 'list' command), static-list " "(check the " STONITH_ATTR_HOSTLIST " attribute), status " "(query the device via the 'status' command), none (assume " "every device can fence every machine)\n"); printf(" \n"); printf(" \n"); printf(" \n", STONITH_ATTR_DELAY_MAX); printf (" Enable a random delay for stonith actions and specify the maximum of random delay.\n"); printf (" This prevents double fencing when using slow devices such as sbd.\n" "Use this to enable a random delay for stonith actions.\n" "The overall delay is derived from this random delay value adding a static delay so that the sum is kept below the maximum delay.\n"); printf(" \n"); printf(" \n"); printf(" \n", STONITH_ATTR_DELAY_BASE); printf (" Enable a base delay for stonith actions and specify base delay value.\n"); printf (" This prevents double fencing when different delays are configured on the nodes.\n" "Use this to enable a static delay for stonith actions.\n" "The overall delay is derived from a random delay value adding this static delay so that the sum is kept below the maximum delay.\n"); printf(" \n"); printf(" \n"); printf(" \n", STONITH_ATTR_ACTION_LIMIT); printf (" The maximum number of actions can be performed in parallel on this device\n"); printf (" Cluster property concurrent-fencing=true needs to be configured first.\n" "Then use this to specify the maximum number of actions can be performed in parallel on this device. -1 is unlimited.\n"); printf(" \n"); printf(" \n"); for (lpc = 0; lpc < DIMOF(actions); lpc++) { printf(" \n", actions[lpc]); printf (" Advanced use only: An alternate command to run instead of '%s'\n", actions[lpc]); printf (" Some devices do not support the standard commands or may provide additional ones.\n" "Use this to specify an alternate, device-specific, command that implements the '%s' action.\n", actions[lpc]); printf(" \n", actions[lpc]); printf(" \n"); printf(" \n", actions[lpc]); printf (" Advanced use only: Specify an alternate timeout to use for %s actions instead of stonith-timeout\n", actions[lpc]); printf (" Some devices need much more/less time to complete than normal.\n" "Use this to specify an alternate, device-specific, timeout for '%s' actions.\n", actions[lpc]); printf(" \n"); printf(" \n"); printf(" \n", actions[lpc]); printf (" Advanced use only: The maximum number of times to retry the '%s' command within the timeout period\n", actions[lpc]); printf(" Some devices do not support multiple connections." " Operations may 'fail' if the device is busy with another task so Pacemaker will automatically retry the operation, if there is time remaining." " Use this option to alter the number of times Pacemaker retries '%s' actions before giving up." "\n", actions[lpc]); printf(" \n"); printf(" \n"); } printf(" \n"); printf("\n"); return CRM_EX_OK; } if (optind != argc) { ++argerr; } if (argerr) { crm_help('?', CRM_EX_USAGE); } crm_log_init(NULL, LOG_INFO, TRUE, FALSE, argc, argv, FALSE); crm_notice("Starting Pacemaker fencer"); old_instance = crm_ipc_new("stonith-ng", 0); if (crm_ipc_connect(old_instance)) { /* IPC end-point already up */ crm_ipc_close(old_instance); crm_ipc_destroy(old_instance); crm_err("pacemaker-fenced is already active, aborting startup"); crm_exit(CRM_EX_OK); } else { /* not up or not authentic, we'll proceed either way */ crm_ipc_destroy(old_instance); old_instance = NULL; } mainloop_add_signal(SIGTERM, stonith_shutdown); crm_peer_init(); fenced_data_set = pe_new_working_set(); CRM_ASSERT(fenced_data_set != NULL); + set_bit(fenced_data_set->flags, pe_flag_no_counts); if (stand_alone == FALSE) { if (is_corosync_cluster()) { #if SUPPORT_COROSYNC cluster.destroy = stonith_peer_cs_destroy; cluster.cpg.cpg_deliver_fn = stonith_peer_ais_callback; cluster.cpg.cpg_confchg_fn = pcmk_cpg_membership; #endif } crm_set_status_callback(&st_peer_update_callback); if (crm_cluster_connect(&cluster) == FALSE) { crm_crit("Cannot sign in to the cluster... terminating"); crm_exit(CRM_EX_FATAL); } stonith_our_uname = cluster.uname; stonith_our_uuid = cluster.uuid; if (no_cib_connect == FALSE) { setup_cib(); } } else { stonith_our_uname = strdup("localhost"); } init_device_list(); init_topology_list(); if(stonith_watchdog_timeout_ms > 0) { int rc; xmlNode *xml; stonith_key_value_t *params = NULL; params = stonith_key_value_add(params, STONITH_ATTR_HOSTLIST, stonith_our_uname); xml = create_device_registration_xml("watchdog", st_namespace_internal, STONITH_WATCHDOG_AGENT, params, NULL); stonith_key_value_freeall(params, 1, 1); rc = stonith_device_register(xml, NULL, FALSE); free_xml(xml); if (rc != pcmk_ok) { crm_crit("Cannot register watchdog pseudo fence agent"); crm_exit(CRM_EX_FATAL); } } stonith_ipc_server_init(&ipcs, &ipc_callbacks); /* Create the mainloop and run it... */ mainloop = g_main_loop_new(NULL, FALSE); crm_notice("Pacemaker fencer successfully started and accepting connections"); g_main_loop_run(mainloop); stonith_cleanup(); pe_free_working_set(fenced_data_set); crm_exit(CRM_EX_OK); } diff --git a/daemons/schedulerd/pacemaker-schedulerd.c b/daemons/schedulerd/pacemaker-schedulerd.c index 878ec59ba2..91a6335362 100644 --- a/daemons/schedulerd/pacemaker-schedulerd.c +++ b/daemons/schedulerd/pacemaker-schedulerd.c @@ -1,351 +1,352 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define OPTARGS "hVc" static GMainLoop *mainloop = NULL; static qb_ipcs_service_t *ipcs = NULL; static pe_working_set_t *sched_data_set = NULL; #define get_series() was_processing_error?1:was_processing_warning?2:3 typedef struct series_s { const char *name; const char *param; int wrap; } series_t; series_t series[] = { {"pe-unknown", "_do_not_match_anything_", -1}, {"pe-error", "pe-error-series-max", -1}, {"pe-warn", "pe-warn-series-max", 200}, {"pe-input", "pe-input-series-max", 400}, }; void pengine_shutdown(int nsig); static gboolean process_pe_message(xmlNode * msg, xmlNode * xml_data, crm_client_t * sender) { static char *last_digest = NULL; static char *filename = NULL; const char *sys_to = crm_element_value(msg, F_CRM_SYS_TO); const char *op = crm_element_value(msg, F_CRM_TASK); const char *ref = crm_element_value(msg, F_CRM_REFERENCE); crm_trace("Processing %s op (ref=%s)...", op, ref); if (op == NULL) { /* error */ } else if (strcasecmp(op, CRM_OP_HELLO) == 0) { /* ignore */ } else if (safe_str_eq(crm_element_value(msg, F_CRM_MSG_TYPE), XML_ATTR_RESPONSE)) { /* ignore */ } else if (sys_to == NULL || strcasecmp(sys_to, CRM_SYSTEM_PENGINE) != 0) { crm_trace("Bad sys-to %s", crm_str(sys_to)); return FALSE; } else if (strcasecmp(op, CRM_OP_PECALC) == 0) { int seq = -1; int series_id = 0; int series_wrap = 0; char *digest = NULL; const char *value = NULL; time_t execution_date = time(NULL); xmlNode *converted = NULL; xmlNode *reply = NULL; gboolean is_repoke = FALSE; gboolean process = TRUE; crm_config_error = FALSE; crm_config_warning = FALSE; was_processing_error = FALSE; was_processing_warning = FALSE; if (sched_data_set == NULL) { sched_data_set = pe_new_working_set(); CRM_ASSERT(sched_data_set != NULL); + set_bit(sched_data_set->flags, pe_flag_no_counts); } digest = calculate_xml_versioned_digest(xml_data, FALSE, FALSE, CRM_FEATURE_SET); converted = copy_xml(xml_data); if (cli_config_update(&converted, NULL, TRUE) == FALSE) { sched_data_set->graph = create_xml_node(NULL, XML_TAG_GRAPH); crm_xml_add_int(sched_data_set->graph, "transition_id", 0); crm_xml_add_int(sched_data_set->graph, "cluster-delay", 0); process = FALSE; free(digest); } else if (safe_str_eq(digest, last_digest)) { crm_info("Input has not changed since last time, not saving to disk"); is_repoke = TRUE; free(digest); } else { free(last_digest); last_digest = digest; } if (process) { pcmk__schedule_actions(sched_data_set, converted, NULL); } series_id = get_series(); series_wrap = series[series_id].wrap; value = pe_pref(sched_data_set->config_hash, series[series_id].param); if (value != NULL) { series_wrap = crm_int_helper(value, NULL); if (errno != 0) { series_wrap = series[series_id].wrap; } } else { crm_config_warn("No value specified for cluster" " preference: %s", series[series_id].param); } seq = get_last_sequence(PE_STATE_DIR, series[series_id].name); crm_trace("Series %s: wrap=%d, seq=%d, pref=%s", series[series_id].name, series_wrap, seq, value); sched_data_set->input = NULL; reply = create_reply(msg, sched_data_set->graph); CRM_ASSERT(reply != NULL); if (is_repoke == FALSE) { free(filename); filename = generate_series_filename(PE_STATE_DIR, series[series_id].name, seq, TRUE); } crm_xml_add(reply, F_CRM_TGRAPH_INPUT, filename); crm_xml_add_int(reply, "graph-errors", was_processing_error); crm_xml_add_int(reply, "graph-warnings", was_processing_warning); crm_xml_add_int(reply, "config-errors", crm_config_error); crm_xml_add_int(reply, "config-warnings", crm_config_warning); if (crm_ipcs_send(sender, 0, reply, crm_ipc_server_event) == FALSE) { int graph_file_fd = 0; char *graph_file = NULL; umask(S_IWGRP | S_IWOTH | S_IROTH); graph_file = crm_strdup_printf("%s/pengine.graph.XXXXXX", PE_STATE_DIR); graph_file_fd = mkstemp(graph_file); crm_err("Couldn't send transition graph to peer, writing to %s instead", graph_file); crm_xml_add(reply, F_CRM_TGRAPH, graph_file); write_xml_fd(sched_data_set->graph, graph_file, graph_file_fd, FALSE); free(graph_file); free_xml(first_named_child(reply, F_CRM_DATA)); CRM_ASSERT(crm_ipcs_send(sender, 0, reply, crm_ipc_server_event)); } free_xml(reply); pe_reset_working_set(sched_data_set); pcmk__log_transition_summary(filename); if (is_repoke == FALSE && series_wrap != 0) { unlink(filename); crm_xml_add_ll(xml_data, "execution-date", (long long) execution_date); write_xml_file(xml_data, filename, TRUE); write_last_sequence(PE_STATE_DIR, series[series_id].name, seq + 1, series_wrap); } else { crm_trace("Not writing out %s: %d & %d", filename, is_repoke, series_wrap); } free_xml(converted); } return TRUE; } static int32_t pe_ipc_accept(qb_ipcs_connection_t * c, uid_t uid, gid_t gid) { crm_trace("Connection %p", c); if (crm_client_new(c, uid, gid) == NULL) { return -EIO; } return 0; } static void pe_ipc_created(qb_ipcs_connection_t * c) { crm_trace("Connection %p", c); } gboolean process_pe_message(xmlNode * msg, xmlNode * xml_data, crm_client_t * sender); static int32_t pe_ipc_dispatch(qb_ipcs_connection_t * qbc, void *data, size_t size) { uint32_t id = 0; uint32_t flags = 0; crm_client_t *c = crm_client_get(qbc); xmlNode *msg = crm_ipcs_recv(c, data, size, &id, &flags); crm_ipcs_send_ack(c, id, flags, "ack", __FUNCTION__, __LINE__); if (msg != NULL) { xmlNode *data_xml = get_message_xml(msg, F_CRM_DATA); process_pe_message(msg, data_xml, c); free_xml(msg); } return 0; } /* Error code means? */ static int32_t pe_ipc_closed(qb_ipcs_connection_t * c) { crm_client_t *client = crm_client_get(c); if (client == NULL) { return 0; } crm_trace("Connection %p", c); crm_client_destroy(client); return 0; } static void pe_ipc_destroy(qb_ipcs_connection_t * c) { crm_trace("Connection %p", c); pe_ipc_closed(c); } struct qb_ipcs_service_handlers ipc_callbacks = { .connection_accept = pe_ipc_accept, .connection_created = pe_ipc_created, .msg_process = pe_ipc_dispatch, .connection_closed = pe_ipc_closed, .connection_destroyed = pe_ipc_destroy }; /* *INDENT-OFF* */ static struct crm_option long_options[] = { /* Top-level Options */ {"help", 0, 0, '?', "\tThis text"}, {"verbose", 0, 0, 'V', "\tIncrease debug output"}, {0, 0, 0, 0} }; /* *INDENT-ON* */ int main(int argc, char **argv) { int flag; int index = 0; int argerr = 0; crm_log_preinit(NULL, argc, argv); crm_set_options(NULL, "[options]", long_options, "Daemon for calculating the cluster's response to events"); mainloop_add_signal(SIGTERM, pengine_shutdown); while (1) { flag = crm_get_option(argc, argv, &index); if (flag == -1) break; switch (flag) { case 'V': crm_bump_log_level(argc, argv); break; case 'h': /* Help message */ crm_help('?', CRM_EX_OK); break; default: ++argerr; break; } } if (argc - optind == 1 && safe_str_eq("metadata", argv[optind])) { pe_metadata(); return CRM_EX_OK; } if (optind > argc) { ++argerr; } if (argerr) { crm_help('?', CRM_EX_USAGE); } crm_log_init(NULL, LOG_INFO, TRUE, FALSE, argc, argv, FALSE); crm_notice("Starting Pacemaker scheduler"); if (pcmk__daemon_can_write(PE_STATE_DIR, NULL) == FALSE) { crm_err("Terminating due to bad permissions on " PE_STATE_DIR); fprintf(stderr, "ERROR: Bad permissions on " PE_STATE_DIR " (see logs for details)\n"); fflush(stderr); return CRM_EX_FATAL; } ipcs = mainloop_add_ipc_server(CRM_SYSTEM_PENGINE, QB_IPC_SHM, &ipc_callbacks); if (ipcs == NULL) { crm_err("Failed to create IPC server: shutting down and inhibiting respawn"); crm_exit(CRM_EX_FATAL); } /* Create the mainloop and run it... */ mainloop = g_main_loop_new(NULL, FALSE); crm_notice("Pacemaker scheduler successfully started and accepting connections"); g_main_loop_run(mainloop); pe_free_working_set(sched_data_set); crm_info("Exiting %s", crm_system_name); crm_exit(CRM_EX_OK); } void pengine_shutdown(int nsig) { mainloop_del_ipc_server(ipcs); pe_free_working_set(sched_data_set); crm_exit(CRM_EX_OK); } diff --git a/include/crm/pengine/internal.h b/include/crm/pengine/internal.h index e61afdd4b2..86c48156bb 100644 --- a/include/crm/pengine/internal.h +++ b/include/crm/pengine/internal.h @@ -1,401 +1,403 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU Lesser General Public License * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY. */ #ifndef PE_INTERNAL__H # define PE_INTERNAL__H # include # include # include # include # define pe_rsc_info(rsc, fmt, args...) crm_log_tag(LOG_INFO, rsc ? rsc->id : "", fmt, ##args) # define pe_rsc_debug(rsc, fmt, args...) crm_log_tag(LOG_DEBUG, rsc ? rsc->id : "", fmt, ##args) # define pe_rsc_trace(rsc, fmt, args...) crm_log_tag(LOG_TRACE, rsc ? rsc->id : "", fmt, ##args) # define pe_err(fmt...) { was_processing_error = TRUE; crm_config_error = TRUE; crm_err(fmt); } # define pe_warn(fmt...) { was_processing_warning = TRUE; crm_config_warning = TRUE; crm_warn(fmt); } # define pe_proc_err(fmt...) { was_processing_error = TRUE; crm_err(fmt); } # define pe_proc_warn(fmt...) { was_processing_warning = TRUE; crm_warn(fmt); } # define pe_set_action_bit(action, bit) action->flags = crm_set_bit(__FUNCTION__, __LINE__, action->uuid, action->flags, bit) # define pe_clear_action_bit(action, bit) action->flags = crm_clear_bit(__FUNCTION__, __LINE__, action->uuid, action->flags, bit) typedef struct pe__location_constraint_s { char *id; // Constraint XML ID pe_resource_t *rsc_lh; // Resource being located enum rsc_role_e role_filter; // Role to locate enum pe_discover_e discover_mode; // Resource discovery GListPtr node_list_rh; // List of pe_node_t* } pe__location_t; typedef struct pe__order_constraint_s { int id; enum pe_ordering type; void *lh_opaque; resource_t *lh_rsc; action_t *lh_action; char *lh_action_task; void *rh_opaque; resource_t *rh_rsc; action_t *rh_action; char *rh_action_task; } pe__ordering_t; typedef struct notify_data_s { GSList *keys; // Environment variable name/value pairs const char *action; action_t *pre; action_t *post; action_t *pre_done; action_t *post_done; GListPtr active; /* notify_entry_t* */ GListPtr inactive; /* notify_entry_t* */ GListPtr start; /* notify_entry_t* */ GListPtr stop; /* notify_entry_t* */ GListPtr demote; /* notify_entry_t* */ GListPtr promote; /* notify_entry_t* */ GListPtr master; /* notify_entry_t* */ GListPtr slave; /* notify_entry_t* */ GHashTable *allowed_nodes; } notify_data_t; bool pe_can_fence(pe_working_set_t *data_set, node_t *node); int merge_weights(int w1, int w2); void add_hash_param(GHashTable * hash, const char *name, const char *value); char *native_parameter(resource_t * rsc, node_t * node, gboolean create, const char *name, pe_working_set_t * data_set); pe_node_t *native_location(const pe_resource_t *rsc, GList **list, int current); void pe_metadata(void); void verify_pe_options(GHashTable * options); void common_update_score(resource_t * rsc, const char *id, int score); void native_add_running(resource_t * rsc, node_t * node, pe_working_set_t * data_set); gboolean native_unpack(resource_t * rsc, pe_working_set_t * data_set); gboolean group_unpack(resource_t * rsc, pe_working_set_t * data_set); gboolean clone_unpack(resource_t * rsc, pe_working_set_t * data_set); gboolean pe__unpack_bundle(pe_resource_t *rsc, pe_working_set_t *data_set); resource_t *native_find_rsc(resource_t *rsc, const char *id, const node_t *node, int flags); gboolean native_active(resource_t * rsc, gboolean all); gboolean group_active(resource_t * rsc, gboolean all); gboolean clone_active(resource_t * rsc, gboolean all); gboolean pe__bundle_active(pe_resource_t *rsc, gboolean all); void native_print(resource_t * rsc, const char *pre_text, long options, void *print_data); void group_print(resource_t * rsc, const char *pre_text, long options, void *print_data); void clone_print(resource_t * rsc, const char *pre_text, long options, void *print_data); void pe__print_bundle(pe_resource_t *rsc, const char *pre_text, long options, void *print_data); int pe__name_and_nvpairs_xml(pcmk__output_t *out, bool is_list, const char *tag_name , size_t pairs_count, ...); int pe__clone_xml(pcmk__output_t *out, va_list args); int pe__clone_html(pcmk__output_t *out, va_list args); int pe__clone_text(pcmk__output_t *out, va_list args); int pe__group_xml(pcmk__output_t *out, va_list args); int pe__group_html(pcmk__output_t *out, va_list args); int pe__group_text(pcmk__output_t *out, va_list args); int pe__bundle_xml(pcmk__output_t *out, va_list args); int pe__bundle_html(pcmk__output_t *out, va_list args); int pe__bundle_text(pcmk__output_t *out, va_list args); int pe__resource_xml(pcmk__output_t *out, va_list args); int pe__resource_html(pcmk__output_t *out, va_list args); int pe__resource_text(pcmk__output_t *out, va_list args); void native_free(resource_t * rsc); void group_free(resource_t * rsc); void clone_free(resource_t * rsc); void pe__free_bundle(pe_resource_t *rsc); enum rsc_role_e native_resource_state(const resource_t * rsc, gboolean current); enum rsc_role_e group_resource_state(const resource_t * rsc, gboolean current); enum rsc_role_e clone_resource_state(const resource_t * rsc, gboolean current); enum rsc_role_e pe__bundle_resource_state(const pe_resource_t *rsc, gboolean current); +void pe__count_common(pe_resource_t *rsc); +void pe__count_bundle(pe_resource_t *rsc); + gboolean common_unpack(xmlNode * xml_obj, resource_t ** rsc, resource_t * parent, pe_working_set_t * data_set); void common_free(resource_t * rsc); extern node_t *node_copy(const node_t *this_node); extern time_t get_effective_time(pe_working_set_t * data_set); /* Failure handling utilities (from failcounts.c) */ // bit flags for fail count handling options enum pe_fc_flags_e { pe_fc_default = 0x00, pe_fc_effective = 0x01, // don't count expired failures pe_fc_fillers = 0x02, // if container, include filler failures in count }; int pe_get_failcount(node_t *node, resource_t *rsc, time_t *last_failure, uint32_t flags, xmlNode *xml_op, pe_working_set_t *data_set); pe_action_t *pe__clear_failcount(pe_resource_t *rsc, pe_node_t *node, const char *reason, pe_working_set_t *data_set); /* Functions for finding/counting a resource's active nodes */ pe_node_t *pe__find_active_on(const pe_resource_t *rsc, unsigned int *count_all, unsigned int *count_clean); pe_node_t *pe__find_active_requires(const pe_resource_t *rsc, unsigned int *count); static inline pe_node_t * pe__current_node(const pe_resource_t *rsc) { return pe__find_active_on(rsc, NULL, NULL); } /* Binary like operators for lists of nodes */ extern void node_list_exclude(GHashTable * list, GListPtr list2, gboolean merge_scores); extern GListPtr node_list_dup(GListPtr list, gboolean reset, gboolean filter); extern GHashTable *node_hash_from_list(GListPtr list); static inline gpointer pe_hash_table_lookup(GHashTable * hash, gconstpointer key) { if (hash) { return g_hash_table_lookup(hash, key); } return NULL; } extern action_t *get_pseudo_op(const char *name, pe_working_set_t * data_set); extern gboolean order_actions(action_t * lh_action, action_t * rh_action, enum pe_ordering order); GHashTable *node_hash_dup(GHashTable * hash); /* Printing functions for debug */ extern void print_node(const char *pre_text, node_t * node, gboolean details); extern void print_str_str(gpointer key, gpointer value, gpointer user_data); extern void pe__output_node(node_t * node, gboolean details, pcmk__output_t *out); -extern void print_resource(int log_level, const char *pre_text, resource_t * rsc, gboolean details); -extern void pe__output_resource(int log_level, resource_t * rsc, gboolean details, pcmk__output_t *out); - extern void dump_node_scores_worker(int level, const char *file, const char *function, int line, resource_t * rsc, const char *comment, GHashTable * nodes); extern void dump_node_capacity(int level, const char *comment, node_t * node); extern void dump_rsc_utilization(int level, const char *comment, resource_t * rsc, node_t * node); # define dump_node_scores(level, rsc, text, nodes) do { \ dump_node_scores_worker(level, __FILE__, __FUNCTION__, __LINE__, rsc, text, nodes); \ } while(0) /* Sorting functions */ extern gint sort_rsc_priority(gconstpointer a, gconstpointer b); extern gint sort_rsc_index(gconstpointer a, gconstpointer b); extern xmlNode *find_rsc_op_entry(resource_t * rsc, const char *key); extern action_t *custom_action(resource_t * rsc, char *key, const char *task, node_t * on_node, gboolean optional, gboolean foo, pe_working_set_t * data_set); # define delete_key(rsc) generate_op_key(rsc->id, CRMD_ACTION_DELETE, 0) # define delete_action(rsc, node, optional) custom_action( \ rsc, delete_key(rsc), CRMD_ACTION_DELETE, node, \ optional, TRUE, data_set); # define stopped_key(rsc) generate_op_key(rsc->id, CRMD_ACTION_STOPPED, 0) # define stopped_action(rsc, node, optional) custom_action( \ rsc, stopped_key(rsc), CRMD_ACTION_STOPPED, node, \ optional, TRUE, data_set); # define stop_key(rsc) generate_op_key(rsc->id, CRMD_ACTION_STOP, 0) # define stop_action(rsc, node, optional) custom_action( \ rsc, stop_key(rsc), CRMD_ACTION_STOP, node, \ optional, TRUE, data_set); # define reload_key(rsc) generate_op_key(rsc->id, CRMD_ACTION_RELOAD, 0) # define start_key(rsc) generate_op_key(rsc->id, CRMD_ACTION_START, 0) # define start_action(rsc, node, optional) custom_action( \ rsc, start_key(rsc), CRMD_ACTION_START, node, \ optional, TRUE, data_set) # define started_key(rsc) generate_op_key(rsc->id, CRMD_ACTION_STARTED, 0) # define started_action(rsc, node, optional) custom_action( \ rsc, started_key(rsc), CRMD_ACTION_STARTED, node, \ optional, TRUE, data_set) # define promote_key(rsc) generate_op_key(rsc->id, CRMD_ACTION_PROMOTE, 0) # define promote_action(rsc, node, optional) custom_action( \ rsc, promote_key(rsc), CRMD_ACTION_PROMOTE, node, \ optional, TRUE, data_set) # define promoted_key(rsc) generate_op_key(rsc->id, CRMD_ACTION_PROMOTED, 0) # define promoted_action(rsc, node, optional) custom_action( \ rsc, promoted_key(rsc), CRMD_ACTION_PROMOTED, node, \ optional, TRUE, data_set) # define demote_key(rsc) generate_op_key(rsc->id, CRMD_ACTION_DEMOTE, 0) # define demote_action(rsc, node, optional) custom_action( \ rsc, demote_key(rsc), CRMD_ACTION_DEMOTE, node, \ optional, TRUE, data_set) # define demoted_key(rsc) generate_op_key(rsc->id, CRMD_ACTION_DEMOTED, 0) # define demoted_action(rsc, node, optional) custom_action( \ rsc, demoted_key(rsc), CRMD_ACTION_DEMOTED, node, \ optional, TRUE, data_set) extern int pe_get_configured_timeout(resource_t *rsc, const char *action, pe_working_set_t *data_set); extern action_t *find_first_action(GListPtr input, const char *uuid, const char *task, node_t * on_node); extern enum action_tasks get_complex_task(resource_t * rsc, const char *name, gboolean allow_non_atomic); extern GListPtr find_actions(GListPtr input, const char *key, const node_t *on_node); GList *find_actions_exact(GList *input, const char *key, const pe_node_t *on_node); extern GListPtr find_recurring_actions(GListPtr input, node_t * not_on_node); GList *pe__resource_actions(const pe_resource_t *rsc, const pe_node_t *node, const char *task, bool require_node); extern void pe_free_action(action_t * action); extern void resource_location(resource_t * rsc, node_t * node, int score, const char *tag, pe_working_set_t * data_set); extern gint sort_op_by_callid(gconstpointer a, gconstpointer b); extern gboolean get_target_role(resource_t * rsc, enum rsc_role_e *role); extern resource_t *find_clone_instance(resource_t * rsc, const char *sub_id, pe_working_set_t * data_set); extern void destroy_ticket(gpointer data); extern ticket_t *ticket_new(const char *ticket_id, pe_working_set_t * data_set); // Resources for manipulating resource names const char *pe_base_name_end(const char *id); char *clone_strip(const char *last_rsc_id); char *clone_zero(const char *last_rsc_id); static inline bool pe_base_name_eq(resource_t *rsc, const char *id) { if (id && rsc && rsc->id) { // Number of characters in rsc->id before any clone suffix size_t base_len = pe_base_name_end(rsc->id) - rsc->id + 1; return (strlen(id) == base_len) && !strncmp(id, rsc->id, base_len); } return FALSE; } int pe__target_rc_from_xml(xmlNode *xml_op); gint sort_node_uname(gconstpointer a, gconstpointer b); bool is_set_recursive(resource_t * rsc, long long flag, bool any); enum rsc_digest_cmp_val { /*! Digests are the same */ RSC_DIGEST_MATCH = 0, /*! Params that require a restart changed */ RSC_DIGEST_RESTART, /*! Some parameter changed. */ RSC_DIGEST_ALL, /*! rsc op didn't have a digest associated with it, so * it is unknown if parameters changed or not. */ RSC_DIGEST_UNKNOWN, }; typedef struct op_digest_cache_s { enum rsc_digest_cmp_val rc; xmlNode *params_all; xmlNode *params_secure; xmlNode *params_restart; char *digest_all_calc; char *digest_secure_calc; char *digest_restart_calc; } op_digest_cache_t; op_digest_cache_t *rsc_action_digest_cmp(resource_t * rsc, xmlNode * xml_op, node_t * node, pe_working_set_t * data_set); action_t *pe_fence_op(node_t * node, const char *op, bool optional, const char *reason, pe_working_set_t * data_set); void trigger_unfencing( resource_t * rsc, node_t *node, const char *reason, action_t *dependency, pe_working_set_t * data_set); void pe_action_set_reason(pe_action_t *action, const char *reason, bool overwrite); void pe_action_set_flag_reason(const char *function, long line, pe_action_t *action, pe_action_t *reason, const char *text, enum pe_action_flags flags, bool overwrite); #define pe_action_required(action, reason, text) pe_action_set_flag_reason(__FUNCTION__, __LINE__, action, reason, text, pe_action_optional, FALSE) #define pe_action_implies(action, reason, flag) pe_action_set_flag_reason(__FUNCTION__, __LINE__, action, reason, NULL, flag, FALSE) void set_bit_recursive(resource_t * rsc, unsigned long long flag); void clear_bit_recursive(resource_t * rsc, unsigned long long flag); gboolean add_tag_ref(GHashTable * tags, const char * tag_name, const char * obj_ref); void print_rscs_brief(GListPtr rsc_list, const char * pre_text, long options, void * print_data, gboolean print_all); void pe__rscs_brief_output(pcmk__output_t *out, GListPtr rsc_list, long options, gboolean print_all); void pe_fence_node(pe_working_set_t * data_set, node_t * node, const char *reason); node_t *pe_create_node(const char *id, const char *uname, const char *type, const char *score, pe_working_set_t * data_set); bool remote_id_conflict(const char *remote_name, pe_working_set_t *data); void common_print(resource_t * rsc, const char *pre_text, const char *name, node_t *node, long options, void *print_data); void pe__common_output_text(pcmk__output_t *out, resource_t * rsc, const char *name, node_t *node, long options); void pe__common_output_html(pcmk__output_t *out, resource_t * rsc, const char *name, node_t *node, long options); pe_resource_t *pe__find_bundle_replica(const pe_resource_t *bundle, const pe_node_t *node); bool pe__bundle_needs_remote_name(pe_resource_t *rsc); const char *pe__add_bundle_remote_name(pe_resource_t *rsc, xmlNode *xml, const char *field); const char *pe_node_attribute_calculated(const pe_node_t *node, const char *name, const resource_t *rsc); const char *pe_node_attribute_raw(pe_node_t *node, const char *name); bool pe__is_universal_clone(pe_resource_t *rsc, pe_working_set_t *data_set); void pe__add_param_check(xmlNode *rsc_op, pe_resource_t *rsc, pe_node_t *node, enum pe_check_parameters, pe_working_set_t *data_set); void pe__foreach_param_check(pe_working_set_t *data_set, void (*cb)(pe_resource_t*, pe_node_t*, xmlNode*, enum pe_check_parameters, pe_working_set_t*)); void pe__free_param_checks(pe_working_set_t *data_set); bool pe__shutdown_requested(pe_node_t *node); void pe__update_recheck_time(time_t recheck, pe_working_set_t *data_set); #define BOOL2STR(x) ((x) ? "true" : "false") /*! * \internal * \brief Register xml formatting message functions. */ void pe__register_messages(pcmk__output_t *out); void pe__unpack_dataset_nvpairs(xmlNode *xml_obj, const char *set_name, GHashTable *node_hash, GHashTable *hash, const char *always_first, gboolean overwrite, pe_working_set_t *data_set); +bool pe__resource_is_disabled(pe_resource_t *rsc); + #endif diff --git a/include/crm/pengine/pe_types.h b/include/crm/pengine/pe_types.h index 3569adb537..23e1c46b9c 100644 --- a/include/crm/pengine/pe_types.h +++ b/include/crm/pengine/pe_types.h @@ -1,503 +1,508 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU Lesser General Public License * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY. */ #ifndef PE_TYPES__H # define PE_TYPES__H #ifdef __cplusplus extern "C" { #endif /*! * \file * \brief Data types for cluster status * \ingroup pengine */ # include // bool # include // time_t # include // gboolean, guint, GList, GHashTable # include // GListPtr # include # include typedef struct pe_node_s pe_node_t; typedef struct pe_action_s pe_action_t; typedef struct pe_resource_s pe_resource_t; typedef struct pe_working_set_s pe_working_set_t; enum pe_obj_types { pe_unknown = -1, pe_native = 0, pe_group = 1, pe_clone = 2, pe_container = 3, }; typedef struct resource_object_functions_s { gboolean (*unpack) (pe_resource_t*, pe_working_set_t*); pe_resource_t *(*find_rsc) (pe_resource_t *parent, const char *search, const pe_node_t *node, int flags); /* parameter result must be free'd */ char *(*parameter) (pe_resource_t*, pe_node_t*, gboolean, const char*, pe_working_set_t*); void (*print) (pe_resource_t*, const char*, long, void*); gboolean (*active) (pe_resource_t*, gboolean); enum rsc_role_e (*state) (const pe_resource_t*, gboolean); pe_node_t *(*location) (const pe_resource_t*, GList**, int); void (*free) (pe_resource_t*); + void (*count) (pe_resource_t*); } resource_object_functions_t; typedef struct resource_alloc_functions_s resource_alloc_functions_t; enum pe_quorum_policy { no_quorum_freeze, no_quorum_stop, no_quorum_ignore, no_quorum_suicide }; enum node_type { node_ping, node_member, node_remote }; //! \deprecated will be removed in a future release enum pe_restart { pe_restart_restart, pe_restart_ignore }; //! Determine behavior of pe_find_resource_with_flags() enum pe_find { pe_find_renamed = 0x001, //!< match resource ID or LRM history ID pe_find_anon = 0x002, //!< match base name of anonymous clone instances pe_find_clone = 0x004, //!< match only clone instances pe_find_current = 0x008, //!< match resource active on specified node pe_find_inactive = 0x010, //!< match resource not running anywhere pe_find_any = 0x020, //!< match base name of any clone instance }; # define pe_flag_have_quorum 0x00000001ULL # define pe_flag_symmetric_cluster 0x00000002ULL # define pe_flag_maintenance_mode 0x00000008ULL # define pe_flag_stonith_enabled 0x00000010ULL # define pe_flag_have_stonith_resource 0x00000020ULL # define pe_flag_enable_unfencing 0x00000040ULL # define pe_flag_concurrent_fencing 0x00000080ULL # define pe_flag_stop_rsc_orphans 0x00000100ULL # define pe_flag_stop_action_orphans 0x00000200ULL # define pe_flag_stop_everything 0x00000400ULL # define pe_flag_start_failure_fatal 0x00001000ULL # define pe_flag_remove_after_stop 0x00002000ULL # define pe_flag_startup_fencing 0x00004000ULL # define pe_flag_startup_probes 0x00010000ULL # define pe_flag_have_status 0x00020000ULL # define pe_flag_have_remote_nodes 0x00040000ULL # define pe_flag_quick_location 0x00100000ULL # define pe_flag_sanitized 0x00200000ULL # define pe_flag_stdout 0x00400000ULL +//! Don't count total, disabled and blocked resource instances +# define pe_flag_no_counts 0x00800000ULL + struct pe_working_set_s { xmlNode *input; crm_time_t *now; /* options extracted from the input */ char *dc_uuid; pe_node_t *dc_node; const char *stonith_action; const char *placement_strategy; unsigned long long flags; int stonith_timeout; enum pe_quorum_policy no_quorum_policy; GHashTable *config_hash; GHashTable *tickets; // Actions for which there can be only one (e.g. fence nodeX) GHashTable *singletons; GListPtr nodes; GListPtr resources; GListPtr placement_constraints; GListPtr ordering_constraints; GListPtr colocation_constraints; GListPtr ticket_constraints; GListPtr actions; xmlNode *failed; xmlNode *op_defaults; xmlNode *rsc_defaults; /* stats */ int num_synapse; int max_valid_nodes; int order_id; int action_id; /* final output */ xmlNode *graph; GHashTable *template_rsc_sets; const char *localhost; GHashTable *tags; int blocked_resources; int disabled_resources; GList *param_check; // History entries that need to be checked GList *stop_needed; // Containers that need stop actions time_t recheck_by; // Hint to controller to re-run scheduler by this time + int ninstances; // Total number of resource instances }; enum pe_check_parameters { /* Clear fail count if parameters changed for un-expired start or monitor * last_failure. */ pe_check_last_failure, /* Clear fail count if parameters changed for start, monitor, promote, or * migrate_from actions for active resources. */ pe_check_active, }; struct pe_node_shared_s { const char *id; const char *uname; enum node_type type; /* @TODO convert these flags into a bitfield */ gboolean online; gboolean standby; gboolean standby_onfail; gboolean pending; gboolean unclean; gboolean unseen; gboolean shutdown; gboolean expected_up; gboolean is_dc; gboolean maintenance; gboolean rsc_discovery_enabled; gboolean remote_requires_reset; gboolean remote_was_fenced; gboolean remote_maintenance; /* what the remote-rsc is thinking */ gboolean unpacked; int num_resources; pe_resource_t *remote_rsc; GListPtr running_rsc; /* pe_resource_t* */ GListPtr allocated_rsc; /* pe_resource_t* */ GHashTable *attrs; /* char* => char* */ GHashTable *utilization; GHashTable *digest_cache; //!< cache of calculated resource digests }; struct pe_node_s { int weight; gboolean fixed; int count; struct pe_node_shared_s *details; int rsc_discover_mode; }; # define pe_rsc_orphan 0x00000001ULL # define pe_rsc_managed 0x00000002ULL # define pe_rsc_block 0x00000004ULL # define pe_rsc_orphan_container_filler 0x00000008ULL # define pe_rsc_notify 0x00000010ULL # define pe_rsc_unique 0x00000020ULL # define pe_rsc_fence_device 0x00000040ULL # define pe_rsc_promotable 0x00000080ULL # define pe_rsc_provisional 0x00000100ULL # define pe_rsc_allocating 0x00000200ULL # define pe_rsc_merging 0x00000400ULL # define pe_rsc_reload 0x00002000ULL # define pe_rsc_allow_remote_remotes 0x00004000ULL # define pe_rsc_failed 0x00010000ULL # define pe_rsc_runnable 0x00040000ULL # define pe_rsc_start_pending 0x00080000ULL # define pe_rsc_starting 0x00100000ULL # define pe_rsc_stopping 0x00200000ULL # define pe_rsc_allow_migrate 0x00800000ULL # define pe_rsc_failure_ignored 0x01000000ULL # define pe_rsc_maintenance 0x04000000ULL # define pe_rsc_is_container 0x08000000ULL # define pe_rsc_needs_quorum 0x10000000ULL # define pe_rsc_needs_fencing 0x20000000ULL # define pe_rsc_needs_unfencing 0x40000000ULL enum pe_graph_flags { pe_graph_none = 0x00000, pe_graph_updated_first = 0x00001, pe_graph_updated_then = 0x00002, pe_graph_disable = 0x00004, }; /* *INDENT-OFF* */ enum pe_action_flags { pe_action_pseudo = 0x00001, pe_action_runnable = 0x00002, pe_action_optional = 0x00004, pe_action_print_always = 0x00008, pe_action_have_node_attrs = 0x00010, pe_action_implied_by_stonith = 0x00040, pe_action_migrate_runnable = 0x00080, pe_action_dumped = 0x00100, pe_action_processed = 0x00200, pe_action_clear = 0x00400, pe_action_dangle = 0x00800, /* This action requires one or more of its dependencies to be runnable. * We use this to clear the runnable flag before checking dependencies. */ pe_action_requires_any = 0x01000, pe_action_reschedule = 0x02000, pe_action_tracking = 0x04000, pe_action_dedup = 0x08000, //! Internal state tracking when creating graph }; /* *INDENT-ON* */ struct pe_resource_s { char *id; char *clone_name; xmlNode *xml; xmlNode *orig_xml; xmlNode *ops_xml; pe_working_set_t *cluster; pe_resource_t *parent; enum pe_obj_types variant; void *variant_opaque; resource_object_functions_t *fns; resource_alloc_functions_t *cmds; enum rsc_recovery_type recovery_type; // @TODO only pe_restart_restart is of interest, so merge into flags enum pe_restart restart_type; //!< \deprecated will be removed in future release int priority; int stickiness; int sort_index; int failure_timeout; int migration_threshold; guint remote_reconnect_ms; char *pending_task; unsigned long long flags; // @TODO merge these into flags gboolean is_remote_node; gboolean exclusive_discover; //!@{ //! This field should be treated as internal to Pacemaker GListPtr rsc_cons_lhs; // List of rsc_colocation_t* GListPtr rsc_cons; // List of rsc_colocation_t* GListPtr rsc_location; // List of pe__location_t* GListPtr actions; // List of pe_action_t* GListPtr rsc_tickets; // List of rsc_ticket* //!@} pe_node_t *allocated_to; pe_node_t *partial_migration_target; pe_node_t *partial_migration_source; GListPtr running_on; /* pe_node_t* */ GHashTable *known_on; /* pe_node_t* */ GHashTable *allowed_nodes; /* pe_node_t* */ enum rsc_role_e role; enum rsc_role_e next_role; GHashTable *meta; GHashTable *parameters; GHashTable *utilization; GListPtr children; /* pe_resource_t* */ GListPtr dangling_migrations; /* pe_node_t* */ pe_resource_t *container; GListPtr fillers; pe_node_t *pending_node; // Node on which pending_task is happening #if ENABLE_VERSIONED_ATTRS xmlNode *versioned_parameters; #endif }; #if ENABLE_VERSIONED_ATTRS // Used as action->action_details if action->rsc is not NULL typedef struct pe_rsc_action_details_s { xmlNode *versioned_parameters; xmlNode *versioned_meta; } pe_rsc_action_details_t; #endif struct pe_action_s { int id; int priority; pe_resource_t *rsc; pe_node_t *node; xmlNode *op_entry; char *task; char *uuid; char *cancel_task; char *reason; enum pe_action_flags flags; enum rsc_start_requirement needs; enum action_fail_response on_fail; enum rsc_role_e fail_role; GHashTable *meta; GHashTable *extra; /* * These two varables are associated with the constraint logic * that involves first having one or more actions runnable before * then allowing this action to execute. * * These varables are used with features such as 'clone-min' which * requires at minimum X number of cloned instances to be running * before an order dependency can run. Another option that uses * this is 'require-all=false' in ordering constrants. This option * says "only require one instance of a resource to start before * allowing dependencies to start" -- basically, require-all=false is * the same as clone-min=1. */ /* current number of known runnable actions in the before list. */ int runnable_before; /* the number of "before" runnable actions required for this action * to be considered runnable */ int required_runnable_before; GListPtr actions_before; /* pe_action_wrapper_t* */ GListPtr actions_after; /* pe_action_wrapper_t* */ /* Some of the above fields could be moved to the details, * except for API backward compatibility. */ void *action_details; // varies by type of action }; typedef struct pe_ticket_s { char *id; gboolean granted; time_t last_granted; gboolean standby; GHashTable *state; } pe_ticket_t; typedef struct pe_tag_s { char *id; GListPtr refs; } pe_tag_t; //! Internal tracking for transition graph creation enum pe_link_state { pe_link_not_dumped, //! Internal tracking for transition graph creation pe_link_dumped, //! Internal tracking for transition graph creation pe_link_dup, //! \deprecated No longer used by Pacemaker }; enum pe_discover_e { pe_discover_always = 0, pe_discover_never, pe_discover_exclusive, }; /* *INDENT-OFF* */ enum pe_ordering { pe_order_none = 0x0, /* deleted */ pe_order_optional = 0x1, /* pure ordering, nothing implied */ pe_order_apply_first_non_migratable = 0x2, /* Only apply this constraint's ordering if first is not migratable. */ pe_order_implies_first = 0x10, /* If 'then' is required, ensure 'first' is too */ pe_order_implies_then = 0x20, /* If 'first' is required, ensure 'then' is too */ pe_order_implies_first_master = 0x40, /* Imply 'first' is required when 'then' is required and then's rsc holds Master role. */ /* first requires then to be both runnable and migrate runnable. */ pe_order_implies_first_migratable = 0x80, pe_order_runnable_left = 0x100, /* 'then' requires 'first' to be runnable */ pe_order_pseudo_left = 0x200, /* 'then' can only be pseudo if 'first' is runnable */ pe_order_implies_then_on_node = 0x400, /* If 'first' is required on 'nodeX', * ensure instances of 'then' on 'nodeX' are too. * Only really useful if 'then' is a clone and 'first' is not */ pe_order_probe = 0x800, /* If 'first->rsc' is * - running but about to stop, ignore the constraint * - otherwise, behave as runnable_left */ pe_order_restart = 0x1000, /* 'then' is runnable if 'first' is optional or runnable */ pe_order_stonith_stop = 0x2000, /* only applies if the action is non-pseudo */ pe_order_serialize_only = 0x4000, /* serialize */ pe_order_same_node = 0x8000, /* applies only if 'first' and 'then' are on same node */ pe_order_implies_first_printed = 0x10000, /* Like ..implies_first but only ensures 'first' is printed, not mandatory */ pe_order_implies_then_printed = 0x20000, /* Like ..implies_then but only ensures 'then' is printed, not mandatory */ pe_order_asymmetrical = 0x100000, /* Indicates asymmetrical one way ordering constraint. */ pe_order_load = 0x200000, /* Only relevant if... */ pe_order_one_or_more = 0x400000, /* 'then' is runnable only if one or more of its dependencies are too */ pe_order_anti_colocation = 0x800000, pe_order_preserve = 0x1000000, /* Hack for breaking user ordering constraints with container resources */ pe_order_then_cancels_first = 0x2000000, // if 'then' becomes required, 'first' becomes optional pe_order_trace = 0x4000000, /* test marker */ }; /* *INDENT-ON* */ typedef struct pe_action_wrapper_s { enum pe_ordering type; enum pe_link_state state; pe_action_t *action; } pe_action_wrapper_t; // Deprecated type aliases typedef struct pe_action_s action_t; //!< \deprecated Use pe_action_t instead typedef struct pe_action_wrapper_s action_wrapper_t; //!< \deprecated Use pe_action_wrapper_t instead typedef struct pe_node_s node_t; //!< \deprecated Use pe_node_t instead typedef struct pe_resource_s resource_t; //!< \deprecated Use pe_resource_t instead typedef struct pe_tag_s tag_t; //!< \deprecated Use pe_tag_t instead typedef struct pe_ticket_s ticket_t; //!< \deprecated Use pe_ticket_t instead typedef enum pe_quorum_policy no_quorum_policy_t; //!< \deprecated Use enum pe_quorum_policy instead #ifdef __cplusplus } #endif #endif // PE_TYPES__H diff --git a/lib/pacemaker/pcmk_sched_group.c b/lib/pacemaker/pcmk_sched_group.c index ea27029fa4..e12774f033 100644 --- a/lib/pacemaker/pcmk_sched_group.c +++ b/lib/pacemaker/pcmk_sched_group.c @@ -1,517 +1,517 @@ /* * Copyright 2004-2018 Andrew Beekhof * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #define VARIANT_GROUP 1 #include node_t * group_color(resource_t * rsc, node_t * prefer, pe_working_set_t * data_set) { node_t *node = NULL; node_t *group_node = NULL; GListPtr gIter = NULL; group_variant_data_t *group_data = NULL; get_group_variant_data(group_data, rsc); if (is_not_set(rsc->flags, pe_rsc_provisional)) { return rsc->allocated_to; } pe_rsc_trace(rsc, "Processing %s", rsc->id); if (is_set(rsc->flags, pe_rsc_allocating)) { pe_rsc_debug(rsc, "Dependency loop detected involving %s", rsc->id); return NULL; } if (group_data->first_child == NULL) { /* nothign to allocate */ clear_bit(rsc->flags, pe_rsc_provisional); return NULL; } set_bit(rsc->flags, pe_rsc_allocating); rsc->role = group_data->first_child->role; group_data->first_child->rsc_cons = g_list_concat(group_data->first_child->rsc_cons, rsc->rsc_cons); rsc->rsc_cons = NULL; group_data->last_child->rsc_cons_lhs = g_list_concat(group_data->last_child->rsc_cons_lhs, rsc->rsc_cons_lhs); rsc->rsc_cons_lhs = NULL; dump_node_scores(show_scores ? 0 : scores_log_level, rsc, __FUNCTION__, rsc->allowed_nodes); gIter = rsc->children; for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; node = child_rsc->cmds->allocate(child_rsc, prefer, data_set); if (group_node == NULL) { group_node = node; } } rsc->next_role = group_data->first_child->next_role; clear_bit(rsc->flags, pe_rsc_allocating); clear_bit(rsc->flags, pe_rsc_provisional); if (group_data->colocated) { return group_node; } return NULL; } void group_update_pseudo_status(resource_t * parent, resource_t * child); void group_create_actions(resource_t * rsc, pe_working_set_t * data_set) { action_t *op = NULL; const char *value = NULL; GListPtr gIter = rsc->children; pe_rsc_trace(rsc, "Creating actions for %s", rsc->id); for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; child_rsc->cmds->create_actions(child_rsc, data_set); group_update_pseudo_status(rsc, child_rsc); } op = start_action(rsc, NULL, TRUE /* !group_data->child_starting */ ); set_bit(op->flags, pe_action_pseudo | pe_action_runnable); op = custom_action(rsc, started_key(rsc), RSC_STARTED, NULL, TRUE /* !group_data->child_starting */ , TRUE, data_set); set_bit(op->flags, pe_action_pseudo | pe_action_runnable); op = stop_action(rsc, NULL, TRUE /* !group_data->child_stopping */ ); set_bit(op->flags, pe_action_pseudo | pe_action_runnable); op = custom_action(rsc, stopped_key(rsc), RSC_STOPPED, NULL, TRUE /* !group_data->child_stopping */ , TRUE, data_set); set_bit(op->flags, pe_action_pseudo | pe_action_runnable); value = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_PROMOTABLE); if (crm_is_true(value)) { op = custom_action(rsc, demote_key(rsc), RSC_DEMOTE, NULL, TRUE, TRUE, data_set); set_bit(op->flags, pe_action_pseudo); set_bit(op->flags, pe_action_runnable); op = custom_action(rsc, demoted_key(rsc), RSC_DEMOTED, NULL, TRUE, TRUE, data_set); set_bit(op->flags, pe_action_pseudo); set_bit(op->flags, pe_action_runnable); op = custom_action(rsc, promote_key(rsc), RSC_PROMOTE, NULL, TRUE, TRUE, data_set); set_bit(op->flags, pe_action_pseudo); set_bit(op->flags, pe_action_runnable); op = custom_action(rsc, promoted_key(rsc), RSC_PROMOTED, NULL, TRUE, TRUE, data_set); set_bit(op->flags, pe_action_pseudo); set_bit(op->flags, pe_action_runnable); } } void group_update_pseudo_status(resource_t * parent, resource_t * child) { GListPtr gIter = child->actions; group_variant_data_t *group_data = NULL; get_group_variant_data(group_data, parent); if (group_data->ordered == FALSE) { /* If this group is not ordered, then leave the meta-actions as optional */ return; } if (group_data->child_stopping && group_data->child_starting) { return; } for (; gIter != NULL; gIter = gIter->next) { action_t *action = (action_t *) gIter->data; if (is_set(action->flags, pe_action_optional)) { continue; } if (safe_str_eq(RSC_STOP, action->task) && is_set(action->flags, pe_action_runnable)) { group_data->child_stopping = TRUE; pe_rsc_trace(action->rsc, "Based on %s the group is stopping", action->uuid); } else if (safe_str_eq(RSC_START, action->task) && is_set(action->flags, pe_action_runnable)) { group_data->child_starting = TRUE; pe_rsc_trace(action->rsc, "Based on %s the group is starting", action->uuid); } } } void group_internal_constraints(resource_t * rsc, pe_working_set_t * data_set) { GListPtr gIter = rsc->children; resource_t *last_rsc = NULL; resource_t *last_active = NULL; resource_t *top = uber_parent(rsc); group_variant_data_t *group_data = NULL; get_group_variant_data(group_data, rsc); new_rsc_order(rsc, RSC_STOPPED, rsc, RSC_START, pe_order_optional, data_set); new_rsc_order(rsc, RSC_START, rsc, RSC_STARTED, pe_order_runnable_left, data_set); new_rsc_order(rsc, RSC_STOP, rsc, RSC_STOPPED, pe_order_runnable_left, data_set); for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; int stop = pe_order_none; int stopped = pe_order_implies_then_printed; int start = pe_order_implies_then | pe_order_runnable_left; int started = pe_order_runnable_left | pe_order_implies_then | pe_order_implies_then_printed; child_rsc->cmds->internal_constraints(child_rsc, data_set); if (last_rsc == NULL) { if (group_data->ordered) { stop |= pe_order_optional; stopped = pe_order_implies_then; } } else if (group_data->colocated) { rsc_colocation_new("group:internal_colocation", NULL, INFINITY, child_rsc, last_rsc, NULL, NULL, data_set); } if (is_set(top->flags, pe_rsc_promotable)) { new_rsc_order(rsc, RSC_DEMOTE, child_rsc, RSC_DEMOTE, stop | pe_order_implies_first_printed, data_set); new_rsc_order(child_rsc, RSC_DEMOTE, rsc, RSC_DEMOTED, stopped, data_set); new_rsc_order(child_rsc, RSC_PROMOTE, rsc, RSC_PROMOTED, started, data_set); new_rsc_order(rsc, RSC_PROMOTE, child_rsc, RSC_PROMOTE, pe_order_implies_first_printed, data_set); } order_start_start(rsc, child_rsc, pe_order_implies_first_printed); order_stop_stop(rsc, child_rsc, stop | pe_order_implies_first_printed); new_rsc_order(child_rsc, RSC_STOP, rsc, RSC_STOPPED, stopped, data_set); new_rsc_order(child_rsc, RSC_START, rsc, RSC_STARTED, started, data_set); if (group_data->ordered == FALSE) { order_start_start(rsc, child_rsc, start | pe_order_implies_first_printed); if (is_set(top->flags, pe_rsc_promotable)) { new_rsc_order(rsc, RSC_PROMOTE, child_rsc, RSC_PROMOTE, start | pe_order_implies_first_printed, data_set); } } else if (last_rsc != NULL) { child_rsc->restart_type = pe_restart_restart; order_start_start(last_rsc, child_rsc, start); order_stop_stop(child_rsc, last_rsc, pe_order_optional | pe_order_restart); if (is_set(top->flags, pe_rsc_promotable)) { new_rsc_order(last_rsc, RSC_PROMOTE, child_rsc, RSC_PROMOTE, start, data_set); new_rsc_order(child_rsc, RSC_DEMOTE, last_rsc, RSC_DEMOTE, pe_order_optional, data_set); } } else { /* If anyone in the group is starting, then * pe_order_implies_then will cause _everyone_ in the group * to be sent a start action * But this is safe since starting something that is already * started is required to be "safe" */ int flags = pe_order_none; order_start_start(rsc, child_rsc, flags); if (is_set(top->flags, pe_rsc_promotable)) { new_rsc_order(rsc, RSC_PROMOTE, child_rsc, RSC_PROMOTE, flags, data_set); } } /* Look for partially active groups * Make sure they still shut down in sequence */ if (child_rsc->running_on) { if (group_data->ordered && last_rsc && last_rsc->running_on == NULL && last_active && last_active->running_on) { order_stop_stop(child_rsc, last_active, pe_order_optional); } last_active = child_rsc; } last_rsc = child_rsc; } if (group_data->ordered && last_rsc != NULL) { int stop_stop_flags = pe_order_implies_then; int stop_stopped_flags = pe_order_optional; order_stop_stop(rsc, last_rsc, stop_stop_flags); new_rsc_order(last_rsc, RSC_STOP, rsc, RSC_STOPPED, stop_stopped_flags, data_set); if (is_set(top->flags, pe_rsc_promotable)) { new_rsc_order(rsc, RSC_DEMOTE, last_rsc, RSC_DEMOTE, stop_stop_flags, data_set); new_rsc_order(last_rsc, RSC_DEMOTE, rsc, RSC_DEMOTED, stop_stopped_flags, data_set); } } } void group_rsc_colocation_lh(pe_resource_t *rsc_lh, pe_resource_t *rsc_rh, rsc_colocation_t *constraint, pe_working_set_t *data_set) { GListPtr gIter = NULL; group_variant_data_t *group_data = NULL; if (rsc_lh == NULL) { pe_err("rsc_lh was NULL for %s", constraint->id); return; } else if (rsc_rh == NULL) { pe_err("rsc_rh was NULL for %s", constraint->id); return; } gIter = rsc_lh->children; pe_rsc_trace(rsc_lh, "Processing constraints from %s", rsc_lh->id); get_group_variant_data(group_data, rsc_lh); if (group_data->colocated) { group_data->first_child->cmds->rsc_colocation_lh(group_data->first_child, rsc_rh, constraint, data_set); return; } else if (constraint->score >= INFINITY) { crm_config_err("%s: Cannot perform mandatory colocation" " between non-colocated group and %s", rsc_lh->id, rsc_rh->id); return; } for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; child_rsc->cmds->rsc_colocation_lh(child_rsc, rsc_rh, constraint, data_set); } } void group_rsc_colocation_rh(pe_resource_t *rsc_lh, pe_resource_t *rsc_rh, rsc_colocation_t *constraint, pe_working_set_t *data_set) { GListPtr gIter = rsc_rh->children; group_variant_data_t *group_data = NULL; get_group_variant_data(group_data, rsc_rh); CRM_CHECK(rsc_lh->variant == pe_native, return); - pe_rsc_trace(rsc_rh, "Processing RH of constraint %s", constraint->id); - print_resource(LOG_TRACE, "LHS", rsc_lh, TRUE); + pe_rsc_trace(rsc_rh, "Processing RH %s of constraint %s (LH is %s)", + rsc_rh->id, constraint->id, rsc_lh->id); if (is_set(rsc_rh->flags, pe_rsc_provisional)) { return; } else if (group_data->colocated && group_data->first_child) { if (constraint->score >= INFINITY) { /* Ensure RHS is _fully_ up before can start LHS */ group_data->last_child->cmds->rsc_colocation_rh(rsc_lh, group_data->last_child, constraint, data_set); } else { /* A partially active RHS is fine */ group_data->first_child->cmds->rsc_colocation_rh(rsc_lh, group_data->first_child, constraint, data_set); } return; } else if (constraint->score >= INFINITY) { crm_config_err("%s: Cannot perform mandatory colocation with" " non-colocated group: %s", rsc_lh->id, rsc_rh->id); return; } for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; child_rsc->cmds->rsc_colocation_rh(rsc_lh, child_rsc, constraint, data_set); } } enum pe_action_flags group_action_flags(action_t * action, node_t * node) { GListPtr gIter = NULL; enum pe_action_flags flags = (pe_action_optional | pe_action_runnable | pe_action_pseudo); for (gIter = action->rsc->children; gIter != NULL; gIter = gIter->next) { resource_t *child = (resource_t *) gIter->data; enum action_tasks task = get_complex_task(child, action->task, TRUE); const char *task_s = task2text(task); action_t *child_action = find_first_action(child->actions, NULL, task_s, node); if (child_action) { enum pe_action_flags child_flags = child->cmds->action_flags(child_action, node); if (is_set(flags, pe_action_optional) && is_set(child_flags, pe_action_optional) == FALSE) { pe_rsc_trace(action->rsc, "%s is mandatory because of %s", action->uuid, child_action->uuid); clear_bit(flags, pe_action_optional); pe_clear_action_bit(action, pe_action_optional); } if (safe_str_neq(task_s, action->task) && is_set(flags, pe_action_runnable) && is_set(child_flags, pe_action_runnable) == FALSE) { pe_rsc_trace(action->rsc, "%s is not runnable because of %s", action->uuid, child_action->uuid); clear_bit(flags, pe_action_runnable); pe_clear_action_bit(action, pe_action_runnable); } } else if (task != stop_rsc && task != action_demote) { pe_rsc_trace(action->rsc, "%s is not runnable because of %s (not found in %s)", action->uuid, task_s, child->id); clear_bit(flags, pe_action_runnable); } } return flags; } enum pe_graph_flags group_update_actions(pe_action_t *first, pe_action_t *then, pe_node_t *node, enum pe_action_flags flags, enum pe_action_flags filter, enum pe_ordering type, pe_working_set_t *data_set) { GListPtr gIter = then->rsc->children; enum pe_graph_flags changed = pe_graph_none; CRM_ASSERT(then->rsc != NULL); changed |= native_update_actions(first, then, node, flags, filter, type, data_set); for (; gIter != NULL; gIter = gIter->next) { resource_t *child = (resource_t *) gIter->data; action_t *child_action = find_first_action(child->actions, NULL, then->task, node); if (child_action) { changed |= child->cmds->update_actions(first, child_action, node, flags, filter, type, data_set); } } return changed; } void group_rsc_location(pe_resource_t *rsc, pe__location_t *constraint) { GListPtr gIter = rsc->children; GListPtr saved = constraint->node_list_rh; GListPtr zero = node_list_dup(constraint->node_list_rh, TRUE, FALSE); gboolean reset_scores = TRUE; group_variant_data_t *group_data = NULL; get_group_variant_data(group_data, rsc); pe_rsc_debug(rsc, "Processing rsc_location %s for %s", constraint->id, rsc->id); native_rsc_location(rsc, constraint); for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; child_rsc->cmds->rsc_location(child_rsc, constraint); if (group_data->colocated && reset_scores) { reset_scores = FALSE; constraint->node_list_rh = zero; } } constraint->node_list_rh = saved; g_list_free_full(zero, free); } void group_expand(resource_t * rsc, pe_working_set_t * data_set) { CRM_CHECK(rsc != NULL, return); pe_rsc_trace(rsc, "Processing actions from %s", rsc->id); native_expand(rsc, data_set); for (GListPtr gIter = rsc->children; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; child_rsc->cmds->expand(child_rsc, data_set); } } GHashTable * group_merge_weights(resource_t * rsc, const char *rhs, GHashTable * nodes, const char *attr, float factor, enum pe_weights flags) { GListPtr gIter = rsc->rsc_cons_lhs; group_variant_data_t *group_data = NULL; get_group_variant_data(group_data, rsc); if (is_set(rsc->flags, pe_rsc_merging)) { pe_rsc_info(rsc, "Breaking dependency loop with %s at %s", rsc->id, rhs); return nodes; } set_bit(rsc->flags, pe_rsc_merging); nodes = group_data->first_child->cmds->merge_weights(group_data->first_child, rhs, nodes, attr, factor, flags); for (; gIter != NULL; gIter = gIter->next) { rsc_colocation_t *constraint = (rsc_colocation_t *) gIter->data; nodes = native_merge_weights(constraint->rsc_lh, rsc->id, nodes, constraint->node_attribute, (float)constraint->score / INFINITY, flags); } clear_bit(rsc->flags, pe_rsc_merging); return nodes; } void group_append_meta(resource_t * rsc, xmlNode * xml) { } diff --git a/lib/pacemaker/pcmk_sched_native.c b/lib/pacemaker/pcmk_sched_native.c index 71d4486b39..bbf3eb7332 100644 --- a/lib/pacemaker/pcmk_sched_native.c +++ b/lib/pacemaker/pcmk_sched_native.c @@ -1,3329 +1,3326 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include // The controller removes the resource from the CIB, making this redundant // #define DELETE_THEN_REFRESH 1 #define INFINITY_HACK (INFINITY * -100) #define VARIANT_NATIVE 1 #include void native_rsc_colocation_rh_must(resource_t * rsc_lh, gboolean update_lh, resource_t * rsc_rh, gboolean update_rh); void native_rsc_colocation_rh_mustnot(resource_t * rsc_lh, gboolean update_lh, resource_t * rsc_rh, gboolean update_rh); static void Recurring(resource_t *rsc, action_t *start, node_t *node, pe_working_set_t *data_set); static void RecurringOp(resource_t *rsc, action_t *start, node_t *node, xmlNode *operation, pe_working_set_t *data_set); static void Recurring_Stopped(resource_t *rsc, action_t *start, node_t *node, pe_working_set_t *data_set); static void RecurringOp_Stopped(resource_t *rsc, action_t *start, node_t *node, xmlNode *operation, pe_working_set_t *data_set); void ReloadRsc(resource_t * rsc, node_t *node, pe_working_set_t * data_set); gboolean DeleteRsc(resource_t * rsc, node_t * node, gboolean optional, pe_working_set_t * data_set); gboolean StopRsc(resource_t * rsc, node_t * next, gboolean optional, pe_working_set_t * data_set); gboolean StartRsc(resource_t * rsc, node_t * next, gboolean optional, pe_working_set_t * data_set); gboolean DemoteRsc(resource_t * rsc, node_t * next, gboolean optional, pe_working_set_t * data_set); gboolean PromoteRsc(resource_t * rsc, node_t * next, gboolean optional, pe_working_set_t * data_set); gboolean RoleError(resource_t * rsc, node_t * next, gboolean optional, pe_working_set_t * data_set); gboolean NullOp(resource_t * rsc, node_t * next, gboolean optional, pe_working_set_t * data_set); /* *INDENT-OFF* */ enum rsc_role_e rsc_state_matrix[RSC_ROLE_MAX][RSC_ROLE_MAX] = { /* Current State */ /* Next State: Unknown Stopped Started Slave Master */ /* Unknown */ { RSC_ROLE_UNKNOWN, RSC_ROLE_STOPPED, RSC_ROLE_STOPPED, RSC_ROLE_STOPPED, RSC_ROLE_STOPPED, }, /* Stopped */ { RSC_ROLE_STOPPED, RSC_ROLE_STOPPED, RSC_ROLE_STARTED, RSC_ROLE_SLAVE, RSC_ROLE_SLAVE, }, /* Started */ { RSC_ROLE_STOPPED, RSC_ROLE_STOPPED, RSC_ROLE_STARTED, RSC_ROLE_SLAVE, RSC_ROLE_MASTER, }, /* Slave */ { RSC_ROLE_STOPPED, RSC_ROLE_STOPPED, RSC_ROLE_STOPPED, RSC_ROLE_SLAVE, RSC_ROLE_MASTER, }, /* Master */ { RSC_ROLE_STOPPED, RSC_ROLE_SLAVE, RSC_ROLE_SLAVE, RSC_ROLE_SLAVE, RSC_ROLE_MASTER, }, }; gboolean (*rsc_action_matrix[RSC_ROLE_MAX][RSC_ROLE_MAX])(resource_t*,node_t*,gboolean,pe_working_set_t*) = { /* Current State */ /* Next State: Unknown Stopped Started Slave Master */ /* Unknown */ { RoleError, StopRsc, RoleError, RoleError, RoleError, }, /* Stopped */ { RoleError, NullOp, StartRsc, StartRsc, RoleError, }, /* Started */ { RoleError, StopRsc, NullOp, NullOp, PromoteRsc, }, /* Slave */ { RoleError, StopRsc, StopRsc, NullOp, PromoteRsc, }, /* Master */ { RoleError, DemoteRsc, DemoteRsc, DemoteRsc, NullOp, }, }; /* *INDENT-ON* */ static gboolean native_choose_node(resource_t * rsc, node_t * prefer, pe_working_set_t * data_set) { GListPtr nodes = NULL; node_t *chosen = NULL; node_t *best = NULL; int multiple = 1; int length = 0; gboolean result = FALSE; process_utilization(rsc, &prefer, data_set); if (is_not_set(rsc->flags, pe_rsc_provisional)) { return rsc->allocated_to ? TRUE : FALSE; } // Sort allowed nodes by weight if (rsc->allowed_nodes) { length = g_hash_table_size(rsc->allowed_nodes); } if (length > 0) { nodes = g_hash_table_get_values(rsc->allowed_nodes); nodes = sort_nodes_by_weight(nodes, pe__current_node(rsc), data_set); // First node in sorted list has the best score best = g_list_nth_data(nodes, 0); } if (prefer && nodes) { chosen = g_hash_table_lookup(rsc->allowed_nodes, prefer->details->id); if (chosen == NULL) { pe_rsc_trace(rsc, "Preferred node %s for %s was unknown", prefer->details->uname, rsc->id); /* Favor the preferred node as long as its weight is at least as good as * the best allowed node's. * * An alternative would be to favor the preferred node even if the best * node is better, when the best node's weight is less than INFINITY. */ } else if ((chosen->weight < 0) || (chosen->weight < best->weight)) { pe_rsc_trace(rsc, "Preferred node %s for %s was unsuitable", chosen->details->uname, rsc->id); chosen = NULL; } else if (!can_run_resources(chosen)) { pe_rsc_trace(rsc, "Preferred node %s for %s was unavailable", chosen->details->uname, rsc->id); chosen = NULL; } else { pe_rsc_trace(rsc, "Chose preferred node %s for %s (ignoring %d candidates)", chosen->details->uname, rsc->id, length); } } if ((chosen == NULL) && nodes) { /* Either there is no preferred node, or the preferred node is not * available, but there are other nodes allowed to run the resource. */ chosen = best; pe_rsc_trace(rsc, "Chose node %s for %s from %d candidates", chosen ? chosen->details->uname : "", rsc->id, length); if (!pe_rsc_is_unique_clone(rsc->parent) && chosen && (chosen->weight > 0) && can_run_resources(chosen)) { /* If the resource is already running on a node, prefer that node if * it is just as good as the chosen node. * * We don't do this for unique clone instances, because * distribute_children() has already assigned instances to their * running nodes when appropriate, and if we get here, we don't want * remaining unallocated instances to prefer a node that's already * running another instance. */ node_t *running = pe__current_node(rsc); if (running && (can_run_resources(running) == FALSE)) { pe_rsc_trace(rsc, "Current node for %s (%s) can't run resources", rsc->id, running->details->uname); } else if (running) { for (GList *iter = nodes->next; iter; iter = iter->next) { node_t *tmp = (node_t *) iter->data; if (tmp->weight != chosen->weight) { // The nodes are sorted by weight, so no more are equal break; } if (tmp->details == running->details) { // Scores are equal, so prefer the current node chosen = tmp; } multiple++; } } } } if (multiple > 1) { static char score[33]; int log_level = (chosen->weight >= INFINITY)? LOG_WARNING : LOG_INFO; score2char_stack(chosen->weight, score, sizeof(score)); do_crm_log(log_level, "Chose node %s for %s from %d nodes with score %s", chosen->details->uname, rsc->id, multiple, score); } result = native_assign_node(rsc, nodes, chosen, FALSE); g_list_free(nodes); return result; } static int node_list_attr_score(GHashTable * list, const char *attr, const char *value) { GHashTableIter iter; node_t *node = NULL; int best_score = -INFINITY; const char *best_node = NULL; if (attr == NULL) { attr = CRM_ATTR_UNAME; } g_hash_table_iter_init(&iter, list); while (g_hash_table_iter_next(&iter, NULL, (void **)&node)) { int weight = node->weight; if (can_run_resources(node) == FALSE) { weight = -INFINITY; } if (weight > best_score || best_node == NULL) { const char *tmp = pe_node_attribute_raw(node, attr); if (safe_str_eq(value, tmp)) { best_score = weight; best_node = node->details->uname; } } } if (safe_str_neq(attr, CRM_ATTR_UNAME)) { crm_info("Best score for %s=%s was %s with %d", attr, value, best_node ? best_node : "", best_score); } return best_score; } static void node_hash_update(GHashTable * list1, GHashTable * list2, const char *attr, float factor, gboolean only_positive) { int score = 0; int new_score = 0; GHashTableIter iter; node_t *node = NULL; if (attr == NULL) { attr = CRM_ATTR_UNAME; } g_hash_table_iter_init(&iter, list1); while (g_hash_table_iter_next(&iter, NULL, (void **)&node)) { float weight_f = 0; int weight = 0; CRM_LOG_ASSERT(node != NULL); if(node == NULL) { continue; }; score = node_list_attr_score(list2, attr, pe_node_attribute_raw(node, attr)); weight_f = factor * score; /* Round the number */ /* http://c-faq.com/fp/round.html */ weight = (int)(weight_f < 0 ? weight_f - 0.5 : weight_f + 0.5); new_score = merge_weights(weight, node->weight); if (factor < 0 && score < 0) { /* Negative preference for a node with a negative score * should not become a positive preference * * TODO - Decide if we want to filter only if weight == -INFINITY * */ crm_trace("%s: Filtering %d + %f*%d (factor * score)", node->details->uname, node->weight, factor, score); } else if (node->weight == INFINITY_HACK) { crm_trace("%s: Filtering %d + %f*%d (node < 0)", node->details->uname, node->weight, factor, score); } else if (only_positive && new_score < 0 && node->weight > 0) { node->weight = INFINITY_HACK; crm_trace("%s: Filtering %d + %f*%d (score > 0)", node->details->uname, node->weight, factor, score); } else if (only_positive && new_score < 0 && node->weight == 0) { crm_trace("%s: Filtering %d + %f*%d (score == 0)", node->details->uname, node->weight, factor, score); } else { crm_trace("%s: %d + %f*%d", node->details->uname, node->weight, factor, score); node->weight = new_score; } } } GHashTable * node_hash_dup(GHashTable * hash) { /* Hack! */ GListPtr list = g_hash_table_get_values(hash); GHashTable *result = node_hash_from_list(list); g_list_free(list); return result; } GHashTable * native_merge_weights(resource_t * rsc, const char *rhs, GHashTable * nodes, const char *attr, float factor, enum pe_weights flags) { return rsc_merge_weights(rsc, rhs, nodes, attr, factor, flags); } GHashTable * rsc_merge_weights(resource_t * rsc, const char *rhs, GHashTable * nodes, const char *attr, float factor, enum pe_weights flags) { GHashTable *work = NULL; int multiplier = 1; if (factor < 0) { multiplier = -1; } if (is_set(rsc->flags, pe_rsc_merging)) { pe_rsc_info(rsc, "%s: Breaking dependency loop at %s", rhs, rsc->id); return nodes; } set_bit(rsc->flags, pe_rsc_merging); if (is_set(flags, pe_weights_init)) { if (rsc->variant == pe_group && rsc->children) { GListPtr last = rsc->children; while (last->next != NULL) { last = last->next; } pe_rsc_trace(rsc, "Merging %s as a group %p %p", rsc->id, rsc->children, last); work = rsc_merge_weights(last->data, rhs, NULL, attr, factor, flags); } else { work = node_hash_dup(rsc->allowed_nodes); } clear_bit(flags, pe_weights_init); } else if (rsc->variant == pe_group && rsc->children) { GListPtr iter = rsc->children; pe_rsc_trace(rsc, "%s: Combining scores from %d children of %s", rhs, g_list_length(iter), rsc->id); work = node_hash_dup(nodes); for(iter = rsc->children; iter->next != NULL; iter = iter->next) { work = rsc_merge_weights(iter->data, rhs, work, attr, factor, flags); } } else { pe_rsc_trace(rsc, "%s: Combining scores from %s", rhs, rsc->id); work = node_hash_dup(nodes); node_hash_update(work, rsc->allowed_nodes, attr, factor, is_set(flags, pe_weights_positive)); } if (is_set(flags, pe_weights_rollback) && can_run_any(work) == FALSE) { pe_rsc_info(rsc, "%s: Rolling back scores from %s", rhs, rsc->id); g_hash_table_destroy(work); clear_bit(rsc->flags, pe_rsc_merging); return nodes; } if (can_run_any(work)) { GListPtr gIter = NULL; if (is_set(flags, pe_weights_forward)) { gIter = rsc->rsc_cons; crm_trace("Checking %d additional colocation constraints", g_list_length(gIter)); } else if(rsc->variant == pe_group && rsc->children) { GListPtr last = rsc->children; while (last->next != NULL) { last = last->next; } gIter = ((resource_t*)last->data)->rsc_cons_lhs; crm_trace("Checking %d additional optional group colocation constraints from %s", g_list_length(gIter), ((resource_t*)last->data)->id); } else { gIter = rsc->rsc_cons_lhs; crm_trace("Checking %d additional optional colocation constraints %s", g_list_length(gIter), rsc->id); } for (; gIter != NULL; gIter = gIter->next) { resource_t *other = NULL; rsc_colocation_t *constraint = (rsc_colocation_t *) gIter->data; if (is_set(flags, pe_weights_forward)) { other = constraint->rsc_rh; } else { other = constraint->rsc_lh; } pe_rsc_trace(rsc, "Applying %s (%s)", constraint->id, other->id); work = rsc_merge_weights(other, rhs, work, constraint->node_attribute, multiplier * (float)constraint->score / INFINITY, flags|pe_weights_rollback); dump_node_scores(LOG_TRACE, NULL, rhs, work); } } if (is_set(flags, pe_weights_positive)) { node_t *node = NULL; GHashTableIter iter; g_hash_table_iter_init(&iter, work); while (g_hash_table_iter_next(&iter, NULL, (void **)&node)) { if (node->weight == INFINITY_HACK) { node->weight = 1; } } } if (nodes) { g_hash_table_destroy(nodes); } clear_bit(rsc->flags, pe_rsc_merging); return work; } static inline bool node_has_been_unfenced(node_t *node) { const char *unfenced = pe_node_attribute_raw(node, CRM_ATTR_UNFENCED); return unfenced && strcmp("0", unfenced); } static inline bool is_unfence_device(resource_t *rsc, pe_working_set_t *data_set) { return is_set(rsc->flags, pe_rsc_fence_device) && is_set(data_set->flags, pe_flag_enable_unfencing); } node_t * native_color(resource_t * rsc, node_t * prefer, pe_working_set_t * data_set) { GListPtr gIter = NULL; int alloc_details = scores_log_level + 1; if (rsc->parent && is_not_set(rsc->parent->flags, pe_rsc_allocating)) { /* never allocate children on their own */ pe_rsc_debug(rsc, "Escalating allocation of %s to its parent: %s", rsc->id, rsc->parent->id); rsc->parent->cmds->allocate(rsc->parent, prefer, data_set); } if (is_not_set(rsc->flags, pe_rsc_provisional)) { return rsc->allocated_to; } if (is_set(rsc->flags, pe_rsc_allocating)) { pe_rsc_debug(rsc, "Dependency loop detected involving %s", rsc->id); return NULL; } set_bit(rsc->flags, pe_rsc_allocating); - print_resource(alloc_details, "Allocating: ", rsc, FALSE); dump_node_scores(alloc_details, rsc, "Pre-alloc", rsc->allowed_nodes); for (gIter = rsc->rsc_cons; gIter != NULL; gIter = gIter->next) { rsc_colocation_t *constraint = (rsc_colocation_t *) gIter->data; GHashTable *archive = NULL; resource_t *rsc_rh = constraint->rsc_rh; pe_rsc_trace(rsc, "%s: Pre-Processing %s (%s, %d, %s)", rsc->id, constraint->id, rsc_rh->id, constraint->score, role2text(constraint->role_lh)); if (constraint->role_lh >= RSC_ROLE_MASTER || (constraint->score < 0 && constraint->score > -INFINITY)) { archive = node_hash_dup(rsc->allowed_nodes); } rsc_rh->cmds->allocate(rsc_rh, NULL, data_set); rsc->cmds->rsc_colocation_lh(rsc, rsc_rh, constraint, data_set); if (archive && can_run_any(rsc->allowed_nodes) == FALSE) { pe_rsc_info(rsc, "%s: Rolling back scores from %s", rsc->id, rsc_rh->id); g_hash_table_destroy(rsc->allowed_nodes); rsc->allowed_nodes = archive; archive = NULL; } if (archive) { g_hash_table_destroy(archive); } } dump_node_scores(alloc_details, rsc, "Post-coloc", rsc->allowed_nodes); for (gIter = rsc->rsc_cons_lhs; gIter != NULL; gIter = gIter->next) { rsc_colocation_t *constraint = (rsc_colocation_t *) gIter->data; rsc->allowed_nodes = constraint->rsc_lh->cmds->merge_weights(constraint->rsc_lh, rsc->id, rsc->allowed_nodes, constraint->node_attribute, (float)constraint->score / INFINITY, pe_weights_rollback); } - print_resource(LOG_TRACE, "Allocating: ", rsc, FALSE); if (rsc->next_role == RSC_ROLE_STOPPED) { pe_rsc_trace(rsc, "Making sure %s doesn't get allocated", rsc->id); /* make sure it doesn't come up again */ resource_location(rsc, NULL, -INFINITY, XML_RSC_ATTR_TARGET_ROLE, data_set); } else if(rsc->next_role > rsc->role && is_set(data_set->flags, pe_flag_have_quorum) == FALSE && data_set->no_quorum_policy == no_quorum_freeze) { crm_notice("Resource %s cannot be elevated from %s to %s: no-quorum-policy=freeze", rsc->id, role2text(rsc->role), role2text(rsc->next_role)); rsc->next_role = rsc->role; } dump_node_scores(show_scores ? 0 : scores_log_level, rsc, __FUNCTION__, rsc->allowed_nodes); if (is_set(data_set->flags, pe_flag_stonith_enabled) && is_set(data_set->flags, pe_flag_have_stonith_resource) == FALSE) { clear_bit(rsc->flags, pe_rsc_managed); } if (is_not_set(rsc->flags, pe_rsc_managed)) { const char *reason = NULL; node_t *assign_to = NULL; rsc->next_role = rsc->role; assign_to = pe__current_node(rsc); if (assign_to == NULL) { reason = "inactive"; } else if (rsc->role == RSC_ROLE_MASTER) { reason = "master"; } else if (is_set(rsc->flags, pe_rsc_failed)) { reason = "failed"; } else { reason = "active"; } pe_rsc_info(rsc, "Unmanaged resource %s allocated to %s: %s", rsc->id, (assign_to? assign_to->details->uname : "no node"), reason); native_assign_node(rsc, NULL, assign_to, TRUE); } else if (is_set(data_set->flags, pe_flag_stop_everything)) { pe_rsc_debug(rsc, "Forcing %s to stop", rsc->id); native_assign_node(rsc, NULL, NULL, TRUE); } else if (is_set(rsc->flags, pe_rsc_provisional) && native_choose_node(rsc, prefer, data_set)) { pe_rsc_trace(rsc, "Allocated resource %s to %s", rsc->id, rsc->allocated_to->details->uname); } else if (rsc->allocated_to == NULL) { if (is_not_set(rsc->flags, pe_rsc_orphan)) { pe_rsc_info(rsc, "Resource %s cannot run anywhere", rsc->id); } else if (rsc->running_on != NULL) { pe_rsc_info(rsc, "Stopping orphan resource %s", rsc->id); } } else { pe_rsc_debug(rsc, "Pre-Allocated resource %s to %s", rsc->id, rsc->allocated_to->details->uname); } clear_bit(rsc->flags, pe_rsc_allocating); - print_resource(LOG_TRACE, "Allocated ", rsc, TRUE); if (rsc->is_remote_node) { node_t *remote_node = pe_find_node(data_set->nodes, rsc->id); CRM_ASSERT(remote_node != NULL); if (rsc->allocated_to && rsc->next_role != RSC_ROLE_STOPPED) { crm_trace("Setting Pacemaker Remote node %s to ONLINE", remote_node->details->id); remote_node->details->online = TRUE; /* We shouldn't consider an unseen remote-node unclean if we are going * to try and connect to it. Otherwise we get an unnecessary fence */ if (remote_node->details->unseen == TRUE) { remote_node->details->unclean = FALSE; } } else { crm_trace("Setting Pacemaker Remote node %s to SHUTDOWN (next role %s, %sallocated)", remote_node->details->id, role2text(rsc->next_role), (rsc->allocated_to? "" : "un")); remote_node->details->shutdown = TRUE; } } return rsc->allocated_to; } static gboolean is_op_dup(resource_t *rsc, const char *name, guint interval_ms) { gboolean dup = FALSE; const char *id = NULL; const char *value = NULL; xmlNode *operation = NULL; guint interval2_ms = 0; CRM_ASSERT(rsc); for (operation = __xml_first_child_element(rsc->ops_xml); operation != NULL; operation = __xml_next_element(operation)) { if (crm_str_eq((const char *)operation->name, "op", TRUE)) { value = crm_element_value(operation, "name"); if (safe_str_neq(value, name)) { continue; } value = crm_element_value(operation, XML_LRM_ATTR_INTERVAL); interval2_ms = crm_parse_interval_spec(value); if (interval_ms != interval2_ms) { continue; } if (id == NULL) { id = ID(operation); } else { crm_config_err("Operation %s is a duplicate of %s", ID(operation), id); crm_config_err ("Do not use the same (name, interval) combination more than once per resource"); dup = TRUE; } } } return dup; } static bool op_cannot_recur(const char *name) { return safe_str_eq(name, RSC_STOP) || safe_str_eq(name, RSC_START) || safe_str_eq(name, RSC_DEMOTE) || safe_str_eq(name, RSC_PROMOTE); } static void RecurringOp(resource_t * rsc, action_t * start, node_t * node, xmlNode * operation, pe_working_set_t * data_set) { char *key = NULL; const char *name = NULL; const char *role = NULL; const char *interval_spec = NULL; const char *node_uname = node? node->details->uname : "n/a"; guint interval_ms = 0; action_t *mon = NULL; gboolean is_optional = TRUE; GListPtr possible_matches = NULL; CRM_ASSERT(rsc); /* Only process for the operations without role="Stopped" */ role = crm_element_value(operation, "role"); if (role && text2role(role) == RSC_ROLE_STOPPED) { return; } interval_spec = crm_element_value(operation, XML_LRM_ATTR_INTERVAL); interval_ms = crm_parse_interval_spec(interval_spec); if (interval_ms == 0) { return; } name = crm_element_value(operation, "name"); if (is_op_dup(rsc, name, interval_ms)) { crm_trace("Not creating duplicate recurring action %s for %dms %s", ID(operation), interval_ms, name); return; } if (op_cannot_recur(name)) { crm_config_err("Ignoring %s because action '%s' cannot be recurring", ID(operation), name); return; } key = generate_op_key(rsc->id, name, interval_ms); if (find_rsc_op_entry(rsc, key) == NULL) { crm_trace("Not creating recurring action %s for disabled resource %s", ID(operation), rsc->id); free(key); return; } pe_rsc_trace(rsc, "Creating recurring action %s for %s in role %s on %s", ID(operation), rsc->id, role2text(rsc->next_role), node_uname); if (start != NULL) { pe_rsc_trace(rsc, "Marking %s %s due to %s", key, is_set(start->flags, pe_action_optional) ? "optional" : "mandatory", start->uuid); is_optional = (rsc->cmds->action_flags(start, NULL) & pe_action_optional); } else { pe_rsc_trace(rsc, "Marking %s optional", key); is_optional = TRUE; } /* start a monitor for an already active resource */ possible_matches = find_actions_exact(rsc->actions, key, node); if (possible_matches == NULL) { is_optional = FALSE; pe_rsc_trace(rsc, "Marking %s mandatory: not active", key); } else { GListPtr gIter = NULL; for (gIter = possible_matches; gIter != NULL; gIter = gIter->next) { action_t *op = (action_t *) gIter->data; if (is_set(op->flags, pe_action_reschedule)) { is_optional = FALSE; break; } } g_list_free(possible_matches); } if ((rsc->next_role == RSC_ROLE_MASTER && role == NULL) || (role != NULL && text2role(role) != rsc->next_role)) { int log_level = LOG_TRACE; const char *result = "Ignoring"; if (is_optional) { char *after_key = NULL; action_t *cancel_op = NULL; // It's running, so cancel it log_level = LOG_INFO; result = "Cancelling"; cancel_op = pe_cancel_op(rsc, name, interval_ms, node, data_set); switch (rsc->role) { case RSC_ROLE_SLAVE: case RSC_ROLE_STARTED: if (rsc->next_role == RSC_ROLE_MASTER) { after_key = promote_key(rsc); } else if (rsc->next_role == RSC_ROLE_STOPPED) { after_key = stop_key(rsc); } break; case RSC_ROLE_MASTER: after_key = demote_key(rsc); break; default: break; } if (after_key) { custom_action_order(rsc, NULL, cancel_op, rsc, after_key, NULL, pe_order_runnable_left, data_set); } } do_crm_log(log_level, "%s action %s (%s vs. %s)", result, key, role ? role : role2text(RSC_ROLE_SLAVE), role2text(rsc->next_role)); free(key); return; } mon = custom_action(rsc, key, name, node, is_optional, TRUE, data_set); key = mon->uuid; if (is_optional) { pe_rsc_trace(rsc, "%s\t %s (optional)", node_uname, mon->uuid); } if (start == NULL || is_set(start->flags, pe_action_runnable) == FALSE) { pe_rsc_debug(rsc, "%s\t %s (cancelled : start un-runnable)", node_uname, mon->uuid); update_action_flags(mon, pe_action_runnable | pe_action_clear, __FUNCTION__, __LINE__); } else if (node == NULL || node->details->online == FALSE || node->details->unclean) { pe_rsc_debug(rsc, "%s\t %s (cancelled : no node available)", node_uname, mon->uuid); update_action_flags(mon, pe_action_runnable | pe_action_clear, __FUNCTION__, __LINE__); } else if (is_set(mon->flags, pe_action_optional) == FALSE) { pe_rsc_info(rsc, " Start recurring %s (%us) for %s on %s", mon->task, interval_ms / 1000, rsc->id, node_uname); } if (rsc->next_role == RSC_ROLE_MASTER) { char *running_master = crm_itoa(PCMK_OCF_RUNNING_MASTER); add_hash_param(mon->meta, XML_ATTR_TE_TARGET_RC, running_master); free(running_master); } if (node == NULL || is_set(rsc->flags, pe_rsc_managed)) { custom_action_order(rsc, start_key(rsc), NULL, NULL, strdup(key), mon, pe_order_implies_then | pe_order_runnable_left, data_set); custom_action_order(rsc, reload_key(rsc), NULL, NULL, strdup(key), mon, pe_order_implies_then | pe_order_runnable_left, data_set); if (rsc->next_role == RSC_ROLE_MASTER) { custom_action_order(rsc, promote_key(rsc), NULL, rsc, NULL, mon, pe_order_optional | pe_order_runnable_left, data_set); } else if (rsc->role == RSC_ROLE_MASTER) { custom_action_order(rsc, demote_key(rsc), NULL, rsc, NULL, mon, pe_order_optional | pe_order_runnable_left, data_set); } } } static void Recurring(resource_t * rsc, action_t * start, node_t * node, pe_working_set_t * data_set) { if (is_not_set(rsc->flags, pe_rsc_maintenance) && (node == NULL || node->details->maintenance == FALSE)) { xmlNode *operation = NULL; for (operation = __xml_first_child_element(rsc->ops_xml); operation != NULL; operation = __xml_next_element(operation)) { if (crm_str_eq((const char *)operation->name, "op", TRUE)) { RecurringOp(rsc, start, node, operation, data_set); } } } } static void RecurringOp_Stopped(resource_t * rsc, action_t * start, node_t * node, xmlNode * operation, pe_working_set_t * data_set) { char *key = NULL; const char *name = NULL; const char *role = NULL; const char *interval_spec = NULL; const char *node_uname = node? node->details->uname : "n/a"; guint interval_ms = 0; GListPtr possible_matches = NULL; GListPtr gIter = NULL; /* Only process for the operations with role="Stopped" */ role = crm_element_value(operation, "role"); if (role == NULL || text2role(role) != RSC_ROLE_STOPPED) { return; } interval_spec = crm_element_value(operation, XML_LRM_ATTR_INTERVAL); interval_ms = crm_parse_interval_spec(interval_spec); if (interval_ms == 0) { return; } name = crm_element_value(operation, "name"); if (is_op_dup(rsc, name, interval_ms)) { crm_trace("Not creating duplicate recurring action %s for %dms %s", ID(operation), interval_ms, name); return; } if (op_cannot_recur(name)) { crm_config_err("Invalid recurring action %s wth name: '%s'", ID(operation), name); return; } key = generate_op_key(rsc->id, name, interval_ms); if (find_rsc_op_entry(rsc, key) == NULL) { crm_trace("Not creating recurring action %s for disabled resource %s", ID(operation), rsc->id); free(key); return; } // @TODO add support if (is_set(rsc->flags, pe_rsc_unique) == FALSE) { crm_notice("Ignoring %s (recurring monitors for Stopped role are " "not supported for anonymous clones)", ID(operation)); return; } pe_rsc_trace(rsc, "Creating recurring action %s for %s in role %s on nodes where it should not be running", ID(operation), rsc->id, role2text(rsc->next_role)); /* if the monitor exists on the node where the resource will be running, cancel it */ if (node != NULL) { possible_matches = find_actions_exact(rsc->actions, key, node); if (possible_matches) { action_t *cancel_op = NULL; g_list_free(possible_matches); cancel_op = pe_cancel_op(rsc, name, interval_ms, node, data_set); if (rsc->next_role == RSC_ROLE_STARTED || rsc->next_role == RSC_ROLE_SLAVE) { /* rsc->role == RSC_ROLE_STOPPED: cancel the monitor before start */ /* rsc->role == RSC_ROLE_STARTED: for a migration, cancel the monitor on the target node before start */ custom_action_order(rsc, NULL, cancel_op, rsc, start_key(rsc), NULL, pe_order_runnable_left, data_set); } pe_rsc_info(rsc, "Cancel action %s (%s vs. %s) on %s", key, role, role2text(rsc->next_role), node_uname); } } for (gIter = data_set->nodes; gIter != NULL; gIter = gIter->next) { node_t *stop_node = (node_t *) gIter->data; const char *stop_node_uname = stop_node->details->uname; gboolean is_optional = TRUE; gboolean probe_is_optional = TRUE; gboolean stop_is_optional = TRUE; action_t *stopped_mon = NULL; char *rc_inactive = NULL; GListPtr probe_complete_ops = NULL; GListPtr stop_ops = NULL; GListPtr local_gIter = NULL; if (node && safe_str_eq(stop_node_uname, node_uname)) { continue; } pe_rsc_trace(rsc, "Creating recurring action %s for %s on %s", ID(operation), rsc->id, crm_str(stop_node_uname)); /* start a monitor for an already stopped resource */ possible_matches = find_actions_exact(rsc->actions, key, stop_node); if (possible_matches == NULL) { pe_rsc_trace(rsc, "Marking %s mandatory on %s: not active", key, crm_str(stop_node_uname)); is_optional = FALSE; } else { pe_rsc_trace(rsc, "Marking %s optional on %s: already active", key, crm_str(stop_node_uname)); is_optional = TRUE; g_list_free(possible_matches); } stopped_mon = custom_action(rsc, strdup(key), name, stop_node, is_optional, TRUE, data_set); rc_inactive = crm_itoa(PCMK_OCF_NOT_RUNNING); add_hash_param(stopped_mon->meta, XML_ATTR_TE_TARGET_RC, rc_inactive); free(rc_inactive); if (is_set(rsc->flags, pe_rsc_managed)) { GList *probes = pe__resource_actions(rsc, stop_node, RSC_STATUS, FALSE); GListPtr pIter = NULL; for (pIter = probes; pIter != NULL; pIter = pIter->next) { action_t *probe = (action_t *) pIter->data; order_actions(probe, stopped_mon, pe_order_runnable_left); crm_trace("%s then %s on %s", probe->uuid, stopped_mon->uuid, stop_node->details->uname); } g_list_free(probes); } if (probe_complete_ops) { g_list_free(probe_complete_ops); } stop_ops = pe__resource_actions(rsc, stop_node, RSC_STOP, TRUE); for (local_gIter = stop_ops; local_gIter != NULL; local_gIter = local_gIter->next) { action_t *stop = (action_t *) local_gIter->data; if (is_set(stop->flags, pe_action_optional) == FALSE) { stop_is_optional = FALSE; } if (is_set(stop->flags, pe_action_runnable) == FALSE) { crm_debug("%s\t %s (cancelled : stop un-runnable)", crm_str(stop_node_uname), stopped_mon->uuid); update_action_flags(stopped_mon, pe_action_runnable | pe_action_clear, __FUNCTION__, __LINE__); } if (is_set(rsc->flags, pe_rsc_managed)) { custom_action_order(rsc, stop_key(rsc), stop, NULL, strdup(key), stopped_mon, pe_order_implies_then | pe_order_runnable_left, data_set); } } if (stop_ops) { g_list_free(stop_ops); } if (is_optional == FALSE && probe_is_optional && stop_is_optional && is_set(rsc->flags, pe_rsc_managed) == FALSE) { pe_rsc_trace(rsc, "Marking %s optional on %s due to unmanaged", key, crm_str(stop_node_uname)); update_action_flags(stopped_mon, pe_action_optional, __FUNCTION__, __LINE__); } if (is_set(stopped_mon->flags, pe_action_optional)) { pe_rsc_trace(rsc, "%s\t %s (optional)", crm_str(stop_node_uname), stopped_mon->uuid); } if (stop_node->details->online == FALSE || stop_node->details->unclean) { pe_rsc_debug(rsc, "%s\t %s (cancelled : no node available)", crm_str(stop_node_uname), stopped_mon->uuid); update_action_flags(stopped_mon, pe_action_runnable | pe_action_clear, __FUNCTION__, __LINE__); } if (is_set(stopped_mon->flags, pe_action_runnable) && is_set(stopped_mon->flags, pe_action_optional) == FALSE) { crm_notice(" Start recurring %s (%us) for %s on %s", stopped_mon->task, interval_ms / 1000, rsc->id, crm_str(stop_node_uname)); } } free(key); } static void Recurring_Stopped(resource_t * rsc, action_t * start, node_t * node, pe_working_set_t * data_set) { if (is_not_set(rsc->flags, pe_rsc_maintenance) && (node == NULL || node->details->maintenance == FALSE)) { xmlNode *operation = NULL; for (operation = __xml_first_child_element(rsc->ops_xml); operation != NULL; operation = __xml_next_element(operation)) { if (crm_str_eq((const char *)operation->name, "op", TRUE)) { RecurringOp_Stopped(rsc, start, node, operation, data_set); } } } } static void handle_migration_actions(resource_t * rsc, node_t *current, node_t *chosen, pe_working_set_t * data_set) { action_t *migrate_to = NULL; action_t *migrate_from = NULL; action_t *start = NULL; action_t *stop = NULL; gboolean partial = rsc->partial_migration_target ? TRUE : FALSE; pe_rsc_trace(rsc, "Processing migration actions %s moving from %s to %s . partial migration = %s", rsc->id, current->details->id, chosen->details->id, partial ? "TRUE" : "FALSE"); start = start_action(rsc, chosen, TRUE); stop = stop_action(rsc, current, TRUE); if (partial == FALSE) { migrate_to = custom_action(rsc, generate_op_key(rsc->id, RSC_MIGRATE, 0), RSC_MIGRATE, current, TRUE, TRUE, data_set); } migrate_from = custom_action(rsc, generate_op_key(rsc->id, RSC_MIGRATED, 0), RSC_MIGRATED, chosen, TRUE, TRUE, data_set); if ((migrate_to && migrate_from) || (migrate_from && partial)) { set_bit(start->flags, pe_action_migrate_runnable); set_bit(stop->flags, pe_action_migrate_runnable); update_action_flags(start, pe_action_pseudo, __FUNCTION__, __LINE__); /* easier than trying to delete it from the graph */ /* order probes before migrations */ if (partial) { set_bit(migrate_from->flags, pe_action_migrate_runnable); migrate_from->needs = start->needs; custom_action_order(rsc, generate_op_key(rsc->id, RSC_STATUS, 0), NULL, rsc, generate_op_key(rsc->id, RSC_MIGRATED, 0), NULL, pe_order_optional, data_set); } else { set_bit(migrate_from->flags, pe_action_migrate_runnable); set_bit(migrate_to->flags, pe_action_migrate_runnable); migrate_to->needs = start->needs; custom_action_order(rsc, generate_op_key(rsc->id, RSC_STATUS, 0), NULL, rsc, generate_op_key(rsc->id, RSC_MIGRATE, 0), NULL, pe_order_optional, data_set); custom_action_order(rsc, generate_op_key(rsc->id, RSC_MIGRATE, 0), NULL, rsc, generate_op_key(rsc->id, RSC_MIGRATED, 0), NULL, pe_order_optional | pe_order_implies_first_migratable, data_set); } custom_action_order(rsc, generate_op_key(rsc->id, RSC_MIGRATED, 0), NULL, rsc, generate_op_key(rsc->id, RSC_STOP, 0), NULL, pe_order_optional | pe_order_implies_first_migratable, data_set); custom_action_order(rsc, generate_op_key(rsc->id, RSC_MIGRATED, 0), NULL, rsc, generate_op_key(rsc->id, RSC_START, 0), NULL, pe_order_optional | pe_order_implies_first_migratable | pe_order_pseudo_left, data_set); } if (migrate_to) { add_hash_param(migrate_to->meta, XML_LRM_ATTR_MIGRATE_SOURCE, current->details->uname); add_hash_param(migrate_to->meta, XML_LRM_ATTR_MIGRATE_TARGET, chosen->details->uname); /* Pacemaker Remote connections don't require pending to be recorded in * the CIB. We can reduce CIB writes by not setting PENDING for them. */ if (rsc->is_remote_node == FALSE) { /* migrate_to takes place on the source node, but can * have an effect on the target node depending on how * the agent is written. Because of this, we have to maintain * a record that the migrate_to occurred, in case the source node * loses membership while the migrate_to action is still in-flight. */ add_hash_param(migrate_to->meta, XML_OP_ATTR_PENDING, "true"); } } if (migrate_from) { add_hash_param(migrate_from->meta, XML_LRM_ATTR_MIGRATE_SOURCE, current->details->uname); add_hash_param(migrate_from->meta, XML_LRM_ATTR_MIGRATE_TARGET, chosen->details->uname); } } void native_create_actions(resource_t * rsc, pe_working_set_t * data_set) { action_t *start = NULL; node_t *chosen = NULL; node_t *current = NULL; gboolean need_stop = FALSE; gboolean is_moving = FALSE; gboolean allow_migrate = is_set(rsc->flags, pe_rsc_allow_migrate) ? TRUE : FALSE; GListPtr gIter = NULL; unsigned int num_all_active = 0; unsigned int num_clean_active = 0; bool multiply_active = FALSE; enum rsc_role_e role = RSC_ROLE_UNKNOWN; enum rsc_role_e next_role = RSC_ROLE_UNKNOWN; CRM_ASSERT(rsc); chosen = rsc->allocated_to; if (chosen != NULL && rsc->next_role == RSC_ROLE_UNKNOWN) { rsc->next_role = RSC_ROLE_STARTED; pe_rsc_trace(rsc, "Fixed next_role: unknown -> %s", role2text(rsc->next_role)); } else if (rsc->next_role == RSC_ROLE_UNKNOWN) { rsc->next_role = RSC_ROLE_STOPPED; pe_rsc_trace(rsc, "Fixed next_role: unknown -> %s", role2text(rsc->next_role)); } pe_rsc_trace(rsc, "Processing state transition for %s %p: %s->%s", rsc->id, rsc, role2text(rsc->role), role2text(rsc->next_role)); current = pe__find_active_on(rsc, &num_all_active, &num_clean_active); for (gIter = rsc->dangling_migrations; gIter != NULL; gIter = gIter->next) { node_t *dangling_source = (node_t *) gIter->data; action_t *stop = stop_action(rsc, dangling_source, FALSE); set_bit(stop->flags, pe_action_dangle); pe_rsc_trace(rsc, "Forcing a cleanup of %s on %s", rsc->id, dangling_source->details->uname); if (is_set(data_set->flags, pe_flag_remove_after_stop)) { DeleteRsc(rsc, dangling_source, FALSE, data_set); } } if ((num_all_active == 2) && (num_clean_active == 2) && chosen && rsc->partial_migration_source && rsc->partial_migration_target && (current->details == rsc->partial_migration_source->details) && (chosen->details == rsc->partial_migration_target->details)) { /* The chosen node is still the migration target from a partial * migration. Attempt to continue the migration instead of recovering * by stopping the resource everywhere and starting it on a single node. */ pe_rsc_trace(rsc, "Will attempt to continue with a partial migration to target %s from %s", rsc->partial_migration_target->details->id, rsc->partial_migration_source->details->id); } else if (is_not_set(rsc->flags, pe_rsc_needs_fencing)) { /* If a resource has "requires" set to nothing or quorum, don't consider * it active on unclean nodes (similar to how all resources behave when * stonith-enabled is false). We can start such resources elsewhere * before fencing completes, and if we considered the resource active on * the failed node, we would attempt recovery for being active on * multiple nodes. */ multiply_active = (num_clean_active > 1); } else { multiply_active = (num_all_active > 1); } if (multiply_active) { if (rsc->partial_migration_target && rsc->partial_migration_source) { // Migration was in progress, but we've chosen a different target crm_notice("Resource %s can no longer migrate to %s. Stopping on %s too", rsc->id, rsc->partial_migration_target->details->uname, rsc->partial_migration_source->details->uname); } else { // Resource was incorrectly multiply active pe_proc_err("Resource %s is active on %u nodes (%s)", rsc->id, num_all_active, recovery2text(rsc->recovery_type)); crm_notice("See https://wiki.clusterlabs.org/wiki/FAQ#Resource_is_Too_Active for more information"); } if (rsc->recovery_type == recovery_stop_start) { need_stop = TRUE; } /* If by chance a partial migration is in process, but the migration * target is not chosen still, clear all partial migration data. */ rsc->partial_migration_source = rsc->partial_migration_target = NULL; allow_migrate = FALSE; } if (is_set(rsc->flags, pe_rsc_start_pending)) { start = start_action(rsc, chosen, TRUE); set_bit(start->flags, pe_action_print_always); } if (current && chosen && current->details != chosen->details) { pe_rsc_trace(rsc, "Moving %s", rsc->id); is_moving = TRUE; need_stop = TRUE; } else if (is_set(rsc->flags, pe_rsc_failed)) { pe_rsc_trace(rsc, "Recovering %s", rsc->id); need_stop = TRUE; } else if (is_set(rsc->flags, pe_rsc_block)) { pe_rsc_trace(rsc, "Block %s", rsc->id); need_stop = TRUE; } else if (rsc->role > RSC_ROLE_STARTED && current != NULL && chosen != NULL) { /* Recovery of a promoted resource */ start = start_action(rsc, chosen, TRUE); if (is_set(start->flags, pe_action_optional) == FALSE) { pe_rsc_trace(rsc, "Forced start %s", rsc->id); need_stop = TRUE; } } pe_rsc_trace(rsc, "Creating actions for %s: %s->%s", rsc->id, role2text(rsc->role), role2text(rsc->next_role)); /* Create any additional actions required when bringing resource down and * back up to same level. */ role = rsc->role; while (role != RSC_ROLE_STOPPED) { next_role = rsc_state_matrix[role][RSC_ROLE_STOPPED]; pe_rsc_trace(rsc, "Down: Executing: %s->%s (%s)%s", role2text(role), role2text(next_role), rsc->id, need_stop ? " required" : ""); if (rsc_action_matrix[role][next_role] (rsc, current, !need_stop, data_set) == FALSE) { break; } role = next_role; } while (rsc->role <= rsc->next_role && role != rsc->role && is_not_set(rsc->flags, pe_rsc_block)) { next_role = rsc_state_matrix[role][rsc->role]; pe_rsc_trace(rsc, "Up: Executing: %s->%s (%s)%s", role2text(role), role2text(next_role), rsc->id, need_stop ? " required" : ""); if (rsc_action_matrix[role][next_role] (rsc, chosen, !need_stop, data_set) == FALSE) { break; } role = next_role; } role = rsc->role; /* Required steps from this role to the next */ while (role != rsc->next_role) { next_role = rsc_state_matrix[role][rsc->next_role]; pe_rsc_trace(rsc, "Role: Executing: %s->%s = (%s on %s)", role2text(role), role2text(next_role), rsc->id, chosen?chosen->details->uname:"NA"); if (rsc_action_matrix[role][next_role] (rsc, chosen, FALSE, data_set) == FALSE) { break; } role = next_role; } if(is_set(rsc->flags, pe_rsc_block)) { pe_rsc_trace(rsc, "No monitor additional ops for blocked resource"); } else if (rsc->next_role != RSC_ROLE_STOPPED || is_set(rsc->flags, pe_rsc_managed) == FALSE) { pe_rsc_trace(rsc, "Monitor ops for active resource"); start = start_action(rsc, chosen, TRUE); Recurring(rsc, start, chosen, data_set); Recurring_Stopped(rsc, start, chosen, data_set); } else { pe_rsc_trace(rsc, "Monitor ops for inactive resource"); Recurring_Stopped(rsc, NULL, NULL, data_set); } /* if we are stuck in a partial migration, where the target * of the partial migration no longer matches the chosen target. * A full stop/start is required */ if (rsc->partial_migration_target && (chosen == NULL || rsc->partial_migration_target->details != chosen->details)) { pe_rsc_trace(rsc, "Not allowing partial migration to continue. %s", rsc->id); allow_migrate = FALSE; } else if (is_moving == FALSE || is_not_set(rsc->flags, pe_rsc_managed) || is_set(rsc->flags, pe_rsc_failed) || is_set(rsc->flags, pe_rsc_start_pending) || (current && current->details->unclean) || rsc->next_role < RSC_ROLE_STARTED) { allow_migrate = FALSE; } if (allow_migrate) { handle_migration_actions(rsc, current, chosen, data_set); } } static void rsc_avoids_remote_nodes(resource_t *rsc) { GHashTableIter iter; node_t *node = NULL; g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&node)) { if (node->details->remote_rsc) { node->weight = -INFINITY; } } } /*! * \internal * \brief Return allowed nodes as (possibly sorted) list * * Convert a resource's hash table of allowed nodes to a list. If printing to * stdout, sort the list, to keep action ID numbers consistent for regression * test output (while avoiding the performance hit on a live cluster). * * \param[in] rsc Resource to check for allowed nodes * \param[in] data_set Cluster working set * * \return List of resource's allowed nodes * \note Callers should take care not to rely on the list being sorted. */ static GList * allowed_nodes_as_list(pe_resource_t *rsc, pe_working_set_t *data_set) { GList *allowed_nodes = NULL; if (rsc->allowed_nodes) { allowed_nodes = g_hash_table_get_values(rsc->allowed_nodes); } if (is_set(data_set->flags, pe_flag_stdout)) { allowed_nodes = g_list_sort(allowed_nodes, sort_node_uname); } return allowed_nodes; } void native_internal_constraints(resource_t * rsc, pe_working_set_t * data_set) { /* This function is on the critical path and worth optimizing as much as possible */ pe_resource_t *top = NULL; GList *allowed_nodes = NULL; bool check_unfencing = FALSE; bool check_utilization = FALSE; if (is_not_set(rsc->flags, pe_rsc_managed)) { pe_rsc_trace(rsc, "Skipping native constraints for unmanaged resource: %s", rsc->id); return; } top = uber_parent(rsc); // Whether resource requires unfencing check_unfencing = is_not_set(rsc->flags, pe_rsc_fence_device) && is_set(data_set->flags, pe_flag_enable_unfencing) && is_set(rsc->flags, pe_rsc_needs_unfencing); // Whether a non-default placement strategy is used check_utilization = (g_hash_table_size(rsc->utilization) > 0) && safe_str_neq(data_set->placement_strategy, "default"); // Order stops before starts (i.e. restart) custom_action_order(rsc, generate_op_key(rsc->id, RSC_STOP, 0), NULL, rsc, generate_op_key(rsc->id, RSC_START, 0), NULL, pe_order_optional | pe_order_implies_then | pe_order_restart, data_set); // Promotable ordering: demote before stop, start before promote if (is_set(top->flags, pe_rsc_promotable) || (rsc->role > RSC_ROLE_SLAVE)) { custom_action_order(rsc, generate_op_key(rsc->id, RSC_DEMOTE, 0), NULL, rsc, generate_op_key(rsc->id, RSC_STOP, 0), NULL, pe_order_implies_first_master, data_set); custom_action_order(rsc, generate_op_key(rsc->id, RSC_START, 0), NULL, rsc, generate_op_key(rsc->id, RSC_PROMOTE, 0), NULL, pe_order_runnable_left, data_set); } // Certain checks need allowed nodes if (check_unfencing || check_utilization || rsc->container) { allowed_nodes = allowed_nodes_as_list(rsc, data_set); } if (check_unfencing) { /* Check if the node needs to be unfenced first */ for (GList *item = allowed_nodes; item; item = item->next) { pe_node_t *node = item->data; pe_action_t *unfence = pe_fence_op(node, "on", TRUE, NULL, data_set); crm_debug("Ordering any stops of %s before %s, and any starts after", rsc->id, unfence->uuid); /* * It would be more efficient to order clone resources once, * rather than order each instance, but ordering the instance * allows us to avoid unnecessary dependencies that might conflict * with user constraints. * * @TODO: This constraint can still produce a transition loop if the * resource has a stop scheduled on the node being unfenced, and * there is a user ordering constraint to start some other resource * (which will be ordered after the unfence) before stopping this * resource. An example is "start some slow-starting cloned service * before stopping an associated virtual IP that may be moving to * it": * stop this -> unfencing -> start that -> stop this */ custom_action_order(rsc, stop_key(rsc), NULL, NULL, strdup(unfence->uuid), unfence, pe_order_optional|pe_order_same_node, data_set); custom_action_order(NULL, strdup(unfence->uuid), unfence, rsc, start_key(rsc), NULL, pe_order_implies_then_on_node|pe_order_same_node, data_set); } } if (check_utilization) { GListPtr gIter = NULL; pe_rsc_trace(rsc, "Creating utilization constraints for %s - strategy: %s", rsc->id, data_set->placement_strategy); for (gIter = rsc->running_on; gIter != NULL; gIter = gIter->next) { node_t *current = (node_t *) gIter->data; char *load_stopped_task = crm_concat(LOAD_STOPPED, current->details->uname, '_'); action_t *load_stopped = get_pseudo_op(load_stopped_task, data_set); if (load_stopped->node == NULL) { load_stopped->node = node_copy(current); update_action_flags(load_stopped, pe_action_optional | pe_action_clear, __FUNCTION__, __LINE__); } custom_action_order(rsc, stop_key(rsc), NULL, NULL, load_stopped_task, load_stopped, pe_order_load, data_set); } for (GList *item = allowed_nodes; item; item = item->next) { pe_node_t *next = item->data; char *load_stopped_task = crm_concat(LOAD_STOPPED, next->details->uname, '_'); action_t *load_stopped = get_pseudo_op(load_stopped_task, data_set); if (load_stopped->node == NULL) { load_stopped->node = node_copy(next); update_action_flags(load_stopped, pe_action_optional | pe_action_clear, __FUNCTION__, __LINE__); } custom_action_order(NULL, strdup(load_stopped_task), load_stopped, rsc, start_key(rsc), NULL, pe_order_load, data_set); custom_action_order(NULL, strdup(load_stopped_task), load_stopped, rsc, generate_op_key(rsc->id, RSC_MIGRATE, 0), NULL, pe_order_load, data_set); free(load_stopped_task); } } if (rsc->container) { resource_t *remote_rsc = NULL; if (rsc->is_remote_node) { // rsc is the implicit remote connection for a guest or bundle node /* Do not allow a guest resource to live on a Pacemaker Remote node, * to avoid nesting remotes. However, allow bundles to run on remote * nodes. */ if (is_not_set(rsc->flags, pe_rsc_allow_remote_remotes)) { rsc_avoids_remote_nodes(rsc->container); } /* If someone cleans up a guest or bundle node's container, we will * likely schedule a (re-)probe of the container and recovery of the * connection. Order the connection stop after the container probe, * so that if we detect the container running, we will trigger a new * transition and avoid the unnecessary recovery. */ new_rsc_order(rsc->container, RSC_STATUS, rsc, RSC_STOP, pe_order_optional, data_set); /* A user can specify that a resource must start on a Pacemaker Remote * node by explicitly configuring it with the container=NODENAME * meta-attribute. This is of questionable merit, since location * constraints can accomplish the same thing. But we support it, so here * we check whether a resource (that is not itself a remote connection) * has container set to a remote node or guest node resource. */ } else if (rsc->container->is_remote_node) { remote_rsc = rsc->container; } else { remote_rsc = pe__resource_contains_guest_node(data_set, rsc->container); } if (remote_rsc) { /* Force the resource on the Pacemaker Remote node instead of * colocating the resource with the container resource. */ for (GList *item = allowed_nodes; item; item = item->next) { pe_node_t *node = item->data; if (node->details->remote_rsc != remote_rsc) { node->weight = -INFINITY; } } } else { /* This resource is either a filler for a container that does NOT * represent a Pacemaker Remote node, or a Pacemaker Remote * connection resource for a guest node or bundle. */ int score; crm_trace("Order and colocate %s relative to its container %s", rsc->id, rsc->container->id); custom_action_order(rsc->container, generate_op_key(rsc->container->id, RSC_START, 0), NULL, rsc, generate_op_key(rsc->id, RSC_START, 0), NULL, pe_order_implies_then | pe_order_runnable_left, data_set); custom_action_order(rsc, generate_op_key(rsc->id, RSC_STOP, 0), NULL, rsc->container, generate_op_key(rsc->container->id, RSC_STOP, 0), NULL, pe_order_implies_first, data_set); if (is_set(rsc->flags, pe_rsc_allow_remote_remotes)) { score = 10000; /* Highly preferred but not essential */ } else { score = INFINITY; /* Force them to run on the same host */ } rsc_colocation_new("resource-with-container", NULL, score, rsc, rsc->container, NULL, NULL, data_set); } } if (rsc->is_remote_node || is_set(rsc->flags, pe_rsc_fence_device)) { /* don't allow remote nodes to run stonith devices * or remote connection resources.*/ rsc_avoids_remote_nodes(rsc); } g_list_free(allowed_nodes); } void native_rsc_colocation_lh(pe_resource_t *rsc_lh, pe_resource_t *rsc_rh, rsc_colocation_t *constraint, pe_working_set_t *data_set) { if (rsc_lh == NULL) { pe_err("rsc_lh was NULL for %s", constraint->id); return; } else if (constraint->rsc_rh == NULL) { pe_err("rsc_rh was NULL for %s", constraint->id); return; } pe_rsc_trace(rsc_lh, "Processing colocation constraint between %s and %s", rsc_lh->id, rsc_rh->id); rsc_rh->cmds->rsc_colocation_rh(rsc_lh, rsc_rh, constraint, data_set); } enum filter_colocation_res filter_colocation_constraint(resource_t * rsc_lh, resource_t * rsc_rh, rsc_colocation_t * constraint, gboolean preview) { if (constraint->score == 0) { return influence_nothing; } /* rh side must be allocated before we can process constraint */ if (preview == FALSE && is_set(rsc_rh->flags, pe_rsc_provisional)) { return influence_nothing; } if ((constraint->role_lh >= RSC_ROLE_SLAVE) && rsc_lh->parent && is_set(rsc_lh->parent->flags, pe_rsc_promotable) && is_not_set(rsc_lh->flags, pe_rsc_provisional)) { /* LH and RH resources have already been allocated, place the correct * priority on LH rsc for the given promotable clone resource role */ return influence_rsc_priority; } if (preview == FALSE && is_not_set(rsc_lh->flags, pe_rsc_provisional)) { // Log an error if we violated a mandatory colocation constraint const pe_node_t *rh_node = rsc_rh->allocated_to; if (rsc_lh->allocated_to == NULL) { // Dependent resource isn't allocated, so constraint doesn't matter return influence_nothing; } if (constraint->score >= INFINITY) { // Dependent resource must colocate with rh_node if ((rh_node == NULL) || (rh_node->details != rsc_lh->allocated_to->details)) { crm_err("%s must be colocated with %s but is not (%s vs. %s)", rsc_lh->id, rsc_rh->id, rsc_lh->allocated_to->details->uname, (rh_node? rh_node->details->uname : "unallocated")); } } else if (constraint->score <= -INFINITY) { // Dependent resource must anti-colocate with rh_node if ((rh_node != NULL) && (rsc_lh->allocated_to->details == rh_node->details)) { crm_err("%s and %s must be anti-colocated but are allocated " "to the same node (%s)", rsc_lh->id, rsc_rh->id, rh_node->details->uname); } } return influence_nothing; } if (constraint->score > 0 && constraint->role_lh != RSC_ROLE_UNKNOWN && constraint->role_lh != rsc_lh->next_role) { crm_trace("LH: Skipping constraint: \"%s\" state filter nextrole is %s", role2text(constraint->role_lh), role2text(rsc_lh->next_role)); return influence_nothing; } if (constraint->score > 0 && constraint->role_rh != RSC_ROLE_UNKNOWN && constraint->role_rh != rsc_rh->next_role) { crm_trace("RH: Skipping constraint: \"%s\" state filter", role2text(constraint->role_rh)); return FALSE; } if (constraint->score < 0 && constraint->role_lh != RSC_ROLE_UNKNOWN && constraint->role_lh == rsc_lh->next_role) { crm_trace("LH: Skipping negative constraint: \"%s\" state filter", role2text(constraint->role_lh)); return influence_nothing; } if (constraint->score < 0 && constraint->role_rh != RSC_ROLE_UNKNOWN && constraint->role_rh == rsc_rh->next_role) { crm_trace("RH: Skipping negative constraint: \"%s\" state filter", role2text(constraint->role_rh)); return influence_nothing; } return influence_rsc_location; } static void influence_priority(resource_t * rsc_lh, resource_t * rsc_rh, rsc_colocation_t * constraint) { const char *rh_value = NULL; const char *lh_value = NULL; const char *attribute = CRM_ATTR_ID; int score_multiplier = 1; if (constraint->node_attribute != NULL) { attribute = constraint->node_attribute; } if (!rsc_rh->allocated_to || !rsc_lh->allocated_to) { return; } lh_value = pe_node_attribute_raw(rsc_lh->allocated_to, attribute); rh_value = pe_node_attribute_raw(rsc_rh->allocated_to, attribute); if (!safe_str_eq(lh_value, rh_value)) { if(constraint->score == INFINITY && constraint->role_lh == RSC_ROLE_MASTER) { rsc_lh->priority = -INFINITY; } return; } if (constraint->role_rh && (constraint->role_rh != rsc_rh->next_role)) { return; } if (constraint->role_lh == RSC_ROLE_SLAVE) { score_multiplier = -1; } rsc_lh->priority = merge_weights(score_multiplier * constraint->score, rsc_lh->priority); } static void colocation_match(resource_t * rsc_lh, resource_t * rsc_rh, rsc_colocation_t * constraint) { const char *tmp = NULL; const char *value = NULL; const char *attribute = CRM_ATTR_ID; GHashTable *work = NULL; gboolean do_check = FALSE; GHashTableIter iter; node_t *node = NULL; if (constraint->node_attribute != NULL) { attribute = constraint->node_attribute; } if (rsc_rh->allocated_to) { value = pe_node_attribute_raw(rsc_rh->allocated_to, attribute); do_check = TRUE; } else if (constraint->score < 0) { /* nothing to do: * anti-colocation with something that is not running */ return; } work = node_hash_dup(rsc_lh->allowed_nodes); g_hash_table_iter_init(&iter, work); while (g_hash_table_iter_next(&iter, NULL, (void **)&node)) { tmp = pe_node_attribute_raw(node, attribute); if (do_check && safe_str_eq(tmp, value)) { if (constraint->score < INFINITY) { pe_rsc_trace(rsc_lh, "%s: %s.%s += %d", constraint->id, rsc_lh->id, node->details->uname, constraint->score); node->weight = merge_weights(constraint->score, node->weight); } } else if (do_check == FALSE || constraint->score >= INFINITY) { pe_rsc_trace(rsc_lh, "%s: %s.%s -= %d (%s)", constraint->id, rsc_lh->id, node->details->uname, constraint->score, do_check ? "failed" : "unallocated"); node->weight = merge_weights(-constraint->score, node->weight); } } if (can_run_any(work) || constraint->score <= -INFINITY || constraint->score >= INFINITY) { g_hash_table_destroy(rsc_lh->allowed_nodes); rsc_lh->allowed_nodes = work; work = NULL; } else { static char score[33]; score2char_stack(constraint->score, score, sizeof(score)); pe_rsc_info(rsc_lh, "%s: Rolling back scores from %s (%d, %s)", rsc_lh->id, rsc_rh->id, do_check, score); } if (work) { g_hash_table_destroy(work); } } void native_rsc_colocation_rh(pe_resource_t *rsc_lh, pe_resource_t *rsc_rh, rsc_colocation_t *constraint, pe_working_set_t *data_set) { enum filter_colocation_res filter_results; CRM_ASSERT(rsc_lh); CRM_ASSERT(rsc_rh); filter_results = filter_colocation_constraint(rsc_lh, rsc_rh, constraint, FALSE); pe_rsc_trace(rsc_lh, "%sColocating %s with %s (%s, weight=%d, filter=%d)", constraint->score >= 0 ? "" : "Anti-", rsc_lh->id, rsc_rh->id, constraint->id, constraint->score, filter_results); switch (filter_results) { case influence_rsc_priority: influence_priority(rsc_lh, rsc_rh, constraint); break; case influence_rsc_location: pe_rsc_trace(rsc_lh, "%sColocating %s with %s (%s, weight=%d)", constraint->score >= 0 ? "" : "Anti-", rsc_lh->id, rsc_rh->id, constraint->id, constraint->score); colocation_match(rsc_lh, rsc_rh, constraint); break; case influence_nothing: default: return; } } static gboolean filter_rsc_ticket(resource_t * rsc_lh, rsc_ticket_t * rsc_ticket) { if (rsc_ticket->role_lh != RSC_ROLE_UNKNOWN && rsc_ticket->role_lh != rsc_lh->role) { pe_rsc_trace(rsc_lh, "LH: Skipping constraint: \"%s\" state filter", role2text(rsc_ticket->role_lh)); return FALSE; } return TRUE; } void rsc_ticket_constraint(resource_t * rsc_lh, rsc_ticket_t * rsc_ticket, pe_working_set_t * data_set) { if (rsc_ticket == NULL) { pe_err("rsc_ticket was NULL"); return; } if (rsc_lh == NULL) { pe_err("rsc_lh was NULL for %s", rsc_ticket->id); return; } if (rsc_ticket->ticket->granted && rsc_ticket->ticket->standby == FALSE) { return; } if (rsc_lh->children) { GListPtr gIter = rsc_lh->children; pe_rsc_trace(rsc_lh, "Processing ticket dependencies from %s", rsc_lh->id); for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; rsc_ticket_constraint(child_rsc, rsc_ticket, data_set); } return; } pe_rsc_trace(rsc_lh, "%s: Processing ticket dependency on %s (%s, %s)", rsc_lh->id, rsc_ticket->ticket->id, rsc_ticket->id, role2text(rsc_ticket->role_lh)); if ((rsc_ticket->ticket->granted == FALSE) && (rsc_lh->running_on != NULL)) { GListPtr gIter = NULL; switch (rsc_ticket->loss_policy) { case loss_ticket_stop: resource_location(rsc_lh, NULL, -INFINITY, "__loss_of_ticket__", data_set); break; case loss_ticket_demote: // Promotion score will be set to -INFINITY in promotion_order() if (rsc_ticket->role_lh != RSC_ROLE_MASTER) { resource_location(rsc_lh, NULL, -INFINITY, "__loss_of_ticket__", data_set); } break; case loss_ticket_fence: if (filter_rsc_ticket(rsc_lh, rsc_ticket) == FALSE) { return; } resource_location(rsc_lh, NULL, -INFINITY, "__loss_of_ticket__", data_set); for (gIter = rsc_lh->running_on; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; pe_fence_node(data_set, node, "deadman ticket was lost"); } break; case loss_ticket_freeze: if (filter_rsc_ticket(rsc_lh, rsc_ticket) == FALSE) { return; } if (rsc_lh->running_on != NULL) { clear_bit(rsc_lh->flags, pe_rsc_managed); set_bit(rsc_lh->flags, pe_rsc_block); } break; } } else if (rsc_ticket->ticket->granted == FALSE) { if (rsc_ticket->role_lh != RSC_ROLE_MASTER || rsc_ticket->loss_policy == loss_ticket_stop) { resource_location(rsc_lh, NULL, -INFINITY, "__no_ticket__", data_set); } } else if (rsc_ticket->ticket->standby) { if (rsc_ticket->role_lh != RSC_ROLE_MASTER || rsc_ticket->loss_policy == loss_ticket_stop) { resource_location(rsc_lh, NULL, -INFINITY, "__ticket_standby__", data_set); } } } enum pe_action_flags native_action_flags(action_t * action, node_t * node) { return action->flags; } static inline bool is_primitive_action(pe_action_t *action) { return action && action->rsc && (action->rsc->variant == pe_native); } /*! * \internal * \brief Set action bits appropriately when pe_restart_order is used * * \param[in] first 'First' action in an ordering with pe_restart_order * \param[in] then 'Then' action in an ordering with pe_restart_order * \param[in] filter What ordering flags to care about * * \note pe_restart_order is set for "stop resource before starting it" and * "stop later group member before stopping earlier group member" */ static void handle_restart_ordering(pe_action_t *first, pe_action_t *then, enum pe_action_flags filter) { const char *reason = NULL; CRM_ASSERT(is_primitive_action(first)); CRM_ASSERT(is_primitive_action(then)); // We need to update the action in two cases: // ... if 'then' is required if (is_set(filter, pe_action_optional) && is_not_set(then->flags, pe_action_optional)) { reason = "restart"; } /* ... if 'then' is unrunnable start of managed resource (if a resource * should restart but can't start, we still want to stop) */ if (is_set(filter, pe_action_runnable) && is_not_set(then->flags, pe_action_runnable) && is_set(then->rsc->flags, pe_rsc_managed) && safe_str_eq(then->task, RSC_START)) { reason = "stop"; } if (reason == NULL) { return; } pe_rsc_trace(first->rsc, "Handling %s -> %s for %s", first->uuid, then->uuid, reason); // Make 'first' required if it is runnable if (is_set(first->flags, pe_action_runnable)) { pe_action_implies(first, then, pe_action_optional); } // Make 'first' required if 'then' is required if (is_not_set(then->flags, pe_action_optional)) { pe_action_implies(first, then, pe_action_optional); } // Make 'first' unmigratable if 'then' is unmigratable if (is_not_set(then->flags, pe_action_migrate_runnable)) { pe_action_implies(first, then, pe_action_migrate_runnable); } // Make 'then' unrunnable if 'first' is required but unrunnable if (is_not_set(first->flags, pe_action_optional) && is_not_set(first->flags, pe_action_runnable)) { pe_action_implies(then, first, pe_action_runnable); } } enum pe_graph_flags native_update_actions(pe_action_t *first, pe_action_t *then, pe_node_t *node, enum pe_action_flags flags, enum pe_action_flags filter, enum pe_ordering type, pe_working_set_t *data_set) { /* flags == get_action_flags(first, then_node) called from update_action() */ enum pe_graph_flags changed = pe_graph_none; enum pe_action_flags then_flags = then->flags; enum pe_action_flags first_flags = first->flags; crm_trace( "Testing %s on %s (0x%.6x) with %s 0x%.6x", first->uuid, first->node ? first->node->details->uname : "[none]", first->flags, then->uuid, then->flags); if (type & pe_order_asymmetrical) { resource_t *then_rsc = then->rsc; enum rsc_role_e then_rsc_role = then_rsc ? then_rsc->fns->state(then_rsc, TRUE) : 0; if (!then_rsc) { /* ignore */ } else if ((then_rsc_role == RSC_ROLE_STOPPED) && safe_str_eq(then->task, RSC_STOP)) { /* ignore... if 'then' is supposed to be stopped after 'first', but * then is already stopped, there is nothing to be done when non-symmetrical. */ } else if ((then_rsc_role >= RSC_ROLE_STARTED) && safe_str_eq(then->task, RSC_START) && is_set(then->flags, pe_action_optional) && then->node && g_list_length(then_rsc->running_on) == 1 && then->node->details == ((node_t *) then_rsc->running_on->data)->details) { /* Ignore. If 'then' is supposed to be started after 'first', but * 'then' is already started, there is nothing to be done when * asymmetrical -- unless the start is mandatory, which indicates * the resource is restarting, and the ordering is still needed. */ } else if (!(first->flags & pe_action_runnable)) { /* prevent 'then' action from happening if 'first' is not runnable and * 'then' has not yet occurred. */ pe_action_implies(then, first, pe_action_optional); pe_action_implies(then, first, pe_action_runnable); pe_rsc_trace(then->rsc, "Unset optional and runnable on %s", then->uuid); } else { /* ignore... then is allowed to start/stop if it wants to. */ } } if (type & pe_order_implies_first) { if (is_set(filter, pe_action_optional) && is_not_set(flags /* Should be then_flags? */, pe_action_optional)) { // Needs is_set(first_flags, pe_action_optional) too? pe_rsc_trace(first->rsc, "Unset optional on %s because of %s", first->uuid, then->uuid); pe_action_implies(first, then, pe_action_optional); } if (is_set(flags, pe_action_migrate_runnable) && is_set(then->flags, pe_action_migrate_runnable) == FALSE && is_set(then->flags, pe_action_optional) == FALSE) { pe_rsc_trace(first->rsc, "Unset migrate runnable on %s because of %s", first->uuid, then->uuid); pe_action_implies(first, then, pe_action_migrate_runnable); } } if (type & pe_order_implies_first_master) { if ((filter & pe_action_optional) && ((then->flags & pe_action_optional) == FALSE) && then->rsc && (then->rsc->role == RSC_ROLE_MASTER)) { pe_action_implies(first, then, pe_action_optional); if (is_set(first->flags, pe_action_migrate_runnable) && is_set(then->flags, pe_action_migrate_runnable) == FALSE) { pe_rsc_trace(first->rsc, "Unset migrate runnable on %s because of %s", first->uuid, then->uuid); pe_action_implies(first, then, pe_action_migrate_runnable); } pe_rsc_trace(then->rsc, "Unset optional on %s because of %s", first->uuid, then->uuid); } } if ((type & pe_order_implies_first_migratable) && is_set(filter, pe_action_optional)) { if (((then->flags & pe_action_migrate_runnable) == FALSE) || ((then->flags & pe_action_runnable) == FALSE)) { pe_rsc_trace(then->rsc, "Unset runnable on %s because %s is neither runnable or migratable", first->uuid, then->uuid); pe_action_implies(first, then, pe_action_runnable); } if ((then->flags & pe_action_optional) == 0) { pe_rsc_trace(then->rsc, "Unset optional on %s because %s is not optional", first->uuid, then->uuid); pe_action_implies(first, then, pe_action_optional); } } if ((type & pe_order_pseudo_left) && is_set(filter, pe_action_optional)) { if ((first->flags & pe_action_runnable) == FALSE) { pe_action_implies(then, first, pe_action_migrate_runnable); pe_clear_action_bit(then, pe_action_pseudo); pe_rsc_trace(then->rsc, "Unset pseudo on %s because %s is not runnable", then->uuid, first->uuid); } } if (is_set(type, pe_order_runnable_left) && is_set(filter, pe_action_runnable) && is_set(then->flags, pe_action_runnable) && is_set(flags, pe_action_runnable) == FALSE) { pe_rsc_trace(then->rsc, "Unset runnable on %s because of %s", then->uuid, first->uuid); pe_action_implies(then, first, pe_action_runnable); pe_action_implies(then, first, pe_action_migrate_runnable); } if (is_set(type, pe_order_implies_then) && is_set(filter, pe_action_optional) && is_set(then->flags, pe_action_optional) && is_set(flags, pe_action_optional) == FALSE) { /* in this case, treat migrate_runnable as if first is optional */ if (is_set(first->flags, pe_action_migrate_runnable) == FALSE) { pe_rsc_trace(then->rsc, "Unset optional on %s because of %s", then->uuid, first->uuid); pe_action_implies(then, first, pe_action_optional); } } if (is_set(type, pe_order_restart)) { handle_restart_ordering(first, then, filter); } if (then_flags != then->flags) { changed |= pe_graph_updated_then; pe_rsc_trace(then->rsc, "Then: Flags for %s on %s are now 0x%.6x (was 0x%.6x) because of %s 0x%.6x", then->uuid, then->node ? then->node->details->uname : "[none]", then->flags, then_flags, first->uuid, first->flags); if(then->rsc && then->rsc->parent) { /* "X_stop then X_start" doesn't get handled for cloned groups unless we do this */ update_action(then, data_set); } } if (first_flags != first->flags) { changed |= pe_graph_updated_first; pe_rsc_trace(first->rsc, "First: Flags for %s on %s are now 0x%.6x (was 0x%.6x) because of %s 0x%.6x", first->uuid, first->node ? first->node->details->uname : "[none]", first->flags, first_flags, then->uuid, then->flags); } return changed; } void native_rsc_location(pe_resource_t *rsc, pe__location_t *constraint) { GListPtr gIter = NULL; GHashTableIter iter; node_t *node = NULL; if (constraint == NULL) { pe_err("Constraint is NULL"); return; } else if (rsc == NULL) { pe_err("LHS of rsc_to_node (%s) is NULL", constraint->id); return; } pe_rsc_trace(rsc, "Applying %s (%s) to %s", constraint->id, role2text(constraint->role_filter), rsc->id); /* take "lifetime" into account */ if (constraint->role_filter > RSC_ROLE_UNKNOWN && constraint->role_filter != rsc->next_role) { pe_rsc_debug(rsc, "Constraint (%s) is not active (role : %s vs. %s)", constraint->id, role2text(constraint->role_filter), role2text(rsc->next_role)); return; } if (constraint->node_list_rh == NULL) { pe_rsc_trace(rsc, "RHS of constraint %s is NULL", constraint->id); return; } for (gIter = constraint->node_list_rh; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; node_t *other_node = NULL; other_node = (node_t *) pe_hash_table_lookup(rsc->allowed_nodes, node->details->id); if (other_node != NULL) { pe_rsc_trace(rsc, "%s + %s: %d + %d", node->details->uname, other_node->details->uname, node->weight, other_node->weight); other_node->weight = merge_weights(other_node->weight, node->weight); } else { other_node = node_copy(node); pe_rsc_trace(rsc, "%s: %d (insert %d)", other_node->details->uname, other_node->weight, constraint->discover_mode); g_hash_table_insert(rsc->allowed_nodes, (gpointer) other_node->details->id, other_node); } if (other_node->rsc_discover_mode < constraint->discover_mode) { if (constraint->discover_mode == pe_discover_exclusive) { rsc->exclusive_discover = TRUE; } /* exclusive > never > always... always is default */ other_node->rsc_discover_mode = constraint->discover_mode; } } g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&node)) { pe_rsc_trace(rsc, "%s + %s : %d", rsc->id, node->details->uname, node->weight); } } void native_expand(resource_t * rsc, pe_working_set_t * data_set) { GListPtr gIter = NULL; CRM_ASSERT(rsc); pe_rsc_trace(rsc, "Processing actions from %s", rsc->id); for (gIter = rsc->actions; gIter != NULL; gIter = gIter->next) { action_t *action = (action_t *) gIter->data; crm_trace("processing action %d for rsc=%s", action->id, rsc->id); graph_element_from_action(action, data_set); } for (gIter = rsc->children; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; child_rsc->cmds->expand(child_rsc, data_set); } } #define log_change(a, fmt, args...) do { \ if(a && a->reason && terminal) { \ printf(" * "fmt" \tdue to %s\n", ##args, a->reason); \ } else if(a && a->reason) { \ crm_notice(fmt" \tdue to %s", ##args, a->reason); \ } else if(terminal) { \ printf(" * "fmt"\n", ##args); \ } else { \ crm_notice(fmt, ##args); \ } \ } while(0) #define STOP_SANITY_ASSERT(lineno) do { \ if(current && current->details->unclean) { \ /* It will be a pseudo op */ \ } else if(stop == NULL) { \ crm_err("%s:%d: No stop action exists for %s", __FUNCTION__, lineno, rsc->id); \ CRM_ASSERT(stop != NULL); \ } else if(is_set(stop->flags, pe_action_optional)) { \ crm_err("%s:%d: Action %s is still optional", __FUNCTION__, lineno, stop->uuid); \ CRM_ASSERT(is_not_set(stop->flags, pe_action_optional)); \ } \ } while(0) static int rsc_width = 5; static int detail_width = 5; static void LogAction(const char *change, resource_t *rsc, pe_node_t *origin, pe_node_t *destination, pe_action_t *action, pe_action_t *source, gboolean terminal) { int len = 0; char *reason = NULL; char *details = NULL; bool same_host = FALSE; bool same_role = FALSE; bool need_role = FALSE; CRM_ASSERT(action); CRM_ASSERT(destination != NULL || origin != NULL); if(source == NULL) { source = action; } len = strlen(rsc->id); if(len > rsc_width) { rsc_width = len + 2; } if(rsc->role > RSC_ROLE_STARTED || rsc->next_role > RSC_ROLE_SLAVE) { need_role = TRUE; } if(origin != NULL && destination != NULL && origin->details == destination->details) { same_host = TRUE; } if(rsc->role == rsc->next_role) { same_role = TRUE; } if(need_role && origin == NULL) { /* Promoting from Stopped */ details = crm_strdup_printf("%s -> %s %s", role2text(rsc->role), role2text(rsc->next_role), destination->details->uname); } else if(need_role && destination == NULL) { /* Demoting a Master or Stopping a Slave */ details = crm_strdup_printf("%s %s", role2text(rsc->role), origin->details->uname); } else if(origin == NULL || destination == NULL) { /* Starting or stopping a resource */ details = crm_strdup_printf("%s", origin?origin->details->uname:destination->details->uname); } else if(need_role && same_role && same_host) { /* Recovering or restarting a promotable clone resource */ details = crm_strdup_printf("%s %s", role2text(rsc->role), origin->details->uname); } else if(same_role && same_host) { /* Recovering or Restarting a normal resource */ details = crm_strdup_printf("%s", origin->details->uname); } else if(same_role && need_role) { /* Moving a promotable clone resource */ details = crm_strdup_printf("%s -> %s %s", origin->details->uname, destination->details->uname, role2text(rsc->role)); } else if(same_role) { /* Moving a normal resource */ details = crm_strdup_printf("%s -> %s", origin->details->uname, destination->details->uname); } else if(same_host) { /* Promoting or demoting a promotable clone resource */ details = crm_strdup_printf("%s -> %s %s", role2text(rsc->role), role2text(rsc->next_role), origin->details->uname); } else { /* Moving and promoting/demoting */ details = crm_strdup_printf("%s %s -> %s %s", role2text(rsc->role), origin->details->uname, role2text(rsc->next_role), destination->details->uname); } len = strlen(details); if(len > detail_width) { detail_width = len; } if(source->reason && is_not_set(action->flags, pe_action_runnable)) { reason = crm_strdup_printf(" due to %s (blocked)", source->reason); } else if(source->reason) { reason = crm_strdup_printf(" due to %s", source->reason); } else if(is_not_set(action->flags, pe_action_runnable)) { reason = strdup(" blocked"); } else { reason = strdup(""); } if(terminal) { printf(" * %-8s %-*s ( %*s ) %s\n", change, rsc_width, rsc->id, detail_width, details, reason); } else { crm_notice(" * %-8s %-*s ( %*s ) %s", change, rsc_width, rsc->id, detail_width, details, reason); } free(details); free(reason); } void LogActions(resource_t * rsc, pe_working_set_t * data_set, gboolean terminal) { node_t *next = NULL; node_t *current = NULL; pe_node_t *start_node = NULL; action_t *stop = NULL; action_t *start = NULL; action_t *demote = NULL; action_t *promote = NULL; char *key = NULL; gboolean moving = FALSE; GListPtr possible_matches = NULL; if(rsc->variant == pe_container) { pcmk__bundle_log_actions(rsc, data_set, terminal); return; } if (rsc->children) { GListPtr gIter = NULL; for (gIter = rsc->children; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; LogActions(child_rsc, data_set, terminal); } return; } next = rsc->allocated_to; if (rsc->running_on) { current = pe__current_node(rsc); if (rsc->role == RSC_ROLE_STOPPED) { /* * This can occur when resources are being recovered * We fiddle with the current role in native_create_actions() */ rsc->role = RSC_ROLE_STARTED; } } if (current == NULL && is_set(rsc->flags, pe_rsc_orphan)) { /* Don't log stopped orphans */ return; } if (is_not_set(rsc->flags, pe_rsc_managed) || (current == NULL && next == NULL)) { pe_rsc_info(rsc, "Leave %s\t(%s%s)", rsc->id, role2text(rsc->role), is_not_set(rsc->flags, pe_rsc_managed) ? " unmanaged" : ""); return; } if (current != NULL && next != NULL && safe_str_neq(current->details->id, next->details->id)) { moving = TRUE; } possible_matches = pe__resource_actions(rsc, next, RSC_START, FALSE); if (possible_matches) { start = possible_matches->data; g_list_free(possible_matches); } if ((start == NULL) || is_not_set(start->flags, pe_action_runnable)) { start_node = NULL; } else { start_node = current; } possible_matches = pe__resource_actions(rsc, start_node, RSC_STOP, FALSE); if (possible_matches) { stop = possible_matches->data; g_list_free(possible_matches); } possible_matches = pe__resource_actions(rsc, next, RSC_PROMOTE, FALSE); if (possible_matches) { promote = possible_matches->data; g_list_free(possible_matches); } possible_matches = pe__resource_actions(rsc, next, RSC_DEMOTE, FALSE); if (possible_matches) { demote = possible_matches->data; g_list_free(possible_matches); } if (rsc->role == rsc->next_role) { action_t *migrate_op = NULL; possible_matches = pe__resource_actions(rsc, next, RSC_MIGRATED, FALSE); if (possible_matches) { migrate_op = possible_matches->data; } CRM_CHECK(next != NULL,); if (next == NULL) { } else if (migrate_op && is_set(migrate_op->flags, pe_action_runnable) && current) { LogAction("Migrate", rsc, current, next, start, NULL, terminal); } else if (is_set(rsc->flags, pe_rsc_reload)) { LogAction("Reload", rsc, current, next, start, NULL, terminal); } else if (start == NULL || is_set(start->flags, pe_action_optional)) { pe_rsc_info(rsc, "Leave %s\t(%s %s)", rsc->id, role2text(rsc->role), next->details->uname); } else if (start && is_set(start->flags, pe_action_runnable) == FALSE) { LogAction("Stop", rsc, current, NULL, stop, (stop && stop->reason)? stop : start, terminal); STOP_SANITY_ASSERT(__LINE__); } else if (moving && current) { LogAction(is_set(rsc->flags, pe_rsc_failed) ? "Recover" : "Move", rsc, current, next, stop, NULL, terminal); } else if (is_set(rsc->flags, pe_rsc_failed)) { LogAction("Recover", rsc, current, NULL, stop, NULL, terminal); STOP_SANITY_ASSERT(__LINE__); } else { LogAction("Restart", rsc, current, next, start, NULL, terminal); /* STOP_SANITY_ASSERT(__LINE__); False positive for migrate-fail-7 */ } g_list_free(possible_matches); return; } if(stop && (rsc->next_role == RSC_ROLE_STOPPED || (start && is_not_set(start->flags, pe_action_runnable)))) { GListPtr gIter = NULL; key = stop_key(rsc); for (gIter = rsc->running_on; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; action_t *stop_op = NULL; possible_matches = find_actions(rsc->actions, key, node); if (possible_matches) { stop_op = possible_matches->data; g_list_free(possible_matches); } if (stop_op && (stop_op->flags & pe_action_runnable)) { STOP_SANITY_ASSERT(__LINE__); } LogAction("Stop", rsc, node, NULL, stop_op, (stop_op && stop_op->reason)? stop_op : start, terminal); } free(key); } else if (stop && is_set(rsc->flags, pe_rsc_failed)) { /* 'stop' may be NULL if the failure was ignored */ LogAction("Recover", rsc, current, next, stop, start, terminal); STOP_SANITY_ASSERT(__LINE__); } else if (moving) { LogAction("Move", rsc, current, next, stop, NULL, terminal); STOP_SANITY_ASSERT(__LINE__); } else if (is_set(rsc->flags, pe_rsc_reload)) { LogAction("Reload", rsc, current, next, start, NULL, terminal); } else if (stop != NULL && is_not_set(stop->flags, pe_action_optional)) { LogAction("Restart", rsc, current, next, start, NULL, terminal); STOP_SANITY_ASSERT(__LINE__); } else if (rsc->role == RSC_ROLE_MASTER) { CRM_LOG_ASSERT(current != NULL); LogAction("Demote", rsc, current, next, demote, NULL, terminal); } else if(rsc->next_role == RSC_ROLE_MASTER) { CRM_LOG_ASSERT(next); LogAction("Promote", rsc, current, next, promote, NULL, terminal); } else if (rsc->role == RSC_ROLE_STOPPED && rsc->next_role > RSC_ROLE_STOPPED) { LogAction("Start", rsc, current, next, start, NULL, terminal); } } gboolean StopRsc(resource_t * rsc, node_t * next, gboolean optional, pe_working_set_t * data_set) { GListPtr gIter = NULL; CRM_ASSERT(rsc); pe_rsc_trace(rsc, "%s", rsc->id); for (gIter = rsc->running_on; gIter != NULL; gIter = gIter->next) { node_t *current = (node_t *) gIter->data; action_t *stop; if (rsc->partial_migration_target) { if (rsc->partial_migration_target->details == current->details) { pe_rsc_trace(rsc, "Filtered %s -> %s %s", current->details->uname, next->details->uname, rsc->id); continue; } else { pe_rsc_trace(rsc, "Forced on %s %s", current->details->uname, rsc->id); optional = FALSE; } } pe_rsc_trace(rsc, "%s on %s", rsc->id, current->details->uname); stop = stop_action(rsc, current, optional); if(rsc->allocated_to == NULL) { pe_action_set_reason(stop, "node availability", TRUE); } if (is_not_set(rsc->flags, pe_rsc_managed)) { update_action_flags(stop, pe_action_runnable | pe_action_clear, __FUNCTION__, __LINE__); } if (is_set(data_set->flags, pe_flag_remove_after_stop)) { DeleteRsc(rsc, current, optional, data_set); } if(is_set(rsc->flags, pe_rsc_needs_unfencing)) { action_t *unfence = pe_fence_op(current, "on", TRUE, NULL, data_set); order_actions(stop, unfence, pe_order_implies_first); if (!node_has_been_unfenced(current)) { pe_proc_err("Stopping %s until %s can be unfenced", rsc->id, current->details->uname); } } } return TRUE; } static void order_after_unfencing(resource_t *rsc, pe_node_t *node, action_t *action, enum pe_ordering order, pe_working_set_t *data_set) { /* When unfencing is in use, we order unfence actions before any probe or * start of resources that require unfencing, and also of fence devices. * * This might seem to violate the principle that fence devices require * only quorum. However, fence agents that unfence often don't have enough * information to even probe or start unless the node is first unfenced. */ if (is_unfence_device(rsc, data_set) || is_set(rsc->flags, pe_rsc_needs_unfencing)) { /* Start with an optional ordering. Requiring unfencing would result in * the node being unfenced, and all its resources being stopped, * whenever a new resource is added -- which would be highly suboptimal. */ action_t *unfence = pe_fence_op(node, "on", TRUE, NULL, data_set); order_actions(unfence, action, order); if (!node_has_been_unfenced(node)) { // But unfencing is required if it has never been done char *reason = crm_strdup_printf("required by %s %s", rsc->id, action->task); trigger_unfencing(NULL, node, reason, NULL, data_set); free(reason); } } } gboolean StartRsc(resource_t * rsc, node_t * next, gboolean optional, pe_working_set_t * data_set) { action_t *start = NULL; CRM_ASSERT(rsc); pe_rsc_trace(rsc, "%s on %s %d %d", rsc->id, next ? next->details->uname : "N/A", optional, next ? next->weight : 0); start = start_action(rsc, next, TRUE); order_after_unfencing(rsc, next, start, pe_order_implies_then, data_set); if (is_set(start->flags, pe_action_runnable) && optional == FALSE) { update_action_flags(start, pe_action_optional | pe_action_clear, __FUNCTION__, __LINE__); } return TRUE; } gboolean PromoteRsc(resource_t * rsc, node_t * next, gboolean optional, pe_working_set_t * data_set) { GListPtr gIter = NULL; gboolean runnable = TRUE; GListPtr action_list = NULL; CRM_ASSERT(rsc); CRM_CHECK(next != NULL, return FALSE); pe_rsc_trace(rsc, "%s on %s", rsc->id, next->details->uname); action_list = pe__resource_actions(rsc, next, RSC_START, TRUE); for (gIter = action_list; gIter != NULL; gIter = gIter->next) { action_t *start = (action_t *) gIter->data; if (is_set(start->flags, pe_action_runnable) == FALSE) { runnable = FALSE; } } g_list_free(action_list); if (runnable) { promote_action(rsc, next, optional); return TRUE; } pe_rsc_debug(rsc, "%s\tPromote %s (canceled)", next->details->uname, rsc->id); action_list = pe__resource_actions(rsc, next, RSC_PROMOTE, TRUE); for (gIter = action_list; gIter != NULL; gIter = gIter->next) { action_t *promote = (action_t *) gIter->data; update_action_flags(promote, pe_action_runnable | pe_action_clear, __FUNCTION__, __LINE__); } g_list_free(action_list); return TRUE; } gboolean DemoteRsc(resource_t * rsc, node_t * next, gboolean optional, pe_working_set_t * data_set) { GListPtr gIter = NULL; CRM_ASSERT(rsc); pe_rsc_trace(rsc, "%s", rsc->id); /* CRM_CHECK(rsc->next_role == RSC_ROLE_SLAVE, return FALSE); */ for (gIter = rsc->running_on; gIter != NULL; gIter = gIter->next) { node_t *current = (node_t *) gIter->data; pe_rsc_trace(rsc, "%s on %s", rsc->id, next ? next->details->uname : "N/A"); demote_action(rsc, current, optional); } return TRUE; } gboolean RoleError(resource_t * rsc, node_t * next, gboolean optional, pe_working_set_t * data_set) { CRM_ASSERT(rsc); crm_err("%s on %s", rsc->id, next ? next->details->uname : "N/A"); CRM_CHECK(FALSE, return FALSE); return FALSE; } gboolean NullOp(resource_t * rsc, node_t * next, gboolean optional, pe_working_set_t * data_set) { CRM_ASSERT(rsc); pe_rsc_trace(rsc, "%s", rsc->id); return FALSE; } gboolean DeleteRsc(resource_t * rsc, node_t * node, gboolean optional, pe_working_set_t * data_set) { if (is_set(rsc->flags, pe_rsc_failed)) { pe_rsc_trace(rsc, "Resource %s not deleted from %s: failed", rsc->id, node->details->uname); return FALSE; } else if (node == NULL) { pe_rsc_trace(rsc, "Resource %s not deleted: NULL node", rsc->id); return FALSE; } else if (node->details->unclean || node->details->online == FALSE) { pe_rsc_trace(rsc, "Resource %s not deleted from %s: unrunnable", rsc->id, node->details->uname); return FALSE; } crm_notice("Removing %s from %s", rsc->id, node->details->uname); delete_action(rsc, node, optional); new_rsc_order(rsc, RSC_STOP, rsc, RSC_DELETE, optional ? pe_order_implies_then : pe_order_optional, data_set); new_rsc_order(rsc, RSC_DELETE, rsc, RSC_START, optional ? pe_order_implies_then : pe_order_optional, data_set); return TRUE; } gboolean native_create_probe(resource_t * rsc, node_t * node, action_t * complete, gboolean force, pe_working_set_t * data_set) { enum pe_ordering flags = pe_order_optional; char *key = NULL; action_t *probe = NULL; node_t *running = NULL; node_t *allowed = NULL; resource_t *top = uber_parent(rsc); static const char *rc_master = NULL; static const char *rc_inactive = NULL; if (rc_inactive == NULL) { rc_inactive = crm_itoa(PCMK_OCF_NOT_RUNNING); rc_master = crm_itoa(PCMK_OCF_RUNNING_MASTER); } CRM_CHECK(node != NULL, return FALSE); if (force == FALSE && is_not_set(data_set->flags, pe_flag_startup_probes)) { pe_rsc_trace(rsc, "Skipping active resource detection for %s", rsc->id); return FALSE; } if (pe__is_guest_or_remote_node(node)) { const char *class = crm_element_value(rsc->xml, XML_AGENT_ATTR_CLASS); if (safe_str_eq(class, PCMK_RESOURCE_CLASS_STONITH)) { pe_rsc_trace(rsc, "Skipping probe for %s on %s because Pacemaker Remote nodes cannot run stonith agents", rsc->id, node->details->id); return FALSE; } else if (pe__is_guest_node(node) && pe__resource_contains_guest_node(data_set, rsc)) { pe_rsc_trace(rsc, "Skipping probe for %s on %s because guest nodes cannot run resources containing guest nodes", rsc->id, node->details->id); return FALSE; } else if (rsc->is_remote_node) { pe_rsc_trace(rsc, "Skipping probe for %s on %s because Pacemaker Remote nodes cannot host remote connections", rsc->id, node->details->id); return FALSE; } } if (rsc->children) { GListPtr gIter = NULL; gboolean any_created = FALSE; for (gIter = rsc->children; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; any_created = child_rsc->cmds->create_probe(child_rsc, node, complete, force, data_set) || any_created; } return any_created; } else if ((rsc->container) && (!rsc->is_remote_node)) { pe_rsc_trace(rsc, "Skipping %s: it is within container %s", rsc->id, rsc->container->id); return FALSE; } if (is_set(rsc->flags, pe_rsc_orphan)) { pe_rsc_trace(rsc, "Skipping orphan: %s", rsc->id); return FALSE; } // Check whether resource is already known on node if (!force && g_hash_table_lookup(rsc->known_on, node->details->id)) { pe_rsc_trace(rsc, "Skipping known: %s on %s", rsc->id, node->details->uname); return FALSE; } allowed = g_hash_table_lookup(rsc->allowed_nodes, node->details->id); if (rsc->exclusive_discover || top->exclusive_discover) { if (allowed == NULL) { /* exclusive discover is enabled and this node is not in the allowed list. */ pe_rsc_trace(rsc, "Skipping probe for %s on node %s, A", rsc->id, node->details->id); return FALSE; } else if (allowed->rsc_discover_mode != pe_discover_exclusive) { /* exclusive discover is enabled and this node is not marked * as a node this resource should be discovered on */ pe_rsc_trace(rsc, "Skipping probe for %s on node %s, B", rsc->id, node->details->id); return FALSE; } } if(allowed == NULL && node->rsc_discover_mode == pe_discover_never) { /* If this node was allowed to host this resource it would * have been explicitly added to the 'allowed_nodes' list. * However it wasn't and the node has discovery disabled, so * no need to probe for this resource. */ pe_rsc_trace(rsc, "Skipping probe for %s on node %s, C", rsc->id, node->details->id); return FALSE; } if (allowed && allowed->rsc_discover_mode == pe_discover_never) { /* this resource is marked as not needing to be discovered on this node */ pe_rsc_trace(rsc, "Skipping probe for %s on node %s, discovery mode", rsc->id, node->details->id); return FALSE; } if (pe__is_guest_node(node)) { resource_t *remote = node->details->remote_rsc->container; if(remote->role == RSC_ROLE_STOPPED) { /* If the container is stopped, then we know anything that * might have been inside it is also stopped and there is * no need to probe. * * If we don't know the container's state on the target * either: * * - the container is running, the transition will abort * and we'll end up in a different case next time, or * * - the container is stopped * * Either way there is no need to probe. * */ if(remote->allocated_to && g_hash_table_lookup(remote->known_on, remote->allocated_to->details->id) == NULL) { /* For safety, we order the 'rsc' start after 'remote' * has been probed. * * Using 'top' helps for groups, but we may need to * follow the start's ordering chain backwards. */ custom_action_order(remote, generate_op_key(remote->id, RSC_STATUS, 0), NULL, top, generate_op_key(top->id, RSC_START, 0), NULL, pe_order_optional, data_set); } pe_rsc_trace(rsc, "Skipping probe for %s on node %s, %s is stopped", rsc->id, node->details->id, remote->id); return FALSE; /* Here we really we want to check if remote->stop is required, * but that information doesn't exist yet */ } else if(node->details->remote_requires_reset || node->details->unclean || is_set(remote->flags, pe_rsc_failed) || remote->next_role == RSC_ROLE_STOPPED || (remote->allocated_to && pe_find_node(remote->running_on, remote->allocated_to->details->uname) == NULL) ) { /* The container is stopping or restarting, don't start * 'rsc' until 'remote' stops as this also implies that * 'rsc' is stopped - avoiding the need to probe */ custom_action_order(remote, generate_op_key(remote->id, RSC_STOP, 0), NULL, top, generate_op_key(top->id, RSC_START, 0), NULL, pe_order_optional, data_set); pe_rsc_trace(rsc, "Skipping probe for %s on node %s, %s is stopping, restarting or moving", rsc->id, node->details->id, remote->id); return FALSE; /* } else { * The container is running so there is no problem probing it */ } } key = generate_op_key(rsc->id, RSC_STATUS, 0); probe = custom_action(rsc, key, RSC_STATUS, node, FALSE, TRUE, data_set); update_action_flags(probe, pe_action_optional | pe_action_clear, __FUNCTION__, __LINE__); order_after_unfencing(rsc, node, probe, pe_order_optional, data_set); /* * We need to know if it's running_on (not just known_on) this node * to correctly determine the target rc. */ running = pe_find_node_id(rsc->running_on, node->details->id); if (running == NULL) { add_hash_param(probe->meta, XML_ATTR_TE_TARGET_RC, rc_inactive); } else if (rsc->role == RSC_ROLE_MASTER) { add_hash_param(probe->meta, XML_ATTR_TE_TARGET_RC, rc_master); } crm_debug("Probing %s on %s (%s) %d %p", rsc->id, node->details->uname, role2text(rsc->role), is_set(probe->flags, pe_action_runnable), rsc->running_on); if (is_unfence_device(rsc, data_set) || !pe_rsc_is_clone(top)) { top = rsc; } else { crm_trace("Probing %s on %s (%s) as %s", rsc->id, node->details->uname, role2text(rsc->role), top->id); } if(is_not_set(probe->flags, pe_action_runnable) && rsc->running_on == NULL) { /* Prevent the start from occurring if rsc isn't active, but * don't cause it to stop if it was active already */ flags |= pe_order_runnable_left; } custom_action_order(rsc, NULL, probe, top, generate_op_key(top->id, RSC_START, 0), NULL, flags, data_set); /* Before any reloads, if they exist */ custom_action_order(rsc, NULL, probe, top, reload_key(rsc), NULL, pe_order_optional, data_set); #if 0 // complete is always null currently if (!is_unfence_device(rsc, data_set)) { /* Normally rsc.start depends on probe complete which depends * on rsc.probe. But this can't be the case for fence devices * with unfencing, as it would create graph loops. * * So instead we explicitly order 'rsc.probe then rsc.start' */ order_actions(probe, complete, pe_order_implies_then); } #endif return TRUE; } /*! * \internal * \brief Check whether a resource is known on a particular node * * \param[in] rsc Resource to check * \param[in] node Node to check * * \return TRUE if resource (or parent if an anonymous clone) is known */ static bool rsc_is_known_on(pe_resource_t *rsc, const pe_node_t *node) { if (pe_hash_table_lookup(rsc->known_on, node->details->id)) { return TRUE; } else if ((rsc->variant == pe_native) && pe_rsc_is_anon_clone(rsc->parent) && pe_hash_table_lookup(rsc->parent->known_on, node->details->id)) { /* We check only the parent, not the uber-parent, because we cannot * assume that the resource is known if it is in an anonymously cloned * group (which may be only partially known). */ return TRUE; } return FALSE; } /*! * \internal * \brief Order a resource's start and promote actions relative to fencing * * \param[in] rsc Resource to be ordered * \param[in] stonith_op Fence action * \param[in] data_set Cluster information */ static void native_start_constraints(resource_t * rsc, action_t * stonith_op, pe_working_set_t * data_set) { node_t *target; GListPtr gIter = NULL; CRM_CHECK(stonith_op && stonith_op->node, return); target = stonith_op->node; for (gIter = rsc->actions; gIter != NULL; gIter = gIter->next) { action_t *action = (action_t *) gIter->data; switch (action->needs) { case rsc_req_nothing: // Anything other than start or promote requires nothing break; case rsc_req_stonith: order_actions(stonith_op, action, pe_order_optional); break; case rsc_req_quorum: if (safe_str_eq(action->task, RSC_START) && pe_hash_table_lookup(rsc->allowed_nodes, target->details->id) && !rsc_is_known_on(rsc, target)) { /* If we don't know the status of the resource on the node * we're about to shoot, we have to assume it may be active * there. Order the resource start after the fencing. This * is analogous to waiting for all the probes for a resource * to complete before starting it. * * The most likely explanation is that the DC died and took * its status with it. */ pe_rsc_debug(rsc, "Ordering %s after %s recovery", action->uuid, target->details->uname); order_actions(stonith_op, action, pe_order_optional | pe_order_runnable_left); } break; } } } static void native_stop_constraints(resource_t * rsc, action_t * stonith_op, pe_working_set_t * data_set) { GListPtr gIter = NULL; GListPtr action_list = NULL; bool order_implicit = false; resource_t *top = uber_parent(rsc); pe_action_t *parent_stop = NULL; node_t *target; CRM_CHECK(stonith_op && stonith_op->node, return); target = stonith_op->node; /* Get a list of stop actions potentially implied by the fencing */ action_list = pe__resource_actions(rsc, target, RSC_STOP, FALSE); /* If resource requires fencing, implicit actions must occur after fencing. * * Implied stops and demotes of resources running on guest nodes are always * ordered after fencing, even if the resource does not require fencing, * because guest node "fencing" is actually just a resource stop. */ if (is_set(rsc->flags, pe_rsc_needs_fencing) || pe__is_guest_node(target)) { order_implicit = true; } if (action_list && order_implicit) { parent_stop = find_first_action(top->actions, NULL, RSC_STOP, NULL); } for (gIter = action_list; gIter != NULL; gIter = gIter->next) { action_t *action = (action_t *) gIter->data; // The stop would never complete, so convert it into a pseudo-action. update_action_flags(action, pe_action_pseudo|pe_action_runnable, __FUNCTION__, __LINE__); if (order_implicit) { update_action_flags(action, pe_action_implied_by_stonith, __FUNCTION__, __LINE__); /* Order the stonith before the parent stop (if any). * * Also order the stonith before the resource stop, unless the * resource is inside a bundle -- that would cause a graph loop. * We can rely on the parent stop's ordering instead. * * User constraints must not order a resource in a guest node * relative to the guest node container resource. The * pe_order_preserve flag marks constraints as generated by the * cluster and thus immune to that check (and is irrelevant if * target is not a guest). */ if (!pe_rsc_is_bundled(rsc)) { order_actions(stonith_op, action, pe_order_preserve); } order_actions(stonith_op, parent_stop, pe_order_preserve); } if (is_set(rsc->flags, pe_rsc_failed)) { crm_notice("Stop of failed resource %s is implicit %s %s is fenced", rsc->id, (order_implicit? "after" : "because"), target->details->uname); } else { crm_info("%s is implicit %s %s is fenced", action->uuid, (order_implicit? "after" : "because"), target->details->uname); } if (is_set(rsc->flags, pe_rsc_notify)) { /* Create a second notification that will be delivered * immediately after the node is fenced * * Basic problem: * - C is a clone active on the node to be shot and stopping on another * - R is a resource that depends on C * * + C.stop depends on R.stop * + C.stopped depends on STONITH * + C.notify depends on C.stopped * + C.healthy depends on C.notify * + R.stop depends on C.healthy * * The extra notification here changes * + C.healthy depends on C.notify * into: * + C.healthy depends on C.notify' * + C.notify' depends on STONITH' * thus breaking the loop */ create_secondary_notification(action, rsc, stonith_op, data_set); } /* From Bug #1601, successful fencing must be an input to a failed resources stop action. However given group(rA, rB) running on nodeX and B.stop has failed, A := stop healthy resource (rA.stop) B := stop failed resource (pseudo operation B.stop) C := stonith nodeX A requires B, B requires C, C requires A This loop would prevent the cluster from making progress. This block creates the "C requires A" dependency and therefore must (at least for now) be disabled. Instead, run the block above and treat all resources on nodeX as B would be (marked as a pseudo op depending on the STONITH). TODO: Break the "A requires B" dependency in update_action() and re-enable this block } else if(is_stonith == FALSE) { crm_info("Moving healthy resource %s" " off %s before fencing", rsc->id, node->details->uname); * stop healthy resources before the * stonith op * custom_action_order( rsc, stop_key(rsc), NULL, NULL,strdup(CRM_OP_FENCE),stonith_op, pe_order_optional, data_set); */ } g_list_free(action_list); /* Get a list of demote actions potentially implied by the fencing */ action_list = pe__resource_actions(rsc, target, RSC_DEMOTE, FALSE); for (gIter = action_list; gIter != NULL; gIter = gIter->next) { action_t *action = (action_t *) gIter->data; if (action->node->details->online == FALSE || action->node->details->unclean == TRUE || is_set(rsc->flags, pe_rsc_failed)) { if (is_set(rsc->flags, pe_rsc_failed)) { pe_rsc_info(rsc, "Demote of failed resource %s is implicit after %s is fenced", rsc->id, target->details->uname); } else { pe_rsc_info(rsc, "%s is implicit after %s is fenced", action->uuid, target->details->uname); } /* The demote would never complete and is now implied by the * fencing, so convert it into a pseudo-action. */ update_action_flags(action, pe_action_pseudo|pe_action_runnable, __FUNCTION__, __LINE__); if (pe_rsc_is_bundled(rsc)) { /* Do nothing, let the recovery be ordered after the parent's implied stop */ } else if (order_implicit) { order_actions(stonith_op, action, pe_order_preserve|pe_order_optional); } } } g_list_free(action_list); } void rsc_stonith_ordering(resource_t * rsc, action_t * stonith_op, pe_working_set_t * data_set) { if (rsc->children) { GListPtr gIter = NULL; for (gIter = rsc->children; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; rsc_stonith_ordering(child_rsc, stonith_op, data_set); } } else if (is_not_set(rsc->flags, pe_rsc_managed)) { pe_rsc_trace(rsc, "Skipping fencing constraints for unmanaged resource: %s", rsc->id); } else { native_start_constraints(rsc, stonith_op, data_set); native_stop_constraints(rsc, stonith_op, data_set); } } void ReloadRsc(resource_t * rsc, node_t *node, pe_working_set_t * data_set) { GListPtr gIter = NULL; action_t *reload = NULL; if (rsc->children) { for (gIter = rsc->children; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; ReloadRsc(child_rsc, node, data_set); } return; } else if (rsc->variant > pe_native) { /* Complex resource with no children */ return; } else if (is_not_set(rsc->flags, pe_rsc_managed)) { pe_rsc_trace(rsc, "%s: unmanaged", rsc->id); return; } else if (is_set(rsc->flags, pe_rsc_failed) || is_set(rsc->flags, pe_rsc_start_pending)) { pe_rsc_trace(rsc, "%s: general resource state: flags=0x%.16llx", rsc->id, rsc->flags); stop_action(rsc, node, FALSE); /* Force a full restart, overkill? */ return; } else if (node == NULL) { pe_rsc_trace(rsc, "%s: not active", rsc->id); return; } pe_rsc_trace(rsc, "Processing %s", rsc->id); set_bit(rsc->flags, pe_rsc_reload); reload = custom_action( rsc, reload_key(rsc), CRMD_ACTION_RELOAD, node, FALSE, TRUE, data_set); pe_action_set_reason(reload, "resource definition change", FALSE); custom_action_order(NULL, NULL, reload, rsc, stop_key(rsc), NULL, pe_order_optional|pe_order_then_cancels_first, data_set); custom_action_order(NULL, NULL, reload, rsc, demote_key(rsc), NULL, pe_order_optional|pe_order_then_cancels_first, data_set); } void native_append_meta(resource_t * rsc, xmlNode * xml) { char *value = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_INCARNATION); resource_t *parent; if (value) { char *name = NULL; name = crm_meta_name(XML_RSC_ATTR_INCARNATION); crm_xml_add(xml, name, value); free(name); } value = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_REMOTE_NODE); if (value) { char *name = NULL; name = crm_meta_name(XML_RSC_ATTR_REMOTE_NODE); crm_xml_add(xml, name, value); free(name); } for (parent = rsc; parent != NULL; parent = parent->parent) { if (parent->container) { crm_xml_add(xml, CRM_META"_"XML_RSC_ATTR_CONTAINER, parent->container->id); } } } diff --git a/lib/pengine/bundle.c b/lib/pengine/bundle.c index e5c5fa2b71..d9e26383fb 100644 --- a/lib/pengine/bundle.c +++ b/lib/pengine/bundle.c @@ -1,1915 +1,1939 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU Lesser General Public License * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #define PE__VARIANT_BUNDLE 1 #include "./variant.h" static char * next_ip(const char *last_ip) { unsigned int oct1 = 0; unsigned int oct2 = 0; unsigned int oct3 = 0; unsigned int oct4 = 0; int rc = sscanf(last_ip, "%u.%u.%u.%u", &oct1, &oct2, &oct3, &oct4); if (rc != 4) { /*@ TODO check for IPv6 */ return NULL; } else if (oct3 > 253) { return NULL; } else if (oct4 > 253) { ++oct3; oct4 = 1; } else { ++oct4; } return crm_strdup_printf("%u.%u.%u.%u", oct1, oct2, oct3, oct4); } static int allocate_ip(pe__bundle_variant_data_t *data, pe__bundle_replica_t *replica, char *buffer, int max) { if(data->ip_range_start == NULL) { return 0; } else if(data->ip_last) { replica->ipaddr = next_ip(data->ip_last); } else { replica->ipaddr = strdup(data->ip_range_start); } data->ip_last = replica->ipaddr; switch (data->agent_type) { case PE__CONTAINER_AGENT_DOCKER: case PE__CONTAINER_AGENT_PODMAN: if (data->add_host) { return snprintf(buffer, max, " --add-host=%s-%d:%s", data->prefix, replica->offset, replica->ipaddr); } case PE__CONTAINER_AGENT_RKT: return snprintf(buffer, max, " --hosts-entry=%s=%s-%d", replica->ipaddr, data->prefix, replica->offset); default: // PE__CONTAINER_AGENT_UNKNOWN break; } return 0; } static xmlNode * create_resource(const char *name, const char *provider, const char *kind) { xmlNode *rsc = create_xml_node(NULL, XML_CIB_TAG_RESOURCE); crm_xml_add(rsc, XML_ATTR_ID, name); crm_xml_add(rsc, XML_AGENT_ATTR_CLASS, PCMK_RESOURCE_CLASS_OCF); crm_xml_add(rsc, XML_AGENT_ATTR_PROVIDER, provider); crm_xml_add(rsc, XML_ATTR_TYPE, kind); return rsc; } /*! * \internal * \brief Check whether cluster can manage resource inside container * * \param[in] data Container variant data * * \return TRUE if networking configuration is acceptable, FALSE otherwise * * \note The resource is manageable if an IP range or control port has been * specified. If a control port is used without an IP range, replicas per * host must be 1. */ static bool valid_network(pe__bundle_variant_data_t *data) { if(data->ip_range_start) { return TRUE; } if(data->control_port) { if(data->nreplicas_per_host > 1) { pe_err("Specifying the 'control-port' for %s requires 'replicas-per-host=1'", data->prefix); data->nreplicas_per_host = 1; /* @TODO to be sure: clear_bit(rsc->flags, pe_rsc_unique); */ } return TRUE; } return FALSE; } static bool create_ip_resource(pe_resource_t *parent, pe__bundle_variant_data_t *data, pe__bundle_replica_t *replica, pe_working_set_t *data_set) { if(data->ip_range_start) { char *id = NULL; xmlNode *xml_ip = NULL; xmlNode *xml_obj = NULL; id = crm_strdup_printf("%s-ip-%s", data->prefix, replica->ipaddr); crm_xml_sanitize_id(id); xml_ip = create_resource(id, "heartbeat", "IPaddr2"); free(id); xml_obj = create_xml_node(xml_ip, XML_TAG_ATTR_SETS); crm_xml_set_id(xml_obj, "%s-attributes-%d", data->prefix, replica->offset); crm_create_nvpair_xml(xml_obj, NULL, "ip", replica->ipaddr); if(data->host_network) { crm_create_nvpair_xml(xml_obj, NULL, "nic", data->host_network); } if(data->host_netmask) { crm_create_nvpair_xml(xml_obj, NULL, "cidr_netmask", data->host_netmask); } else { crm_create_nvpair_xml(xml_obj, NULL, "cidr_netmask", "32"); } xml_obj = create_xml_node(xml_ip, "operations"); crm_create_op_xml(xml_obj, ID(xml_ip), "monitor", "60s", NULL); // TODO: Other ops? Timeouts and intervals from underlying resource? if (!common_unpack(xml_ip, &replica->ip, parent, data_set)) { return FALSE; } parent->children = g_list_append(parent->children, replica->ip); } return TRUE; } static bool create_docker_resource(pe_resource_t *parent, pe__bundle_variant_data_t *data, pe__bundle_replica_t *replica, pe_working_set_t *data_set) { int offset = 0, max = 4096; char *buffer = calloc(1, max+1); int doffset = 0, dmax = 1024; char *dbuffer = calloc(1, dmax+1); char *id = NULL; xmlNode *xml_container = NULL; xmlNode *xml_obj = NULL; id = crm_strdup_printf("%s-docker-%d", data->prefix, replica->offset); crm_xml_sanitize_id(id); xml_container = create_resource(id, "heartbeat", PE__CONTAINER_AGENT_DOCKER_S); free(id); xml_obj = create_xml_node(xml_container, XML_TAG_ATTR_SETS); crm_xml_set_id(xml_obj, "%s-attributes-%d", data->prefix, replica->offset); crm_create_nvpair_xml(xml_obj, NULL, "image", data->image); crm_create_nvpair_xml(xml_obj, NULL, "allow_pull", XML_BOOLEAN_TRUE); crm_create_nvpair_xml(xml_obj, NULL, "force_kill", XML_BOOLEAN_FALSE); crm_create_nvpair_xml(xml_obj, NULL, "reuse", XML_BOOLEAN_FALSE); offset += snprintf(buffer+offset, max-offset, " --restart=no"); /* Set a container hostname only if we have an IP to map it to. * The user can set -h or --uts=host themselves if they want a nicer * name for logs, but this makes applications happy who need their * hostname to match the IP they bind to. */ if (data->ip_range_start != NULL) { offset += snprintf(buffer+offset, max-offset, " -h %s-%d", data->prefix, replica->offset); } offset += snprintf(buffer+offset, max-offset, " -e PCMK_stderr=1"); if (data->container_network) { #if 0 offset += snprintf(buffer+offset, max-offset, " --link-local-ip=%s", replica->ipaddr); #endif offset += snprintf(buffer+offset, max-offset, " --net=%s", data->container_network); } if(data->control_port) { offset += snprintf(buffer+offset, max-offset, " -e PCMK_remote_port=%s", data->control_port); } else { offset += snprintf(buffer+offset, max-offset, " -e PCMK_remote_port=%d", DEFAULT_REMOTE_PORT); } for(GListPtr pIter = data->mounts; pIter != NULL; pIter = pIter->next) { pe__bundle_mount_t *mount = pIter->data; if (is_set(mount->flags, pe__bundle_mount_subdir)) { char *source = crm_strdup_printf( "%s/%s-%d", mount->source, data->prefix, replica->offset); if(doffset > 0) { doffset += snprintf(dbuffer+doffset, dmax-doffset, ","); } doffset += snprintf(dbuffer+doffset, dmax-doffset, "%s", source); offset += snprintf(buffer+offset, max-offset, " -v %s:%s", source, mount->target); free(source); } else { offset += snprintf(buffer+offset, max-offset, " -v %s:%s", mount->source, mount->target); } if(mount->options) { offset += snprintf(buffer+offset, max-offset, ":%s", mount->options); } } for(GListPtr pIter = data->ports; pIter != NULL; pIter = pIter->next) { pe__bundle_port_t *port = pIter->data; if (replica->ipaddr) { offset += snprintf(buffer+offset, max-offset, " -p %s:%s:%s", replica->ipaddr, port->source, port->target); } else if(safe_str_neq(data->container_network, "host")) { // No need to do port mapping if net=host offset += snprintf(buffer+offset, max-offset, " -p %s:%s", port->source, port->target); } } if (data->launcher_options) { offset += snprintf(buffer+offset, max-offset, " %s", data->launcher_options); } if (data->container_host_options) { offset += snprintf(buffer + offset, max - offset, " %s", data->container_host_options); } crm_create_nvpair_xml(xml_obj, NULL, "run_opts", buffer); free(buffer); crm_create_nvpair_xml(xml_obj, NULL, "mount_points", dbuffer); free(dbuffer); if (replica->child) { if (data->container_command) { crm_create_nvpair_xml(xml_obj, NULL, "run_cmd", data->container_command); } else { crm_create_nvpair_xml(xml_obj, NULL, "run_cmd", SBIN_DIR "/pacemaker-remoted"); } /* TODO: Allow users to specify their own? * * We just want to know if the container is alive, we'll * monitor the child independently */ crm_create_nvpair_xml(xml_obj, NULL, "monitor_cmd", "/bin/true"); /* } else if(child && data->untrusted) { * Support this use-case? * * The ability to have resources started/stopped by us, but * unable to set attributes, etc. * * Arguably better to control API access this with ACLs like * "normal" remote nodes * * crm_create_nvpair_xml(xml_obj, NULL, * "run_cmd", * "/usr/libexec/pacemaker/pacemaker-execd"); * crm_create_nvpair_xml(xml_obj, NULL, "monitor_cmd", * "/usr/libexec/pacemaker/lrmd_internal_ctl -c poke"); */ } else { if (data->container_command) { crm_create_nvpair_xml(xml_obj, NULL, "run_cmd", data->container_command); } /* TODO: Allow users to specify their own? * * We don't know what's in the container, so we just want * to know if it is alive */ crm_create_nvpair_xml(xml_obj, NULL, "monitor_cmd", "/bin/true"); } xml_obj = create_xml_node(xml_container, "operations"); crm_create_op_xml(xml_obj, ID(xml_container), "monitor", "60s", NULL); // TODO: Other ops? Timeouts and intervals from underlying resource? if (!common_unpack(xml_container, &replica->container, parent, data_set)) { return FALSE; } parent->children = g_list_append(parent->children, replica->container); return TRUE; } static bool create_podman_resource(pe_resource_t *parent, pe__bundle_variant_data_t *data, pe__bundle_replica_t *replica, pe_working_set_t *data_set) { int offset = 0, max = 4096; char *buffer = calloc(1, max+1); int doffset = 0, dmax = 1024; char *dbuffer = calloc(1, dmax+1); char *id = NULL; xmlNode *xml_container = NULL; xmlNode *xml_obj = NULL; id = crm_strdup_printf("%s-podman-%d", data->prefix, replica->offset); crm_xml_sanitize_id(id); xml_container = create_resource(id, "heartbeat", PE__CONTAINER_AGENT_PODMAN_S); free(id); xml_obj = create_xml_node(xml_container, XML_TAG_ATTR_SETS); crm_xml_set_id(xml_obj, "%s-attributes-%d", data->prefix, replica->offset); crm_create_nvpair_xml(xml_obj, NULL, "image", data->image); crm_create_nvpair_xml(xml_obj, NULL, "allow_pull", XML_BOOLEAN_TRUE); crm_create_nvpair_xml(xml_obj, NULL, "force_kill", XML_BOOLEAN_FALSE); crm_create_nvpair_xml(xml_obj, NULL, "reuse", XML_BOOLEAN_FALSE); // FIXME: (bandini 2018-08) podman has no restart policies //offset += snprintf(buffer+offset, max-offset, " --restart=no"); /* Set a container hostname only if we have an IP to map it to. * The user can set -h or --uts=host themselves if they want a nicer * name for logs, but this makes applications happy who need their * hostname to match the IP they bind to. */ if (data->ip_range_start != NULL) { offset += snprintf(buffer+offset, max-offset, " -h %s-%d", data->prefix, replica->offset); } offset += snprintf(buffer+offset, max-offset, " -e PCMK_stderr=1"); if (data->container_network) { #if 0 // podman has no support for --link-local-ip offset += snprintf(buffer+offset, max-offset, " --link-local-ip=%s", replica->ipaddr); #endif offset += snprintf(buffer+offset, max-offset, " --net=%s", data->container_network); } if(data->control_port) { offset += snprintf(buffer+offset, max-offset, " -e PCMK_remote_port=%s", data->control_port); } else { offset += snprintf(buffer+offset, max-offset, " -e PCMK_remote_port=%d", DEFAULT_REMOTE_PORT); } for(GListPtr pIter = data->mounts; pIter != NULL; pIter = pIter->next) { pe__bundle_mount_t *mount = pIter->data; if (is_set(mount->flags, pe__bundle_mount_subdir)) { char *source = crm_strdup_printf( "%s/%s-%d", mount->source, data->prefix, replica->offset); if(doffset > 0) { doffset += snprintf(dbuffer+doffset, dmax-doffset, ","); } doffset += snprintf(dbuffer+doffset, dmax-doffset, "%s", source); offset += snprintf(buffer+offset, max-offset, " -v %s:%s", source, mount->target); free(source); } else { offset += snprintf(buffer+offset, max-offset, " -v %s:%s", mount->source, mount->target); } if(mount->options) { offset += snprintf(buffer+offset, max-offset, ":%s", mount->options); } } for(GListPtr pIter = data->ports; pIter != NULL; pIter = pIter->next) { pe__bundle_port_t *port = pIter->data; if (replica->ipaddr) { offset += snprintf(buffer+offset, max-offset, " -p %s:%s:%s", replica->ipaddr, port->source, port->target); } else if(safe_str_neq(data->container_network, "host")) { // No need to do port mapping if net=host offset += snprintf(buffer+offset, max-offset, " -p %s:%s", port->source, port->target); } } if (data->launcher_options) { offset += snprintf(buffer+offset, max-offset, " %s", data->launcher_options); } if (data->container_host_options) { offset += snprintf(buffer + offset, max - offset, " %s", data->container_host_options); } crm_create_nvpair_xml(xml_obj, NULL, "run_opts", buffer); free(buffer); crm_create_nvpair_xml(xml_obj, NULL, "mount_points", dbuffer); free(dbuffer); if (replica->child) { if (data->container_command) { crm_create_nvpair_xml(xml_obj, NULL, "run_cmd", data->container_command); } else { crm_create_nvpair_xml(xml_obj, NULL, "run_cmd", SBIN_DIR "/pacemaker-remoted"); } /* TODO: Allow users to specify their own? * * We just want to know if the container is alive, we'll * monitor the child independently */ crm_create_nvpair_xml(xml_obj, NULL, "monitor_cmd", "/bin/true"); /* } else if(child && data->untrusted) { * Support this use-case? * * The ability to have resources started/stopped by us, but * unable to set attributes, etc. * * Arguably better to control API access this with ACLs like * "normal" remote nodes * * crm_create_nvpair_xml(xml_obj, NULL, * "run_cmd", * "/usr/libexec/pacemaker/pacemaker-execd"); * crm_create_nvpair_xml(xml_obj, NULL, "monitor_cmd", * "/usr/libexec/pacemaker/lrmd_internal_ctl -c poke"); */ } else { if (data->container_command) { crm_create_nvpair_xml(xml_obj, NULL, "run_cmd", data->container_command); } /* TODO: Allow users to specify their own? * * We don't know what's in the container, so we just want * to know if it is alive */ crm_create_nvpair_xml(xml_obj, NULL, "monitor_cmd", "/bin/true"); } xml_obj = create_xml_node(xml_container, "operations"); crm_create_op_xml(xml_obj, ID(xml_container), "monitor", "60s", NULL); // TODO: Other ops? Timeouts and intervals from underlying resource? if (!common_unpack(xml_container, &replica->container, parent, data_set)) { return FALSE; } parent->children = g_list_append(parent->children, replica->container); return TRUE; } static bool create_rkt_resource(pe_resource_t *parent, pe__bundle_variant_data_t *data, pe__bundle_replica_t *replica, pe_working_set_t *data_set) { int offset = 0, max = 4096; char *buffer = calloc(1, max+1); int doffset = 0, dmax = 1024; char *dbuffer = calloc(1, dmax+1); char *id = NULL; xmlNode *xml_container = NULL; xmlNode *xml_obj = NULL; int volid = 0; id = crm_strdup_printf("%s-rkt-%d", data->prefix, replica->offset); crm_xml_sanitize_id(id); xml_container = create_resource(id, "heartbeat", PE__CONTAINER_AGENT_RKT_S); free(id); xml_obj = create_xml_node(xml_container, XML_TAG_ATTR_SETS); crm_xml_set_id(xml_obj, "%s-attributes-%d", data->prefix, replica->offset); crm_create_nvpair_xml(xml_obj, NULL, "image", data->image); crm_create_nvpair_xml(xml_obj, NULL, "allow_pull", "true"); crm_create_nvpair_xml(xml_obj, NULL, "force_kill", "false"); crm_create_nvpair_xml(xml_obj, NULL, "reuse", "false"); /* Set a container hostname only if we have an IP to map it to. * The user can set -h or --uts=host themselves if they want a nicer * name for logs, but this makes applications happy who need their * hostname to match the IP they bind to. */ if (data->ip_range_start != NULL) { offset += snprintf(buffer+offset, max-offset, " --hostname=%s-%d", data->prefix, replica->offset); } offset += snprintf(buffer+offset, max-offset, " --environment=PCMK_stderr=1"); if (data->container_network) { #if 0 offset += snprintf(buffer+offset, max-offset, " --link-local-ip=%s", replica->ipaddr); #endif offset += snprintf(buffer+offset, max-offset, " --net=%s", data->container_network); } if(data->control_port) { offset += snprintf(buffer+offset, max-offset, " --environment=PCMK_remote_port=%s", data->control_port); } else { offset += snprintf(buffer+offset, max-offset, " --environment=PCMK_remote_port=%d", DEFAULT_REMOTE_PORT); } for(GListPtr pIter = data->mounts; pIter != NULL; pIter = pIter->next) { pe__bundle_mount_t *mount = pIter->data; if (is_set(mount->flags, pe__bundle_mount_subdir)) { char *source = crm_strdup_printf( "%s/%s-%d", mount->source, data->prefix, replica->offset); if(doffset > 0) { doffset += snprintf(dbuffer+doffset, dmax-doffset, ","); } doffset += snprintf(dbuffer+doffset, dmax-doffset, "%s", source); offset += snprintf(buffer+offset, max-offset, " --volume vol%d,kind=host,source=%s", volid, source); if(mount->options) { offset += snprintf(buffer+offset, max-offset, ",%s", mount->options); } offset += snprintf(buffer+offset, max-offset, " --mount volume=vol%d,target=%s", volid, mount->target); free(source); } else { offset += snprintf(buffer+offset, max-offset, " --volume vol%d,kind=host,source=%s", volid, mount->source); if(mount->options) { offset += snprintf(buffer+offset, max-offset, ",%s", mount->options); } offset += snprintf(buffer+offset, max-offset, " --mount volume=vol%d,target=%s", volid, mount->target); } volid++; } for(GListPtr pIter = data->ports; pIter != NULL; pIter = pIter->next) { pe__bundle_port_t *port = pIter->data; if (replica->ipaddr) { offset += snprintf(buffer+offset, max-offset, " --port=%s:%s:%s", port->target, replica->ipaddr, port->source); } else { offset += snprintf(buffer+offset, max-offset, " --port=%s:%s", port->target, port->source); } } if (data->launcher_options) { offset += snprintf(buffer+offset, max-offset, " %s", data->launcher_options); } if (data->container_host_options) { offset += snprintf(buffer + offset, max - offset, " %s", data->container_host_options); } crm_create_nvpair_xml(xml_obj, NULL, "run_opts", buffer); free(buffer); crm_create_nvpair_xml(xml_obj, NULL, "mount_points", dbuffer); free(dbuffer); if (replica->child) { if (data->container_command) { crm_create_nvpair_xml(xml_obj, NULL, "run_cmd", data->container_command); } else { crm_create_nvpair_xml(xml_obj, NULL, "run_cmd", SBIN_DIR "/pacemaker-remoted"); } /* TODO: Allow users to specify their own? * * We just want to know if the container is alive, we'll * monitor the child independently */ crm_create_nvpair_xml(xml_obj, NULL, "monitor_cmd", "/bin/true"); /* } else if(child && data->untrusted) { * Support this use-case? * * The ability to have resources started/stopped by us, but * unable to set attributes, etc. * * Arguably better to control API access this with ACLs like * "normal" remote nodes * * crm_create_nvpair_xml(xml_obj, NULL, * "run_cmd", * "/usr/libexec/pacemaker/pacemaker-execd"); * crm_create_nvpair_xml(xml_obj, NULL, "monitor_cmd", * "/usr/libexec/pacemaker/lrmd_internal_ctl -c poke"); */ } else { if (data->container_command) { crm_create_nvpair_xml(xml_obj, NULL, "run_cmd", data->container_command); } /* TODO: Allow users to specify their own? * * We don't know what's in the container, so we just want * to know if it is alive */ crm_create_nvpair_xml(xml_obj, NULL, "monitor_cmd", "/bin/true"); } xml_obj = create_xml_node(xml_container, "operations"); crm_create_op_xml(xml_obj, ID(xml_container), "monitor", "60s", NULL); // TODO: Other ops? Timeouts and intervals from underlying resource? if (!common_unpack(xml_container, &replica->container, parent, data_set)) { return FALSE; } parent->children = g_list_append(parent->children, replica->container); return TRUE; } /*! * \brief Ban a node from a resource's (and its children's) allowed nodes list * * \param[in,out] rsc Resource to modify * \param[in] uname Name of node to ban */ static void disallow_node(resource_t *rsc, const char *uname) { gpointer match = g_hash_table_lookup(rsc->allowed_nodes, uname); if (match) { ((pe_node_t *) match)->weight = -INFINITY; ((pe_node_t *) match)->rsc_discover_mode = pe_discover_never; } if (rsc->children) { GListPtr child; for (child = rsc->children; child != NULL; child = child->next) { disallow_node((resource_t *) (child->data), uname); } } } static bool create_remote_resource(pe_resource_t *parent, pe__bundle_variant_data_t *data, pe__bundle_replica_t *replica, pe_working_set_t *data_set) { if (replica->child && valid_network(data)) { GHashTableIter gIter; GListPtr rsc_iter = NULL; node_t *node = NULL; xmlNode *xml_remote = NULL; char *id = crm_strdup_printf("%s-%d", data->prefix, replica->offset); char *port_s = NULL; const char *uname = NULL; const char *connect_name = NULL; if (remote_id_conflict(id, data_set)) { free(id); // The biggest hammer we have id = crm_strdup_printf("pcmk-internal-%s-remote-%d", replica->child->id, replica->offset); CRM_ASSERT(remote_id_conflict(id, data_set) == FALSE); } /* REMOTE_CONTAINER_HACK: Using "#uname" as the server name when the * connection does not have its own IP is a magic string that we use to * support nested remotes (i.e. a bundle running on a remote node). */ connect_name = (replica->ipaddr? replica->ipaddr : "#uname"); if (data->control_port == NULL) { port_s = crm_itoa(DEFAULT_REMOTE_PORT); } /* This sets replica->container as replica->remote's container, which is * similar to what happens with guest nodes. This is how the PE knows * that the bundle node is fenced by recovering the container, and that * remote should be ordered relative to the container. */ xml_remote = pe_create_remote_xml(NULL, id, replica->container->id, NULL, NULL, NULL, connect_name, (data->control_port? data->control_port : port_s)); free(port_s); /* Abandon our created ID, and pull the copy from the XML, because we * need something that will get freed during data set cleanup to use as * the node ID and uname. */ free(id); id = NULL; uname = ID(xml_remote); /* Ensure a node has been created for the guest (it may have already * been, if it has a permanent node attribute), and ensure its weight is * -INFINITY so no other resources can run on it. */ node = pe_find_node(data_set->nodes, uname); if (node == NULL) { node = pe_create_node(uname, uname, "remote", "-INFINITY", data_set); } else { node->weight = -INFINITY; } node->rsc_discover_mode = pe_discover_never; /* unpack_remote_nodes() ensures that each remote node and guest node * has a pe_node_t entry. Ideally, it would do the same for bundle nodes. * Unfortunately, a bundle has to be mostly unpacked before it's obvious * what nodes will be needed, so we do it just above. * * Worse, that means that the node may have been utilized while * unpacking other resources, without our weight correction. The most * likely place for this to happen is when common_unpack() calls * resource_location() to set a default score in symmetric clusters. * This adds a node *copy* to each resource's allowed nodes, and these * copies will have the wrong weight. * * As a hacky workaround, fix those copies here. * * @TODO Possible alternative: ensure bundles are unpacked before other * resources, so the weight is correct before any copies are made. */ for (rsc_iter = data_set->resources; rsc_iter; rsc_iter = rsc_iter->next) { disallow_node((resource_t *) (rsc_iter->data), uname); } replica->node = node_copy(node); replica->node->weight = 500; replica->node->rsc_discover_mode = pe_discover_exclusive; /* Ensure the node shows up as allowed and with the correct discovery set */ if (replica->child->allowed_nodes != NULL) { g_hash_table_destroy(replica->child->allowed_nodes); } replica->child->allowed_nodes = g_hash_table_new_full(crm_str_hash, g_str_equal, NULL, free); g_hash_table_insert(replica->child->allowed_nodes, (gpointer) replica->node->details->id, node_copy(replica->node)); { node_t *copy = node_copy(replica->node); copy->weight = -INFINITY; g_hash_table_insert(replica->child->parent->allowed_nodes, (gpointer) replica->node->details->id, copy); } if (!common_unpack(xml_remote, &replica->remote, parent, data_set)) { return FALSE; } g_hash_table_iter_init(&gIter, replica->remote->allowed_nodes); while (g_hash_table_iter_next(&gIter, NULL, (void **)&node)) { if (pe__is_guest_or_remote_node(node)) { /* Remote resources can only run on 'normal' cluster node */ node->weight = -INFINITY; } } replica->node->details->remote_rsc = replica->remote; // Ensure pe__is_guest_node() functions correctly immediately replica->remote->container = replica->container; /* A bundle's #kind is closer to "container" (guest node) than the * "remote" set by pe_create_node(). */ g_hash_table_insert(replica->node->details->attrs, strdup(CRM_ATTR_KIND), strdup("container")); /* One effect of this is that setup_container() will add * replica->remote to replica->container's fillers, which will make * pe__resource_contains_guest_node() true for replica->container. * * replica->child does NOT get added to replica->container's fillers. * The only noticeable effect if it did would be for its fail count to * be taken into account when checking replica->container's migration * threshold. */ parent->children = g_list_append(parent->children, replica->remote); } return TRUE; } static bool create_container(pe_resource_t *parent, pe__bundle_variant_data_t *data, pe__bundle_replica_t *replica, pe_working_set_t *data_set) { switch (data->agent_type) { case PE__CONTAINER_AGENT_DOCKER: if (!create_docker_resource(parent, data, replica, data_set)) { return FALSE; } break; case PE__CONTAINER_AGENT_PODMAN: if (!create_podman_resource(parent, data, replica, data_set)) { return FALSE; } break; case PE__CONTAINER_AGENT_RKT: if (!create_rkt_resource(parent, data, replica, data_set)) { return FALSE; } break; default: // PE__CONTAINER_AGENT_UNKNOWN return FALSE; } if (create_ip_resource(parent, data, replica, data_set) == FALSE) { return FALSE; } if(create_remote_resource(parent, data, replica, data_set) == FALSE) { return FALSE; } if (replica->child && replica->ipaddr) { add_hash_param(replica->child->meta, "external-ip", replica->ipaddr); } if (replica->remote) { /* * Allow the remote connection resource to be allocated to a * different node than the one on which the container is active. * * This makes it possible to have Pacemaker Remote nodes running * containers with pacemaker-remoted inside in order to start * services inside those containers. */ set_bit(replica->remote->flags, pe_rsc_allow_remote_remotes); } return TRUE; } static void mount_add(pe__bundle_variant_data_t *bundle_data, const char *source, const char *target, const char *options, uint32_t flags) { pe__bundle_mount_t *mount = calloc(1, sizeof(pe__bundle_mount_t)); mount->source = strdup(source); mount->target = strdup(target); if (options) { mount->options = strdup(options); } mount->flags = flags; bundle_data->mounts = g_list_append(bundle_data->mounts, mount); } static void mount_free(pe__bundle_mount_t *mount) { free(mount->source); free(mount->target); free(mount->options); free(mount); } static void port_free(pe__bundle_port_t *port) { free(port->source); free(port->target); free(port); } static pe__bundle_replica_t * replica_for_remote(pe_resource_t *remote) { resource_t *top = remote; pe__bundle_variant_data_t *bundle_data = NULL; if (top == NULL) { return NULL; } while (top->parent != NULL) { top = top->parent; } get_bundle_variant_data(bundle_data, top); for (GList *gIter = bundle_data->replicas; gIter != NULL; gIter = gIter->next) { pe__bundle_replica_t *replica = gIter->data; if (replica->remote == remote) { return replica; } } CRM_LOG_ASSERT(FALSE); return NULL; } bool pe__bundle_needs_remote_name(pe_resource_t *rsc) { const char *value; if (rsc == NULL) { return FALSE; } value = g_hash_table_lookup(rsc->parameters, XML_RSC_ATTR_REMOTE_RA_ADDR); if (safe_str_eq(value, "#uname") == FALSE) { return FALSE; } else { const char *match[3][2] = { { XML_ATTR_TYPE, "remote" }, { XML_AGENT_ATTR_CLASS, PCMK_RESOURCE_CLASS_OCF }, { XML_AGENT_ATTR_PROVIDER, "pacemaker" }, }; for (int m = 0; m < 3; m++) { value = crm_element_value(rsc->xml, match[m][0]); if (safe_str_neq(value, match[m][1])) { return FALSE; } } } return TRUE; } const char * pe__add_bundle_remote_name(pe_resource_t *rsc, xmlNode *xml, const char *field) { // REMOTE_CONTAINER_HACK: Allow remote nodes that start containers with pacemaker remote inside pe_node_t *node = NULL; pe__bundle_replica_t *replica = NULL; if (!pe__bundle_needs_remote_name(rsc)) { return NULL; } replica = replica_for_remote(rsc); if (replica == NULL) { return NULL; } node = replica->container->allocated_to; if (node == NULL) { /* If it won't be running anywhere after the * transition, go with where it's running now. */ node = pe__current_node(replica->container); } if(node == NULL) { crm_trace("Cannot determine address for bundle connection %s", rsc->id); return NULL; } crm_trace("Setting address for bundle connection %s to bundle host %s", rsc->id, node->details->uname); if(xml != NULL && field != NULL) { crm_xml_add(xml, field, node->details->uname); } return node->details->uname; } gboolean pe__unpack_bundle(pe_resource_t *rsc, pe_working_set_t *data_set) { const char *value = NULL; xmlNode *xml_obj = NULL; xmlNode *xml_resource = NULL; pe__bundle_variant_data_t *bundle_data = NULL; bool need_log_mount = TRUE; CRM_ASSERT(rsc != NULL); pe_rsc_trace(rsc, "Processing resource %s...", rsc->id); bundle_data = calloc(1, sizeof(pe__bundle_variant_data_t)); rsc->variant_opaque = bundle_data; bundle_data->prefix = strdup(rsc->id); xml_obj = first_named_child(rsc->xml, PE__CONTAINER_AGENT_DOCKER_S); if (xml_obj != NULL) { bundle_data->agent_type = PE__CONTAINER_AGENT_DOCKER; } else { xml_obj = first_named_child(rsc->xml, PE__CONTAINER_AGENT_RKT_S); if (xml_obj != NULL) { bundle_data->agent_type = PE__CONTAINER_AGENT_RKT; } else { xml_obj = first_named_child(rsc->xml, PE__CONTAINER_AGENT_PODMAN_S); if (xml_obj != NULL) { bundle_data->agent_type = PE__CONTAINER_AGENT_PODMAN; } else { return FALSE; } } } value = crm_element_value(xml_obj, XML_RSC_ATTR_PROMOTED_MAX); if (value == NULL) { // @COMPAT deprecated since 2.0.0 value = crm_element_value(xml_obj, "masters"); } bundle_data->promoted_max = crm_parse_int(value, "0"); if (bundle_data->promoted_max < 0) { pe_err("%s for %s must be nonnegative integer, using 0", XML_RSC_ATTR_PROMOTED_MAX, rsc->id); bundle_data->promoted_max = 0; } value = crm_element_value(xml_obj, "replicas"); if ((value == NULL) && bundle_data->promoted_max) { bundle_data->nreplicas = bundle_data->promoted_max; } else { bundle_data->nreplicas = crm_parse_int(value, "1"); } if (bundle_data->nreplicas < 1) { pe_err("'replicas' for %s must be positive integer, using 1", rsc->id); bundle_data->nreplicas = 1; } /* * Communication between containers on the same host via the * floating IPs only works if the container is started with: * --userland-proxy=false --ip-masq=false */ value = crm_element_value(xml_obj, "replicas-per-host"); bundle_data->nreplicas_per_host = crm_parse_int(value, "1"); if (bundle_data->nreplicas_per_host < 1) { pe_err("'replicas-per-host' for %s must be positive integer, using 1", rsc->id); bundle_data->nreplicas_per_host = 1; } if (bundle_data->nreplicas_per_host == 1) { clear_bit(rsc->flags, pe_rsc_unique); } bundle_data->container_command = crm_element_value_copy(xml_obj, "run-command"); bundle_data->launcher_options = crm_element_value_copy(xml_obj, "options"); bundle_data->image = crm_element_value_copy(xml_obj, "image"); bundle_data->container_network = crm_element_value_copy(xml_obj, "network"); xml_obj = first_named_child(rsc->xml, "network"); if(xml_obj) { bundle_data->ip_range_start = crm_element_value_copy(xml_obj, "ip-range-start"); bundle_data->host_netmask = crm_element_value_copy(xml_obj, "host-netmask"); bundle_data->host_network = crm_element_value_copy(xml_obj, "host-interface"); bundle_data->control_port = crm_element_value_copy(xml_obj, "control-port"); value = crm_element_value(xml_obj, "add-host"); if (check_boolean(value) == FALSE) { bundle_data->add_host = TRUE; } else { crm_str_to_boolean(value, &bundle_data->add_host); } for (xmlNode *xml_child = __xml_first_child_element(xml_obj); xml_child != NULL; xml_child = __xml_next_element(xml_child)) { pe__bundle_port_t *port = calloc(1, sizeof(pe__bundle_port_t)); port->source = crm_element_value_copy(xml_child, "port"); if(port->source == NULL) { port->source = crm_element_value_copy(xml_child, "range"); } else { port->target = crm_element_value_copy(xml_child, "internal-port"); } if(port->source != NULL && strlen(port->source) > 0) { if(port->target == NULL) { port->target = strdup(port->source); } bundle_data->ports = g_list_append(bundle_data->ports, port); } else { pe_err("Invalid port directive %s", ID(xml_child)); port_free(port); } } } xml_obj = first_named_child(rsc->xml, "storage"); for (xmlNode *xml_child = __xml_first_child_element(xml_obj); xml_child != NULL; xml_child = __xml_next_element(xml_child)) { const char *source = crm_element_value(xml_child, "source-dir"); const char *target = crm_element_value(xml_child, "target-dir"); const char *options = crm_element_value(xml_child, "options"); int flags = pe__bundle_mount_none; if (source == NULL) { source = crm_element_value(xml_child, "source-dir-root"); set_bit(flags, pe__bundle_mount_subdir); } if (source && target) { mount_add(bundle_data, source, target, options, flags); if (strcmp(target, "/var/log") == 0) { need_log_mount = FALSE; } } else { pe_err("Invalid mount directive %s", ID(xml_child)); } } xml_obj = first_named_child(rsc->xml, "primitive"); if (xml_obj && valid_network(bundle_data)) { char *value = NULL; xmlNode *xml_set = NULL; xml_resource = create_xml_node(NULL, XML_CIB_TAG_INCARNATION); /* @COMPAT We no longer use the tag, but we need to keep it as * part of the resource name, so that bundles don't restart in a rolling * upgrade. (It also avoids needing to change regression tests.) */ crm_xml_set_id(xml_resource, "%s-%s", bundle_data->prefix, (bundle_data->promoted_max? "master" : (const char *)xml_resource->name)); xml_set = create_xml_node(xml_resource, XML_TAG_META_SETS); crm_xml_set_id(xml_set, "%s-%s-meta", bundle_data->prefix, xml_resource->name); crm_create_nvpair_xml(xml_set, NULL, XML_RSC_ATTR_ORDERED, XML_BOOLEAN_TRUE); value = crm_itoa(bundle_data->nreplicas); crm_create_nvpair_xml(xml_set, NULL, XML_RSC_ATTR_INCARNATION_MAX, value); free(value); value = crm_itoa(bundle_data->nreplicas_per_host); crm_create_nvpair_xml(xml_set, NULL, XML_RSC_ATTR_INCARNATION_NODEMAX, value); free(value); crm_create_nvpair_xml(xml_set, NULL, XML_RSC_ATTR_UNIQUE, (bundle_data->nreplicas_per_host > 1)? XML_BOOLEAN_TRUE : XML_BOOLEAN_FALSE); if (bundle_data->promoted_max) { crm_create_nvpair_xml(xml_set, NULL, XML_RSC_ATTR_PROMOTABLE, XML_BOOLEAN_TRUE); value = crm_itoa(bundle_data->promoted_max); crm_create_nvpair_xml(xml_set, NULL, XML_RSC_ATTR_PROMOTED_MAX, value); free(value); } //crm_xml_add(xml_obj, XML_ATTR_ID, bundle_data->prefix); add_node_copy(xml_resource, xml_obj); } else if(xml_obj) { pe_err("Cannot control %s inside %s without either ip-range-start or control-port", rsc->id, ID(xml_obj)); return FALSE; } if(xml_resource) { int lpc = 0; GListPtr childIter = NULL; resource_t *new_rsc = NULL; pe__bundle_port_t *port = NULL; int offset = 0, max = 1024; char *buffer = NULL; if (common_unpack(xml_resource, &new_rsc, rsc, data_set) == FALSE) { pe_err("Failed unpacking resource %s", ID(rsc->xml)); if (new_rsc != NULL && new_rsc->fns != NULL) { new_rsc->fns->free(new_rsc); } return FALSE; } bundle_data->child = new_rsc; /* Currently, we always map the default authentication key location * into the same location inside the container. * * Ideally, we would respect the host's PCMK_authkey_location, but: * - it may be different on different nodes; * - the actual connection will do extra checking to make sure the key * file exists and is readable, that we can't do here on the DC * - tools such as crm_resource and crm_simulate may not have the same * environment variables as the cluster, causing operation digests to * differ * * Always using the default location inside the container is fine, * because we control the pacemaker_remote environment, and it avoids * having to pass another environment variable to the container. * * @TODO A better solution may be to have only pacemaker_remote use the * environment variable, and have the cluster nodes use a new * cluster option for key location. This would introduce the limitation * of the location being the same on all cluster nodes, but that's * reasonable. */ mount_add(bundle_data, DEFAULT_REMOTE_KEY_LOCATION, DEFAULT_REMOTE_KEY_LOCATION, NULL, pe__bundle_mount_none); if (need_log_mount) { mount_add(bundle_data, CRM_BUNDLE_DIR, "/var/log", NULL, pe__bundle_mount_subdir); } port = calloc(1, sizeof(pe__bundle_port_t)); if(bundle_data->control_port) { port->source = strdup(bundle_data->control_port); } else { /* If we wanted to respect PCMK_remote_port, we could use * crm_default_remote_port() here and elsewhere in this file instead * of DEFAULT_REMOTE_PORT. * * However, it gains nothing, since we control both the container * environment and the connection resource parameters, and the user * can use a different port if desired by setting control-port. */ port->source = crm_itoa(DEFAULT_REMOTE_PORT); } port->target = strdup(port->source); bundle_data->ports = g_list_append(bundle_data->ports, port); buffer = calloc(1, max+1); for (childIter = bundle_data->child->children; childIter != NULL; childIter = childIter->next) { pe__bundle_replica_t *replica = calloc(1, sizeof(pe__bundle_replica_t)); replica->child = childIter->data; replica->child->exclusive_discover = TRUE; replica->offset = lpc++; // Ensure the child's notify gets set based on the underlying primitive's value if (is_set(replica->child->flags, pe_rsc_notify)) { set_bit(bundle_data->child->flags, pe_rsc_notify); } offset += allocate_ip(bundle_data, replica, buffer+offset, max-offset); bundle_data->replicas = g_list_append(bundle_data->replicas, replica); bundle_data->attribute_target = g_hash_table_lookup(replica->child->meta, XML_RSC_ATTR_TARGET); } bundle_data->container_host_options = buffer; if (bundle_data->attribute_target) { g_hash_table_replace(rsc->meta, strdup(XML_RSC_ATTR_TARGET), strdup(bundle_data->attribute_target)); g_hash_table_replace(bundle_data->child->meta, strdup(XML_RSC_ATTR_TARGET), strdup(bundle_data->attribute_target)); } } else { // Just a naked container, no pacemaker-remote int offset = 0, max = 1024; char *buffer = calloc(1, max+1); for (int lpc = 0; lpc < bundle_data->nreplicas; lpc++) { pe__bundle_replica_t *replica = calloc(1, sizeof(pe__bundle_replica_t)); replica->offset = lpc; offset += allocate_ip(bundle_data, replica, buffer+offset, max-offset); bundle_data->replicas = g_list_append(bundle_data->replicas, replica); } bundle_data->container_host_options = buffer; } for (GList *gIter = bundle_data->replicas; gIter != NULL; gIter = gIter->next) { pe__bundle_replica_t *replica = gIter->data; if (!create_container(rsc, bundle_data, replica, data_set)) { pe_err("Failed unpacking resource %s", rsc->id); rsc->fns->free(rsc); return FALSE; } } if (bundle_data->child) { rsc->children = g_list_append(rsc->children, bundle_data->child); } return TRUE; } static int replica_resource_active(pe_resource_t *rsc, gboolean all) { if (rsc) { gboolean child_active = rsc->fns->active(rsc, all); if (child_active && !all) { return TRUE; } else if (!child_active && all) { return FALSE; } } return -1; } gboolean pe__bundle_active(pe_resource_t *rsc, gboolean all) { pe__bundle_variant_data_t *bundle_data = NULL; GListPtr iter = NULL; get_bundle_variant_data(bundle_data, rsc); for (iter = bundle_data->replicas; iter != NULL; iter = iter->next) { pe__bundle_replica_t *replica = iter->data; int rsc_active; rsc_active = replica_resource_active(replica->ip, all); if (rsc_active >= 0) { return (gboolean) rsc_active; } rsc_active = replica_resource_active(replica->child, all); if (rsc_active >= 0) { return (gboolean) rsc_active; } rsc_active = replica_resource_active(replica->container, all); if (rsc_active >= 0) { return (gboolean) rsc_active; } rsc_active = replica_resource_active(replica->remote, all); if (rsc_active >= 0) { return (gboolean) rsc_active; } } /* If "all" is TRUE, we've already checked that no resources were inactive, * so return TRUE; if "all" is FALSE, we didn't find any active resources, * so return FALSE. */ return all; } /*! * \internal * \brief Find the bundle replica corresponding to a given node * * \param[in] bundle Top-level bundle resource * \param[in] node Node to search for * * \return Bundle replica if found, NULL otherwise */ pe_resource_t * pe__find_bundle_replica(const pe_resource_t *bundle, const pe_node_t *node) { pe__bundle_variant_data_t *bundle_data = NULL; CRM_ASSERT(bundle && node); get_bundle_variant_data(bundle_data, bundle); for (GList *gIter = bundle_data->replicas; gIter != NULL; gIter = gIter->next) { pe__bundle_replica_t *replica = gIter->data; CRM_ASSERT(replica && replica->node); if (replica->node->details == node->details) { return replica->child; } } return NULL; } static void print_rsc_in_list(resource_t *rsc, const char *pre_text, long options, void *print_data) { if (rsc != NULL) { if (options & pe_print_html) { status_print("
  • "); } rsc->fns->print(rsc, pre_text, options, print_data); if (options & pe_print_html) { status_print("
    • \n"); } } } static const char* container_agent_str(enum pe__container_agent t) { switch (t) { case PE__CONTAINER_AGENT_DOCKER: return PE__CONTAINER_AGENT_DOCKER_S; case PE__CONTAINER_AGENT_RKT: return PE__CONTAINER_AGENT_RKT_S; case PE__CONTAINER_AGENT_PODMAN: return PE__CONTAINER_AGENT_PODMAN_S; default: // PE__CONTAINER_AGENT_UNKNOWN break; } return PE__CONTAINER_AGENT_UNKNOWN_S; } static void bundle_print_xml(pe_resource_t *rsc, const char *pre_text, long options, void *print_data) { pe__bundle_variant_data_t *bundle_data = NULL; char *child_text = NULL; CRM_CHECK(rsc != NULL, return); if (pre_text == NULL) { pre_text = ""; } child_text = crm_concat(pre_text, " ", ' '); get_bundle_variant_data(bundle_data, rsc); status_print("%sid); status_print("type=\"%s\" ", container_agent_str(bundle_data->agent_type)); status_print("image=\"%s\" ", bundle_data->image); status_print("unique=\"%s\" ", is_set(rsc->flags, pe_rsc_unique)? "true" : "false"); status_print("managed=\"%s\" ", is_set(rsc->flags, pe_rsc_managed) ? "true" : "false"); status_print("failed=\"%s\" ", is_set(rsc->flags, pe_rsc_failed) ? "true" : "false"); status_print(">\n"); for (GList *gIter = bundle_data->replicas; gIter != NULL; gIter = gIter->next) { pe__bundle_replica_t *replica = gIter->data; CRM_ASSERT(replica); status_print("%s \n", pre_text, replica->offset); print_rsc_in_list(replica->ip, child_text, options, print_data); print_rsc_in_list(replica->child, child_text, options, print_data); print_rsc_in_list(replica->container, child_text, options, print_data); print_rsc_in_list(replica->remote, child_text, options, print_data); status_print("%s \n", pre_text); } status_print("%s\n", pre_text); free(child_text); } int pe__bundle_xml(pcmk__output_t *out, va_list args) { long options = va_arg(args, int); resource_t *rsc = va_arg(args, resource_t *); pe__bundle_variant_data_t *bundle_data = NULL; int rc = 0; CRM_ASSERT(rsc != NULL); get_bundle_variant_data(bundle_data, rsc); rc = pe__name_and_nvpairs_xml(out, true, "bundle", 6 , "id", rsc->id , "type", container_agent_str(bundle_data->agent_type) , "image", bundle_data->image , "unique", BOOL2STR(is_set(rsc->flags, pe_rsc_unique)) , "managed", BOOL2STR(is_set(rsc->flags, pe_rsc_managed)) , "failed", BOOL2STR(is_set(rsc->flags, pe_rsc_failed))); CRM_ASSERT(rc == 0); for (GList *gIter = bundle_data->replicas; gIter != NULL; gIter = gIter->next) { pe__bundle_replica_t *replica = gIter->data; char *id = crm_itoa(replica->offset); CRM_ASSERT(replica); rc = pe__name_and_nvpairs_xml(out, true, "replica", 1, "id", id); free(id); CRM_ASSERT(rc == 0); if (replica->ip != NULL) { out->message(out, crm_map_element_name(replica->ip->xml), options, replica->ip); } if (replica->child != NULL) { out->message(out, crm_map_element_name(replica->child->xml), options, replica->child); } out->message(out, crm_map_element_name(replica->container->xml), options, replica->container); if (replica->remote != NULL) { out->message(out, crm_map_element_name(replica->remote->xml), options, replica->remote); } pcmk__output_xml_pop_parent(out); // replica } pcmk__output_xml_pop_parent(out); // bundle return rc; } static void pe__bundle_replica_output_html(pcmk__output_t *out, pe__bundle_replica_t *replica, long options) { node_t *node = NULL; pe_resource_t *rsc = replica->child; int offset = 0; char buffer[LINE_MAX]; if(rsc == NULL) { rsc = replica->container; } if (replica->remote) { offset += snprintf(buffer + offset, LINE_MAX - offset, "%s", rsc_printable_id(replica->remote)); } else { offset += snprintf(buffer + offset, LINE_MAX - offset, "%s", rsc_printable_id(replica->container)); } if (replica->ipaddr) { offset += snprintf(buffer + offset, LINE_MAX - offset, " (%s)", replica->ipaddr); } node = pe__current_node(replica->container); pe__common_output_html(out, rsc, buffer, node, options); } int pe__bundle_html(pcmk__output_t *out, va_list args) { long options = va_arg(args, int); resource_t *rsc = va_arg(args, resource_t *); pe__bundle_variant_data_t *bundle_data = NULL; char buffer[LINE_MAX]; CRM_ASSERT(rsc != NULL); get_bundle_variant_data(bundle_data, rsc); pcmk__output_create_xml_node(out, "br"); out->begin_list(out, NULL, NULL, "Container bundle%s: %s [%s]%s%s", (bundle_data->nreplicas > 1)? " set" : "", rsc->id, bundle_data->image, is_set(rsc->flags, pe_rsc_unique) ? " (unique)" : "", is_set(rsc->flags, pe_rsc_managed) ? "" : " (unmanaged)"); for (GList *gIter = bundle_data->replicas; gIter != NULL; gIter = gIter->next) { pe__bundle_replica_t *replica = gIter->data; CRM_ASSERT(replica); pcmk__output_xml_create_parent(out, "li"); if (is_set(options, pe_print_implicit)) { if(g_list_length(bundle_data->replicas) > 1) { snprintf(buffer, LINE_MAX, " Replica[%d]", replica->offset); xmlNodeSetContent(pcmk__output_xml_peek_parent(out), (pcmkXmlStr) buffer); } pcmk__output_create_xml_node(out, "br"); out->begin_list(out, NULL, NULL, NULL); if (replica->ip != NULL) { out->message(out, crm_map_element_name(replica->ip->xml), options, replica->ip); } if (replica->child != NULL) { out->message(out, crm_map_element_name(replica->child->xml), options, replica->child); } out->message(out, crm_map_element_name(replica->container->xml), options, replica->container); if (replica->remote != NULL) { out->message(out, crm_map_element_name(replica->remote->xml), options, replica->remote); } out->end_list(out); } else { pe__bundle_replica_output_html(out, replica, options); } pcmk__output_xml_pop_parent(out); } out->end_list(out); return 0; } static void pe__bundle_replica_output_text(pcmk__output_t *out, pe__bundle_replica_t *replica, long options) { node_t *node = NULL; pe_resource_t *rsc = replica->child; int offset = 0; char buffer[LINE_MAX]; if(rsc == NULL) { rsc = replica->container; } if (replica->remote) { offset += snprintf(buffer + offset, LINE_MAX - offset, "%s", rsc_printable_id(replica->remote)); } else { offset += snprintf(buffer + offset, LINE_MAX - offset, "%s", rsc_printable_id(replica->container)); } if (replica->ipaddr) { offset += snprintf(buffer + offset, LINE_MAX - offset, " (%s)", replica->ipaddr); } node = pe__current_node(replica->container); pe__common_output_text(out, rsc, buffer, node, options); } int pe__bundle_text(pcmk__output_t *out, va_list args) { long options = va_arg(args, int); resource_t *rsc = va_arg(args, resource_t *); pe__bundle_variant_data_t *bundle_data = NULL; CRM_ASSERT(rsc != NULL); get_bundle_variant_data(bundle_data, rsc); out->begin_list(out, NULL, NULL, "Container bundle%s: %s [%s]%s%s", (bundle_data->nreplicas > 1)? " set" : "", rsc->id, bundle_data->image, is_set(rsc->flags, pe_rsc_unique) ? " (unique)" : "", is_set(rsc->flags, pe_rsc_managed) ? "" : " (unmanaged)"); for (GList *gIter = bundle_data->replicas; gIter != NULL; gIter = gIter->next) { pe__bundle_replica_t *replica = gIter->data; CRM_ASSERT(replica); if (is_set(options, pe_print_implicit)) { if(g_list_length(bundle_data->replicas) > 1) { out->list_item(out, NULL, "Replica[%d]", replica->offset); } out->begin_list(out, NULL, NULL, NULL); if (replica->ip != NULL) { out->message(out, crm_map_element_name(replica->ip->xml), options, replica->ip); } if (replica->child != NULL) { out->message(out, crm_map_element_name(replica->child->xml), options, replica->child); } out->message(out, crm_map_element_name(replica->container->xml), options, replica->container); if (replica->remote != NULL) { out->message(out, crm_map_element_name(replica->remote->xml), options, replica->remote); } out->end_list(out); } else { pe__bundle_replica_output_text(out, replica, options); } } out->end_list(out); return 0; } static void print_bundle_replica(pe__bundle_replica_t *replica, const char *pre_text, long options, void *print_data) { node_t *node = NULL; pe_resource_t *rsc = replica->child; int offset = 0; char buffer[LINE_MAX]; if(rsc == NULL) { rsc = replica->container; } if (replica->remote) { offset += snprintf(buffer + offset, LINE_MAX - offset, "%s", rsc_printable_id(replica->remote)); } else { offset += snprintf(buffer + offset, LINE_MAX - offset, "%s", rsc_printable_id(replica->container)); } if (replica->ipaddr) { offset += snprintf(buffer + offset, LINE_MAX - offset, " (%s)", replica->ipaddr); } node = pe__current_node(replica->container); common_print(rsc, pre_text, buffer, node, options, print_data); } void pe__print_bundle(pe_resource_t *rsc, const char *pre_text, long options, void *print_data) { pe__bundle_variant_data_t *bundle_data = NULL; char *child_text = NULL; CRM_CHECK(rsc != NULL, return); if (options & pe_print_xml) { bundle_print_xml(rsc, pre_text, options, print_data); return; } get_bundle_variant_data(bundle_data, rsc); if (pre_text == NULL) { pre_text = " "; } status_print("%sContainer bundle%s: %s [%s]%s%s\n", pre_text, ((bundle_data->nreplicas > 1)? " set" : ""), rsc->id, bundle_data->image, is_set(rsc->flags, pe_rsc_unique) ? " (unique)" : "", is_set(rsc->flags, pe_rsc_managed) ? "" : " (unmanaged)"); if (options & pe_print_html) { status_print("
    \n
      \n"); } for (GList *gIter = bundle_data->replicas; gIter != NULL; gIter = gIter->next) { pe__bundle_replica_t *replica = gIter->data; CRM_ASSERT(replica); if (options & pe_print_html) { status_print("
    • "); } if (is_set(options, pe_print_implicit)) { child_text = crm_strdup_printf(" %s", pre_text); if(g_list_length(bundle_data->replicas) > 1) { status_print(" %sReplica[%d]\n", pre_text, replica->offset); } if (options & pe_print_html) { status_print("
      \n
        \n"); } print_rsc_in_list(replica->ip, child_text, options, print_data); print_rsc_in_list(replica->container, child_text, options, print_data); print_rsc_in_list(replica->remote, child_text, options, print_data); print_rsc_in_list(replica->child, child_text, options, print_data); if (options & pe_print_html) { status_print("
      \n"); } } else { child_text = crm_strdup_printf("%s ", pre_text); print_bundle_replica(replica, child_text, options, print_data); } free(child_text); if (options & pe_print_html) { status_print("
      • \n"); } } if (options & pe_print_html) { status_print("
    \n"); } } static void free_bundle_replica(pe__bundle_replica_t *replica) { if (replica == NULL) { return; } if (replica->node) { free(replica->node); replica->node = NULL; } if (replica->ip) { free_xml(replica->ip->xml); replica->ip->xml = NULL; replica->ip->fns->free(replica->ip); replica->ip = NULL; } if (replica->container) { free_xml(replica->container->xml); replica->container->xml = NULL; replica->container->fns->free(replica->container); replica->container = NULL; } if (replica->remote) { free_xml(replica->remote->xml); replica->remote->xml = NULL; replica->remote->fns->free(replica->remote); replica->remote = NULL; } free(replica->ipaddr); free(replica); } void pe__free_bundle(pe_resource_t *rsc) { pe__bundle_variant_data_t *bundle_data = NULL; CRM_CHECK(rsc != NULL, return); get_bundle_variant_data(bundle_data, rsc); pe_rsc_trace(rsc, "Freeing %s", rsc->id); free(bundle_data->prefix); free(bundle_data->image); free(bundle_data->control_port); free(bundle_data->host_network); free(bundle_data->host_netmask); free(bundle_data->ip_range_start); free(bundle_data->container_network); free(bundle_data->launcher_options); free(bundle_data->container_command); free(bundle_data->container_host_options); g_list_free_full(bundle_data->replicas, (GDestroyNotify) free_bundle_replica); g_list_free_full(bundle_data->mounts, (GDestroyNotify)mount_free); g_list_free_full(bundle_data->ports, (GDestroyNotify)port_free); g_list_free(rsc->children); if(bundle_data->child) { free_xml(bundle_data->child->xml); bundle_data->child->xml = NULL; bundle_data->child->fns->free(bundle_data->child); } common_free(rsc); } enum rsc_role_e pe__bundle_resource_state(const pe_resource_t *rsc, gboolean current) { enum rsc_role_e container_role = RSC_ROLE_UNKNOWN; return container_role; } /*! * \brief Get the number of configured replicas in a bundle * * \param[in] rsc Bundle resource * * \return Number of configured replicas, or 0 on error */ int pe_bundle_replicas(const resource_t *rsc) { if ((rsc == NULL) || (rsc->variant != pe_container)) { return 0; } else { pe__bundle_variant_data_t *bundle_data = NULL; get_bundle_variant_data(bundle_data, rsc); return bundle_data->nreplicas; } } + +void +pe__count_bundle(pe_resource_t *rsc) +{ + pe__bundle_variant_data_t *bundle_data = NULL; + + get_bundle_variant_data(bundle_data, rsc); + for (GList *item = bundle_data->replicas; item != NULL; item = item->next) { + pe__bundle_replica_t *replica = item->data; + + if (replica->ip) { + replica->ip->fns->count(replica->ip); + } + if (replica->child) { + replica->child->fns->count(replica->child); + } + if (replica->container) { + replica->container->fns->count(replica->container); + } + if (replica->remote) { + replica->remote->fns->count(replica->remote); + } + } +} diff --git a/lib/pengine/clone.c b/lib/pengine/clone.c index 57690299c5..b356f54118 100644 --- a/lib/pengine/clone.c +++ b/lib/pengine/clone.c @@ -1,1045 +1,1044 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU Lesser General Public License * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #define VARIANT_CLONE 1 #include "./variant.h" void pe__force_anon(const char *standard, pe_resource_t *rsc, const char *rid, pe_working_set_t *data_set) { if (pe_rsc_is_clone(rsc)) { clone_variant_data_t *clone_data = NULL; get_clone_variant_data(clone_data, rsc); pe_warn("Ignoring " XML_RSC_ATTR_UNIQUE " for %s because %s resources " "such as %s can be used only as anonymous clones", rsc->id, standard, rid); clone_data->clone_node_max = 1; clone_data->clone_max = QB_MIN(clone_data->clone_max, g_list_length(data_set->nodes)); } } resource_t * find_clone_instance(resource_t * rsc, const char *sub_id, pe_working_set_t * data_set) { char *child_id = NULL; resource_t *child = NULL; const char *child_base = NULL; clone_variant_data_t *clone_data = NULL; get_clone_variant_data(clone_data, rsc); child_base = ID(clone_data->xml_obj_child); child_id = crm_concat(child_base, sub_id, ':'); child = pe_find_resource(rsc->children, child_id); free(child_id); return child; } pe_resource_t * pe__create_clone_child(pe_resource_t *rsc, pe_working_set_t *data_set) { gboolean as_orphan = FALSE; char *inc_num = NULL; char *inc_max = NULL; resource_t *child_rsc = NULL; xmlNode *child_copy = NULL; clone_variant_data_t *clone_data = NULL; get_clone_variant_data(clone_data, rsc); CRM_CHECK(clone_data->xml_obj_child != NULL, return FALSE); if (clone_data->total_clones >= clone_data->clone_max) { // If we've already used all available instances, this is an orphan as_orphan = TRUE; } // Allocate instance numbers in numerical order (starting at 0) inc_num = crm_itoa(clone_data->total_clones); inc_max = crm_itoa(clone_data->clone_max); child_copy = copy_xml(clone_data->xml_obj_child); crm_xml_add(child_copy, XML_RSC_ATTR_INCARNATION, inc_num); if (common_unpack(child_copy, &child_rsc, rsc, data_set) == FALSE) { pe_err("Failed unpacking resource %s", crm_element_value(child_copy, XML_ATTR_ID)); child_rsc = NULL; goto bail; } /* child_rsc->globally_unique = rsc->globally_unique; */ CRM_ASSERT(child_rsc); clone_data->total_clones += 1; pe_rsc_trace(child_rsc, "Setting clone attributes for: %s", child_rsc->id); rsc->children = g_list_append(rsc->children, child_rsc); if (as_orphan) { set_bit_recursive(child_rsc, pe_rsc_orphan); } add_hash_param(child_rsc->meta, XML_RSC_ATTR_INCARNATION_MAX, inc_max); - - print_resource(LOG_TRACE, "Added ", child_rsc, FALSE); + pe_rsc_trace(rsc, "Added %s instance %s", rsc->id, child_rsc->id); bail: free(inc_num); free(inc_max); return child_rsc; } gboolean clone_unpack(resource_t * rsc, pe_working_set_t * data_set) { int lpc = 0; xmlNode *a_child = NULL; xmlNode *xml_obj = rsc->xml; clone_variant_data_t *clone_data = NULL; const char *ordered = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_ORDERED); const char *max_clones = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_INCARNATION_MAX); const char *max_clones_node = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_INCARNATION_NODEMAX); pe_rsc_trace(rsc, "Processing resource %s...", rsc->id); clone_data = calloc(1, sizeof(clone_variant_data_t)); rsc->variant_opaque = clone_data; if (is_set(rsc->flags, pe_rsc_promotable)) { const char *promoted_max = NULL; const char *promoted_node_max = NULL; promoted_max = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_PROMOTED_MAX); if (promoted_max == NULL) { // @COMPAT deprecated since 2.0.0 promoted_max = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_MASTER_MAX); } promoted_node_max = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_PROMOTED_NODEMAX); if (promoted_node_max == NULL) { // @COMPAT deprecated since 2.0.0 promoted_node_max = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_MASTER_NODEMAX); } clone_data->promoted_max = crm_parse_int(promoted_max, "1"); clone_data->promoted_node_max = crm_parse_int(promoted_node_max, "1"); } // Implied by calloc() /* clone_data->xml_obj_child = NULL; */ clone_data->clone_node_max = crm_parse_int(max_clones_node, "1"); if (max_clones) { clone_data->clone_max = crm_parse_int(max_clones, "1"); } else if (g_list_length(data_set->nodes) > 0) { clone_data->clone_max = g_list_length(data_set->nodes); } else { clone_data->clone_max = 1; /* Handy during crm_verify */ } clone_data->ordered = crm_is_true(ordered); if ((rsc->flags & pe_rsc_unique) == 0 && clone_data->clone_node_max > 1) { crm_config_err("Anonymous clones (%s) may only support one copy per node", rsc->id); clone_data->clone_node_max = 1; } pe_rsc_trace(rsc, "Options for %s", rsc->id); pe_rsc_trace(rsc, "\tClone max: %d", clone_data->clone_max); pe_rsc_trace(rsc, "\tClone node max: %d", clone_data->clone_node_max); pe_rsc_trace(rsc, "\tClone is unique: %s", is_set(rsc->flags, pe_rsc_unique) ? "true" : "false"); pe_rsc_trace(rsc, "\tClone is promotable: %s", is_set(rsc->flags, pe_rsc_promotable) ? "true" : "false"); // Clones may contain a single group or primitive for (a_child = __xml_first_child_element(xml_obj); a_child != NULL; a_child = __xml_next_element(a_child)) { if (crm_str_eq((const char *)a_child->name, XML_CIB_TAG_RESOURCE, TRUE) || crm_str_eq((const char *)a_child->name, XML_CIB_TAG_GROUP, TRUE)) { clone_data->xml_obj_child = a_child; break; } } if (clone_data->xml_obj_child == NULL) { crm_config_err("%s has nothing to clone", rsc->id); return FALSE; } /* * Make clones ever so slightly sticky by default * * This helps ensure clone instances are not shuffled around the cluster * for no benefit in situations when pre-allocation is not appropriate */ if (g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_STICKINESS) == NULL) { add_hash_param(rsc->meta, XML_RSC_ATTR_STICKINESS, "1"); } /* This ensures that the globally-unique value always exists for children to * inherit when being unpacked, as well as in resource agents' environment. */ add_hash_param(rsc->meta, XML_RSC_ATTR_UNIQUE, is_set(rsc->flags, pe_rsc_unique) ? XML_BOOLEAN_TRUE : XML_BOOLEAN_FALSE); if (clone_data->clone_max <= 0) { /* Create one child instance so that unpack_find_resource() will hook up * any orphans up to the parent correctly. */ if (pe__create_clone_child(rsc, data_set) == NULL) { return FALSE; } } else { // Create a child instance for each available instance number for (lpc = 0; lpc < clone_data->clone_max; lpc++) { if (pe__create_clone_child(rsc, data_set) == NULL) { return FALSE; } } } pe_rsc_trace(rsc, "Added %d children to resource %s...", clone_data->clone_max, rsc->id); return TRUE; } gboolean clone_active(resource_t * rsc, gboolean all) { GListPtr gIter = rsc->children; for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; gboolean child_active = child_rsc->fns->active(child_rsc, all); if (all == FALSE && child_active) { return TRUE; } else if (all && child_active == FALSE) { return FALSE; } } if (all) { return TRUE; } else { return FALSE; } } static void short_print(char *list, const char *prefix, const char *type, const char *suffix, long options, void *print_data) { if(suffix == NULL) { suffix = ""; } if (list) { if (options & pe_print_html) { status_print("
  • "); } status_print("%s%s: [%s ]%s", prefix, type, list, suffix); if (options & pe_print_html) { status_print("
    • \n"); } else if (options & pe_print_suppres_nl) { /* nothing */ } else if ((options & pe_print_printf) || (options & pe_print_ncurses)) { status_print("\n"); } } } static const char * configured_role_str(resource_t * rsc) { const char *target_role = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_TARGET_ROLE); if ((target_role == NULL) && rsc->children && rsc->children->data) { target_role = g_hash_table_lookup(((resource_t*)rsc->children->data)->meta, XML_RSC_ATTR_TARGET_ROLE); } return target_role; } static enum rsc_role_e configured_role(resource_t * rsc) { const char *target_role = configured_role_str(rsc); if (target_role) { return text2role(target_role); } return RSC_ROLE_UNKNOWN; } static void clone_print_xml(resource_t * rsc, const char *pre_text, long options, void *print_data) { char *child_text = crm_concat(pre_text, " ", ' '); const char *target_role = configured_role_str(rsc); GListPtr gIter = rsc->children; status_print("%sid); status_print("multi_state=\"%s\" ", is_set(rsc->flags, pe_rsc_promotable)? "true" : "false"); status_print("unique=\"%s\" ", is_set(rsc->flags, pe_rsc_unique) ? "true" : "false"); status_print("managed=\"%s\" ", is_set(rsc->flags, pe_rsc_managed) ? "true" : "false"); status_print("failed=\"%s\" ", is_set(rsc->flags, pe_rsc_failed) ? "true" : "false"); status_print("failure_ignored=\"%s\" ", is_set(rsc->flags, pe_rsc_failure_ignored) ? "true" : "false"); if (target_role) { status_print("target_role=\"%s\" ", target_role); } status_print(">\n"); for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; child_rsc->fns->print(child_rsc, child_text, options, print_data); } status_print("%s\n", pre_text); free(child_text); } bool is_set_recursive(resource_t * rsc, long long flag, bool any) { GListPtr gIter; bool all = !any; if(is_set(rsc->flags, flag)) { if(any) { return TRUE; } } else if(all) { return FALSE; } for (gIter = rsc->children; gIter != NULL; gIter = gIter->next) { if(is_set_recursive(gIter->data, flag, any)) { if(any) { return TRUE; } } else if(all) { return FALSE; } } if(all) { return TRUE; } return FALSE; } void clone_print(resource_t * rsc, const char *pre_text, long options, void *print_data) { char *list_text = NULL; char *child_text = NULL; char *stopped_list = NULL; GListPtr master_list = NULL; GListPtr started_list = NULL; GListPtr gIter = rsc->children; clone_variant_data_t *clone_data = NULL; int active_instances = 0; if (pre_text == NULL) { pre_text = " "; } if (options & pe_print_xml) { clone_print_xml(rsc, pre_text, options, print_data); return; } get_clone_variant_data(clone_data, rsc); child_text = crm_concat(pre_text, " ", ' '); status_print("%sClone Set: %s [%s]%s%s%s", pre_text ? pre_text : "", rsc->id, ID(clone_data->xml_obj_child), is_set(rsc->flags, pe_rsc_promotable) ? " (promotable)" : "", is_set(rsc->flags, pe_rsc_unique) ? " (unique)" : "", is_set(rsc->flags, pe_rsc_managed) ? "" : " (unmanaged)"); if (options & pe_print_html) { status_print("\n
      \n"); } else if ((options & pe_print_log) == 0) { status_print("\n"); } for (; gIter != NULL; gIter = gIter->next) { gboolean print_full = FALSE; resource_t *child_rsc = (resource_t *) gIter->data; gboolean partially_active = child_rsc->fns->active(child_rsc, FALSE); if (options & pe_print_clone_details) { print_full = TRUE; } if (is_set(rsc->flags, pe_rsc_unique)) { // Print individual instance when unique (except stopped orphans) if (partially_active || is_not_set(rsc->flags, pe_rsc_orphan)) { print_full = TRUE; } // Everything else in this block is for anonymous clones } else if (is_set(options, pe_print_pending) && (child_rsc->pending_task != NULL) && strcmp(child_rsc->pending_task, "probe")) { // Print individual instance when non-probe action is pending print_full = TRUE; } else if (partially_active == FALSE) { // List stopped instances when requested (except orphans) if (is_not_set(child_rsc->flags, pe_rsc_orphan) && is_not_set(options, pe_print_clone_active)) { stopped_list = add_list_element(stopped_list, child_rsc->id); } } else if (is_set_recursive(child_rsc, pe_rsc_orphan, TRUE) || is_set_recursive(child_rsc, pe_rsc_managed, FALSE) == FALSE || is_set_recursive(child_rsc, pe_rsc_failed, TRUE)) { // Print individual instance when active orphaned/unmanaged/failed print_full = TRUE; } else if (child_rsc->fns->active(child_rsc, TRUE)) { // Instance of fully active anonymous clone node_t *location = child_rsc->fns->location(child_rsc, NULL, TRUE); if (location) { // Instance is active on a single node enum rsc_role_e a_role = child_rsc->fns->state(child_rsc, TRUE); if (location->details->online == FALSE && location->details->unclean) { print_full = TRUE; } else if (a_role > RSC_ROLE_SLAVE) { master_list = g_list_append(master_list, location); } else { started_list = g_list_append(started_list, location); } } else { /* uncolocated group - bleh */ print_full = TRUE; } } else { // Instance of partially active anonymous clone print_full = TRUE; } if (print_full) { if (options & pe_print_html) { status_print("
    • \n"); } child_rsc->fns->print(child_rsc, child_text, options, print_data); if (options & pe_print_html) { status_print("
      • \n"); } } } /* Masters */ master_list = g_list_sort(master_list, sort_node_uname); for (gIter = master_list; gIter; gIter = gIter->next) { node_t *host = gIter->data; list_text = add_list_element(list_text, host->details->uname); active_instances++; } short_print(list_text, child_text, "Masters", NULL, options, print_data); g_list_free(master_list); free(list_text); list_text = NULL; /* Started/Slaves */ started_list = g_list_sort(started_list, sort_node_uname); for (gIter = started_list; gIter; gIter = gIter->next) { node_t *host = gIter->data; list_text = add_list_element(list_text, host->details->uname); active_instances++; } if (is_set(rsc->flags, pe_rsc_promotable)) { enum rsc_role_e role = configured_role(rsc); if(role == RSC_ROLE_SLAVE) { short_print(list_text, child_text, "Slaves (target-role)", NULL, options, print_data); } else { short_print(list_text, child_text, "Slaves", NULL, options, print_data); } } else { short_print(list_text, child_text, "Started", NULL, options, print_data); } g_list_free(started_list); free(list_text); list_text = NULL; if (is_not_set(options, pe_print_clone_active)) { const char *state = "Stopped"; enum rsc_role_e role = configured_role(rsc); if (role == RSC_ROLE_STOPPED) { state = "Stopped (disabled)"; } if (is_not_set(rsc->flags, pe_rsc_unique) && (clone_data->clone_max > active_instances)) { GListPtr nIter; GListPtr list = g_hash_table_get_values(rsc->allowed_nodes); /* Custom stopped list for non-unique clones */ free(stopped_list); stopped_list = NULL; if (g_list_length(list) == 0) { /* Clusters with symmetrical=false haven't calculated allowed_nodes yet * If we've not probed for them yet, the Stopped list will be empty */ list = g_hash_table_get_values(rsc->known_on); } list = g_list_sort(list, sort_node_uname); for (nIter = list; nIter != NULL; nIter = nIter->next) { node_t *node = (node_t *)nIter->data; if (pe_find_node(rsc->running_on, node->details->uname) == NULL) { stopped_list = add_list_element(stopped_list, node->details->uname); } } g_list_free(list); } short_print(stopped_list, child_text, state, NULL, options, print_data); free(stopped_list); } if (options & pe_print_html) { status_print("
    \n"); } free(child_text); } int pe__clone_xml(pcmk__output_t *out, va_list args) { long options = va_arg(args, long); resource_t *rsc = va_arg(args, resource_t *); GListPtr gIter = rsc->children; int rc = pe__name_and_nvpairs_xml(out, true, "clone", 7 , "id", rsc->id , "multi_state", BOOL2STR(is_set(rsc->flags, pe_rsc_promotable)) , "unique", BOOL2STR(is_set(rsc->flags, pe_rsc_unique)) , "managed", BOOL2STR(is_set(rsc->flags, pe_rsc_managed)) , "failed", BOOL2STR(is_set(rsc->flags, pe_rsc_failed)) , "failure_ignored", BOOL2STR(is_set(rsc->flags, pe_rsc_failure_ignored)) , "target_role", configured_role_str(rsc)); CRM_ASSERT(rc == 0); for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; out->message(out, crm_map_element_name(child_rsc->xml), options, child_rsc); } pcmk__output_xml_pop_parent(out); return rc; } int pe__clone_html(pcmk__output_t *out, va_list args) { long options = va_arg(args, long); resource_t *rsc = va_arg(args, resource_t *); char *list_text = NULL; char *stopped_list = NULL; GListPtr master_list = NULL; GListPtr started_list = NULL; GListPtr gIter = rsc->children; clone_variant_data_t *clone_data = NULL; int active_instances = 0; get_clone_variant_data(clone_data, rsc); out->begin_list(out, NULL, NULL, "Clone Set: %s [%s]%s%s%s", rsc->id, ID(clone_data->xml_obj_child), is_set(rsc->flags, pe_rsc_promotable) ? " (promotable)" : "", is_set(rsc->flags, pe_rsc_unique) ? " (unique)" : "", is_set(rsc->flags, pe_rsc_managed) ? "" : " (unmanaged)"); for (; gIter != NULL; gIter = gIter->next) { gboolean print_full = FALSE; resource_t *child_rsc = (resource_t *) gIter->data; gboolean partially_active = child_rsc->fns->active(child_rsc, FALSE); if (options & pe_print_clone_details) { print_full = TRUE; } if (is_set(rsc->flags, pe_rsc_unique)) { // Print individual instance when unique (except stopped orphans) if (partially_active || is_not_set(rsc->flags, pe_rsc_orphan)) { print_full = TRUE; } // Everything else in this block is for anonymous clones } else if (is_set(options, pe_print_pending) && (child_rsc->pending_task != NULL) && strcmp(child_rsc->pending_task, "probe")) { // Print individual instance when non-probe action is pending print_full = TRUE; } else if (partially_active == FALSE) { // List stopped instances when requested (except orphans) if (is_not_set(child_rsc->flags, pe_rsc_orphan) && is_not_set(options, pe_print_clone_active)) { stopped_list = add_list_element(stopped_list, child_rsc->id); } } else if (is_set_recursive(child_rsc, pe_rsc_orphan, TRUE) || is_set_recursive(child_rsc, pe_rsc_managed, FALSE) == FALSE || is_set_recursive(child_rsc, pe_rsc_failed, TRUE)) { // Print individual instance when active orphaned/unmanaged/failed print_full = TRUE; } else if (child_rsc->fns->active(child_rsc, TRUE)) { // Instance of fully active anonymous clone node_t *location = child_rsc->fns->location(child_rsc, NULL, TRUE); if (location) { // Instance is active on a single node enum rsc_role_e a_role = child_rsc->fns->state(child_rsc, TRUE); if (location->details->online == FALSE && location->details->unclean) { print_full = TRUE; } else if (a_role > RSC_ROLE_SLAVE) { master_list = g_list_append(master_list, location); } else { started_list = g_list_append(started_list, location); } } else { /* uncolocated group - bleh */ print_full = TRUE; } } else { // Instance of partially active anonymous clone print_full = TRUE; } if (print_full) { out->message(out, crm_map_element_name(child_rsc->xml), options, child_rsc); } } /* Masters */ master_list = g_list_sort(master_list, sort_node_uname); for (gIter = master_list; gIter; gIter = gIter->next) { node_t *host = gIter->data; list_text = add_list_element(list_text, host->details->uname); active_instances++; } if (list_text != NULL) { out->list_item(out, NULL, " Masters: [%s ]", list_text); g_list_free(master_list); free(list_text); list_text = NULL; } /* Started/Slaves */ started_list = g_list_sort(started_list, sort_node_uname); for (gIter = started_list; gIter; gIter = gIter->next) { node_t *host = gIter->data; list_text = add_list_element(list_text, host->details->uname); active_instances++; } if (list_text != NULL) { if (is_set(rsc->flags, pe_rsc_promotable)) { enum rsc_role_e role = configured_role(rsc); if(role == RSC_ROLE_SLAVE) { out->list_item(out, NULL, " Slaves (target-role): [%s ]", list_text); } else { out->list_item(out, NULL, " Slaves: [%s ]", list_text); } } else { out->list_item(out, NULL, " Started: [%s ]", list_text); } g_list_free(started_list); free(list_text); list_text = NULL; } if (is_not_set(options, pe_print_clone_active)) { const char *state = "Stopped"; enum rsc_role_e role = configured_role(rsc); if (role == RSC_ROLE_STOPPED) { state = "Stopped (disabled)"; } if (is_not_set(rsc->flags, pe_rsc_unique) && (clone_data->clone_max > active_instances)) { GListPtr nIter; GListPtr list = g_hash_table_get_values(rsc->allowed_nodes); /* Custom stopped list for non-unique clones */ free(stopped_list); stopped_list = NULL; if (g_list_length(list) == 0) { /* Clusters with symmetrical=false haven't calculated allowed_nodes yet * If we've not probed for them yet, the Stopped list will be empty */ list = g_hash_table_get_values(rsc->known_on); } list = g_list_sort(list, sort_node_uname); for (nIter = list; nIter != NULL; nIter = nIter->next) { node_t *node = (node_t *)nIter->data; if (pe_find_node(rsc->running_on, node->details->uname) == NULL) { stopped_list = add_list_element(stopped_list, node->details->uname); } } g_list_free(list); } if (stopped_list != NULL) { out->list_item(out, NULL, " %s: [%s ]", state, stopped_list); free(stopped_list); } } out->end_list(out); return 0; } int pe__clone_text(pcmk__output_t *out, va_list args) { long options = va_arg(args, long); resource_t *rsc = va_arg(args, resource_t *); char *list_text = NULL; char *stopped_list = NULL; GListPtr master_list = NULL; GListPtr started_list = NULL; GListPtr gIter = rsc->children; clone_variant_data_t *clone_data = NULL; int active_instances = 0; get_clone_variant_data(clone_data, rsc); out->begin_list(out, NULL, NULL, "Clone Set: %s [%s]%s%s%s", rsc->id, ID(clone_data->xml_obj_child), is_set(rsc->flags, pe_rsc_promotable) ? " (promotable)" : "", is_set(rsc->flags, pe_rsc_unique) ? " (unique)" : "", is_set(rsc->flags, pe_rsc_managed) ? "" : " (unmanaged)"); for (; gIter != NULL; gIter = gIter->next) { gboolean print_full = FALSE; resource_t *child_rsc = (resource_t *) gIter->data; gboolean partially_active = child_rsc->fns->active(child_rsc, FALSE); if (options & pe_print_clone_details) { print_full = TRUE; } if (is_set(rsc->flags, pe_rsc_unique)) { // Print individual instance when unique (except stopped orphans) if (partially_active || is_not_set(rsc->flags, pe_rsc_orphan)) { print_full = TRUE; } // Everything else in this block is for anonymous clones } else if (is_set(options, pe_print_pending) && (child_rsc->pending_task != NULL) && strcmp(child_rsc->pending_task, "probe")) { // Print individual instance when non-probe action is pending print_full = TRUE; } else if (partially_active == FALSE) { // List stopped instances when requested (except orphans) if (is_not_set(child_rsc->flags, pe_rsc_orphan) && is_not_set(options, pe_print_clone_active)) { stopped_list = add_list_element(stopped_list, child_rsc->id); } } else if (is_set_recursive(child_rsc, pe_rsc_orphan, TRUE) || is_set_recursive(child_rsc, pe_rsc_managed, FALSE) == FALSE || is_set_recursive(child_rsc, pe_rsc_failed, TRUE)) { // Print individual instance when active orphaned/unmanaged/failed print_full = TRUE; } else if (child_rsc->fns->active(child_rsc, TRUE)) { // Instance of fully active anonymous clone node_t *location = child_rsc->fns->location(child_rsc, NULL, TRUE); if (location) { // Instance is active on a single node enum rsc_role_e a_role = child_rsc->fns->state(child_rsc, TRUE); if (location->details->online == FALSE && location->details->unclean) { print_full = TRUE; } else if (a_role > RSC_ROLE_SLAVE) { master_list = g_list_append(master_list, location); } else { started_list = g_list_append(started_list, location); } } else { /* uncolocated group - bleh */ print_full = TRUE; } } else { // Instance of partially active anonymous clone print_full = TRUE; } if (print_full) { out->message(out, crm_map_element_name(child_rsc->xml), options, child_rsc); } } /* Masters */ master_list = g_list_sort(master_list, sort_node_uname); for (gIter = master_list; gIter; gIter = gIter->next) { node_t *host = gIter->data; list_text = add_list_element(list_text, host->details->uname); active_instances++; } if (list_text != NULL) { out->list_item(out, "Masters", "[%s ]", list_text); g_list_free(master_list); free(list_text); list_text = NULL; } /* Started/Slaves */ started_list = g_list_sort(started_list, sort_node_uname); for (gIter = started_list; gIter; gIter = gIter->next) { node_t *host = gIter->data; list_text = add_list_element(list_text, host->details->uname); active_instances++; } if (list_text != NULL) { if (is_set(rsc->flags, pe_rsc_promotable)) { enum rsc_role_e role = configured_role(rsc); if(role == RSC_ROLE_SLAVE) { out->list_item(out, "Slaves (target-role)", "[%s ]", list_text); } else { out->list_item(out, "Slaves", "[%s ]", list_text); } } else { out->list_item(out, "Started", "[%s ]", list_text); } g_list_free(started_list); free(list_text); list_text = NULL; } if (is_not_set(options, pe_print_clone_active)) { const char *state = "Stopped"; enum rsc_role_e role = configured_role(rsc); if (role == RSC_ROLE_STOPPED) { state = "Stopped (disabled)"; } if (is_not_set(rsc->flags, pe_rsc_unique) && (clone_data->clone_max > active_instances)) { GListPtr nIter; GListPtr list = g_hash_table_get_values(rsc->allowed_nodes); /* Custom stopped list for non-unique clones */ free(stopped_list); stopped_list = NULL; if (g_list_length(list) == 0) { /* Clusters with symmetrical=false haven't calculated allowed_nodes yet * If we've not probed for them yet, the Stopped list will be empty */ list = g_hash_table_get_values(rsc->known_on); } list = g_list_sort(list, sort_node_uname); for (nIter = list; nIter != NULL; nIter = nIter->next) { node_t *node = (node_t *)nIter->data; if (pe_find_node(rsc->running_on, node->details->uname) == NULL) { stopped_list = add_list_element(stopped_list, node->details->uname); } } g_list_free(list); } if (stopped_list != NULL) { out->list_item(out, state, "[%s ]", stopped_list); free(stopped_list); } } out->end_list(out); return 0; } void clone_free(resource_t * rsc) { clone_variant_data_t *clone_data = NULL; get_clone_variant_data(clone_data, rsc); pe_rsc_trace(rsc, "Freeing %s", rsc->id); for (GListPtr gIter = rsc->children; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; CRM_ASSERT(child_rsc); pe_rsc_trace(child_rsc, "Freeing child %s", child_rsc->id); free_xml(child_rsc->xml); child_rsc->xml = NULL; /* There could be a saved unexpanded xml */ free_xml(child_rsc->orig_xml); child_rsc->orig_xml = NULL; child_rsc->fns->free(child_rsc); } g_list_free(rsc->children); if (clone_data) { CRM_ASSERT(clone_data->demote_notify == NULL); CRM_ASSERT(clone_data->stop_notify == NULL); CRM_ASSERT(clone_data->start_notify == NULL); CRM_ASSERT(clone_data->promote_notify == NULL); } common_free(rsc); } enum rsc_role_e clone_resource_state(const resource_t * rsc, gboolean current) { enum rsc_role_e clone_role = RSC_ROLE_UNKNOWN; GListPtr gIter = rsc->children; for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; enum rsc_role_e a_role = child_rsc->fns->state(child_rsc, current); if (a_role > clone_role) { clone_role = a_role; } } pe_rsc_trace(rsc, "%s role: %s", rsc->id, role2text(clone_role)); return clone_role; } /*! * \internal * \brief Check whether a clone has an instance for every node * * \param[in] rsc Clone to check * \param[in] data_set Cluster state */ bool pe__is_universal_clone(pe_resource_t *rsc, pe_working_set_t *data_set) { if (pe_rsc_is_clone(rsc)) { clone_variant_data_t *clone_data = NULL; get_clone_variant_data(clone_data, rsc); if (clone_data->clone_max == g_list_length(data_set->nodes)) { return TRUE; } } return FALSE; } diff --git a/lib/pengine/complex.c b/lib/pengine/complex.c index ef9dacad61..e6b569ee97 100644 --- a/lib/pengine/complex.c +++ b/lib/pengine/complex.c @@ -1,942 +1,966 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU Lesser General Public License * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include void populate_hash(xmlNode * nvpair_list, GHashTable * hash, const char **attrs, int attrs_length); resource_object_functions_t resource_class_functions[] = { { - native_unpack, - native_find_rsc, - native_parameter, - native_print, - native_active, - native_resource_state, - native_location, - native_free + native_unpack, + native_find_rsc, + native_parameter, + native_print, + native_active, + native_resource_state, + native_location, + native_free, + pe__count_common, }, { - group_unpack, - native_find_rsc, - native_parameter, - group_print, - group_active, - group_resource_state, - native_location, - group_free + group_unpack, + native_find_rsc, + native_parameter, + group_print, + group_active, + group_resource_state, + native_location, + group_free, + pe__count_common, }, { - clone_unpack, - native_find_rsc, - native_parameter, - clone_print, - clone_active, - clone_resource_state, - native_location, - clone_free + clone_unpack, + native_find_rsc, + native_parameter, + clone_print, + clone_active, + clone_resource_state, + native_location, + clone_free, + pe__count_common, }, { - pe__unpack_bundle, - native_find_rsc, - native_parameter, - pe__print_bundle, - pe__bundle_active, - pe__bundle_resource_state, - native_location, - pe__free_bundle + pe__unpack_bundle, + native_find_rsc, + native_parameter, + pe__print_bundle, + pe__bundle_active, + pe__bundle_resource_state, + native_location, + pe__free_bundle, + pe__count_bundle, } }; static enum pe_obj_types get_resource_type(const char *name) { if (safe_str_eq(name, XML_CIB_TAG_RESOURCE)) { return pe_native; } else if (safe_str_eq(name, XML_CIB_TAG_GROUP)) { return pe_group; } else if (safe_str_eq(name, XML_CIB_TAG_INCARNATION)) { return pe_clone; } else if (safe_str_eq(name, XML_CIB_TAG_MASTER)) { // @COMPAT deprecated since 2.0.0 return pe_clone; } else if (safe_str_eq(name, XML_CIB_TAG_CONTAINER)) { return pe_container; } return pe_unknown; } static void dup_attr(gpointer key, gpointer value, gpointer user_data) { add_hash_param(user_data, key, value); } void get_meta_attributes(GHashTable * meta_hash, resource_t * rsc, node_t * node, pe_working_set_t * data_set) { GHashTable *node_hash = NULL; if (node) { node_hash = node->details->attrs; } if (rsc->xml) { xmlAttrPtr xIter = NULL; for (xIter = rsc->xml->properties; xIter; xIter = xIter->next) { const char *prop_name = (const char *)xIter->name; const char *prop_value = crm_element_value(rsc->xml, prop_name); add_hash_param(meta_hash, prop_name, prop_value); } } pe__unpack_dataset_nvpairs(rsc->xml, XML_TAG_META_SETS, node_hash, meta_hash, NULL, FALSE, data_set); /* set anything else based on the parent */ if (rsc->parent != NULL) { g_hash_table_foreach(rsc->parent->meta, dup_attr, meta_hash); } /* and finally check the defaults */ pe__unpack_dataset_nvpairs(data_set->rsc_defaults, XML_TAG_META_SETS, node_hash, meta_hash, NULL, FALSE, data_set); } void get_rsc_attributes(GHashTable * meta_hash, resource_t * rsc, node_t * node, pe_working_set_t * data_set) { GHashTable *node_hash = NULL; if (node) { node_hash = node->details->attrs; } pe__unpack_dataset_nvpairs(rsc->xml, XML_TAG_ATTR_SETS, node_hash, meta_hash, NULL, FALSE, data_set); /* set anything else based on the parent */ if (rsc->parent != NULL) { get_rsc_attributes(meta_hash, rsc->parent, node, data_set); } else { /* and finally check the defaults */ pe__unpack_dataset_nvpairs(data_set->rsc_defaults, XML_TAG_ATTR_SETS, node_hash, meta_hash, NULL, FALSE, data_set); } } #if ENABLE_VERSIONED_ATTRS void pe_get_versioned_attributes(xmlNode * meta_hash, resource_t * rsc, node_t * node, pe_working_set_t * data_set) { GHashTable *node_hash = NULL; if (node) { node_hash = node->details->attrs; } pe_unpack_versioned_attributes(data_set->input, rsc->xml, XML_TAG_ATTR_SETS, node_hash, meta_hash, data_set->now, NULL); /* set anything else based on the parent */ if (rsc->parent != NULL) { pe_get_versioned_attributes(meta_hash, rsc->parent, node, data_set); } else { /* and finally check the defaults */ pe_unpack_versioned_attributes(data_set->input, data_set->rsc_defaults, XML_TAG_ATTR_SETS, node_hash, meta_hash, data_set->now, NULL); } } #endif static char * template_op_key(xmlNode * op) { const char *name = crm_element_value(op, "name"); const char *role = crm_element_value(op, "role"); char *key = NULL; if (role == NULL || crm_str_eq(role, RSC_ROLE_STARTED_S, TRUE) || crm_str_eq(role, RSC_ROLE_SLAVE_S, TRUE)) { role = RSC_ROLE_UNKNOWN_S; } key = crm_concat(name, role, '-'); return key; } static gboolean unpack_template(xmlNode * xml_obj, xmlNode ** expanded_xml, pe_working_set_t * data_set) { xmlNode *cib_resources = NULL; xmlNode *template = NULL; xmlNode *new_xml = NULL; xmlNode *child_xml = NULL; xmlNode *rsc_ops = NULL; xmlNode *template_ops = NULL; const char *template_ref = NULL; const char *clone = NULL; const char *id = NULL; if (xml_obj == NULL) { pe_err("No resource object for template unpacking"); return FALSE; } template_ref = crm_element_value(xml_obj, XML_CIB_TAG_RSC_TEMPLATE); if (template_ref == NULL) { return TRUE; } id = ID(xml_obj); if (id == NULL) { pe_err("'%s' object must have a id", crm_element_name(xml_obj)); return FALSE; } if (crm_str_eq(template_ref, id, TRUE)) { pe_err("The resource object '%s' should not reference itself", id); return FALSE; } cib_resources = get_xpath_object("//"XML_CIB_TAG_RESOURCES, data_set->input, LOG_TRACE); if (cib_resources == NULL) { pe_err("No resources configured"); return FALSE; } template = find_entity(cib_resources, XML_CIB_TAG_RSC_TEMPLATE, template_ref); if (template == NULL) { pe_err("No template named '%s'", template_ref); return FALSE; } new_xml = copy_xml(template); xmlNodeSetName(new_xml, xml_obj->name); crm_xml_replace(new_xml, XML_ATTR_ID, id); clone = crm_element_value(xml_obj, XML_RSC_ATTR_INCARNATION); if(clone) { crm_xml_add(new_xml, XML_RSC_ATTR_INCARNATION, clone); } template_ops = find_xml_node(new_xml, "operations", FALSE); for (child_xml = __xml_first_child_element(xml_obj); child_xml != NULL; child_xml = __xml_next_element(child_xml)) { xmlNode *new_child = NULL; new_child = add_node_copy(new_xml, child_xml); if (crm_str_eq((const char *)new_child->name, "operations", TRUE)) { rsc_ops = new_child; } } if (template_ops && rsc_ops) { xmlNode *op = NULL; GHashTable *rsc_ops_hash = g_hash_table_new_full(crm_str_hash, g_str_equal, free, NULL); for (op = __xml_first_child_element(rsc_ops); op != NULL; op = __xml_next_element(op)) { char *key = template_op_key(op); g_hash_table_insert(rsc_ops_hash, key, op); } for (op = __xml_first_child_element(template_ops); op != NULL; op = __xml_next_element(op)) { char *key = template_op_key(op); if (g_hash_table_lookup(rsc_ops_hash, key) == NULL) { add_node_copy(rsc_ops, op); } free(key); } if (rsc_ops_hash) { g_hash_table_destroy(rsc_ops_hash); } free_xml(template_ops); } /*free_xml(*expanded_xml); */ *expanded_xml = new_xml; /* Disable multi-level templates for now */ /*if(unpack_template(new_xml, expanded_xml, data_set) == FALSE) { free_xml(*expanded_xml); *expanded_xml = NULL; return FALSE; } */ return TRUE; } static gboolean add_template_rsc(xmlNode * xml_obj, pe_working_set_t * data_set) { const char *template_ref = NULL; const char *id = NULL; if (xml_obj == NULL) { pe_err("No resource object for processing resource list of template"); return FALSE; } template_ref = crm_element_value(xml_obj, XML_CIB_TAG_RSC_TEMPLATE); if (template_ref == NULL) { return TRUE; } id = ID(xml_obj); if (id == NULL) { pe_err("'%s' object must have a id", crm_element_name(xml_obj)); return FALSE; } if (crm_str_eq(template_ref, id, TRUE)) { pe_err("The resource object '%s' should not reference itself", id); return FALSE; } if (add_tag_ref(data_set->template_rsc_sets, template_ref, id) == FALSE) { return FALSE; } return TRUE; } static bool detect_promotable(resource_t *rsc) { const char *promotable = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_PROMOTABLE); if (crm_is_true(promotable)) { return TRUE; } // @COMPAT deprecated since 2.0.0 if (safe_str_eq(crm_element_name(rsc->xml), XML_CIB_TAG_MASTER)) { /* @TODO in some future version, pe_warn_once() here, * then drop support in even later version */ g_hash_table_insert(rsc->meta, strdup(XML_RSC_ATTR_PROMOTABLE), strdup(XML_BOOLEAN_TRUE)); return TRUE; } return FALSE; } gboolean common_unpack(xmlNode * xml_obj, resource_t ** rsc, resource_t * parent, pe_working_set_t * data_set) { bool isdefault = FALSE; xmlNode *expanded_xml = NULL; xmlNode *ops = NULL; const char *value = NULL; const char *rclass = NULL; /* Look for this after any templates have been expanded */ const char *id = crm_element_value(xml_obj, XML_ATTR_ID); bool guest_node = FALSE; bool remote_node = FALSE; bool has_versioned_params = FALSE; crm_log_xml_trace(xml_obj, "Processing resource input..."); if (id == NULL) { pe_err("Must specify id tag in "); return FALSE; } else if (rsc == NULL) { pe_err("Nowhere to unpack resource into"); return FALSE; } if (unpack_template(xml_obj, &expanded_xml, data_set) == FALSE) { return FALSE; } *rsc = calloc(1, sizeof(resource_t)); (*rsc)->cluster = data_set; if (expanded_xml) { crm_log_xml_trace(expanded_xml, "Expanded resource..."); (*rsc)->xml = expanded_xml; (*rsc)->orig_xml = xml_obj; } else { (*rsc)->xml = xml_obj; (*rsc)->orig_xml = NULL; } /* Do not use xml_obj from here on, use (*rsc)->xml in case templates are involved */ rclass = crm_element_value((*rsc)->xml, XML_AGENT_ATTR_CLASS); (*rsc)->parent = parent; ops = find_xml_node((*rsc)->xml, "operations", FALSE); (*rsc)->ops_xml = expand_idref(ops, data_set->input); (*rsc)->variant = get_resource_type(crm_element_name((*rsc)->xml)); if ((*rsc)->variant == pe_unknown) { pe_err("Unknown resource type: %s", crm_element_name((*rsc)->xml)); free(*rsc); return FALSE; } (*rsc)->parameters = crm_str_table_new(); #if ENABLE_VERSIONED_ATTRS (*rsc)->versioned_parameters = create_xml_node(NULL, XML_TAG_RSC_VER_ATTRS); #endif (*rsc)->meta = crm_str_table_new(); (*rsc)->allowed_nodes = g_hash_table_new_full(crm_str_hash, g_str_equal, NULL, free); (*rsc)->known_on = g_hash_table_new_full(crm_str_hash, g_str_equal, NULL, free); value = crm_element_value((*rsc)->xml, XML_RSC_ATTR_INCARNATION); if (value) { (*rsc)->id = crm_concat(id, value, ':'); add_hash_param((*rsc)->meta, XML_RSC_ATTR_INCARNATION, value); } else { (*rsc)->id = strdup(id); } (*rsc)->fns = &resource_class_functions[(*rsc)->variant]; pe_rsc_trace((*rsc), "Unpacking resource..."); get_meta_attributes((*rsc)->meta, *rsc, NULL, data_set); get_rsc_attributes((*rsc)->parameters, *rsc, NULL, data_set); #if ENABLE_VERSIONED_ATTRS pe_get_versioned_attributes((*rsc)->versioned_parameters, *rsc, NULL, data_set); #endif (*rsc)->flags = 0; set_bit((*rsc)->flags, pe_rsc_runnable); set_bit((*rsc)->flags, pe_rsc_provisional); if (is_not_set(data_set->flags, pe_flag_maintenance_mode)) { set_bit((*rsc)->flags, pe_rsc_managed); } (*rsc)->rsc_cons = NULL; (*rsc)->rsc_tickets = NULL; (*rsc)->actions = NULL; (*rsc)->role = RSC_ROLE_STOPPED; (*rsc)->next_role = RSC_ROLE_UNKNOWN; (*rsc)->recovery_type = recovery_stop_start; (*rsc)->stickiness = 0; (*rsc)->migration_threshold = INFINITY; (*rsc)->failure_timeout = 0; value = g_hash_table_lookup((*rsc)->meta, XML_CIB_ATTR_PRIORITY); (*rsc)->priority = crm_parse_int(value, "0"); value = g_hash_table_lookup((*rsc)->meta, XML_RSC_ATTR_NOTIFY); if (crm_is_true(value)) { set_bit((*rsc)->flags, pe_rsc_notify); } if (xml_contains_remote_node((*rsc)->xml)) { (*rsc)->is_remote_node = TRUE; if (g_hash_table_lookup((*rsc)->meta, XML_RSC_ATTR_CONTAINER)) { guest_node = TRUE; } else { remote_node = TRUE; } } value = g_hash_table_lookup((*rsc)->meta, XML_OP_ATTR_ALLOW_MIGRATE); #if ENABLE_VERSIONED_ATTRS has_versioned_params = xml_has_children((*rsc)->versioned_parameters); #endif if (crm_is_true(value) && has_versioned_params) { pe_rsc_trace((*rsc), "Migration is disabled for resources with versioned parameters"); } else if (crm_is_true(value)) { set_bit((*rsc)->flags, pe_rsc_allow_migrate); } else if ((value == NULL) && remote_node && !has_versioned_params) { /* By default, we want remote nodes to be able * to float around the cluster without having to stop all the * resources within the remote-node before moving. Allowing * migration support enables this feature. If this ever causes * problems, migration support can be explicitly turned off with * allow-migrate=false. * We don't support migration for versioned resources, though. */ set_bit((*rsc)->flags, pe_rsc_allow_migrate); } value = g_hash_table_lookup((*rsc)->meta, XML_RSC_ATTR_MANAGED); if (value != NULL && safe_str_neq("default", value)) { gboolean bool_value = TRUE; crm_str_to_boolean(value, &bool_value); if (bool_value == FALSE) { clear_bit((*rsc)->flags, pe_rsc_managed); } else { set_bit((*rsc)->flags, pe_rsc_managed); } } value = g_hash_table_lookup((*rsc)->meta, XML_RSC_ATTR_MAINTENANCE); if (value != NULL && safe_str_neq("default", value)) { gboolean bool_value = FALSE; crm_str_to_boolean(value, &bool_value); if (bool_value == TRUE) { clear_bit((*rsc)->flags, pe_rsc_managed); set_bit((*rsc)->flags, pe_rsc_maintenance); } } else if (is_set(data_set->flags, pe_flag_maintenance_mode)) { clear_bit((*rsc)->flags, pe_rsc_managed); set_bit((*rsc)->flags, pe_rsc_maintenance); } if (pe_rsc_is_clone(uber_parent(*rsc))) { value = g_hash_table_lookup((*rsc)->meta, XML_RSC_ATTR_UNIQUE); if (crm_is_true(value)) { set_bit((*rsc)->flags, pe_rsc_unique); } if (detect_promotable(*rsc)) { set_bit((*rsc)->flags, pe_rsc_promotable); } } else { set_bit((*rsc)->flags, pe_rsc_unique); } pe_rsc_trace((*rsc), "Options for %s", (*rsc)->id); value = g_hash_table_lookup((*rsc)->meta, XML_RSC_ATTR_RESTART); if (safe_str_eq(value, "restart")) { (*rsc)->restart_type = pe_restart_restart; pe_rsc_trace((*rsc), "\tDependency restart handling: restart"); pe_warn_once(pe_wo_restart_type, "Support for restart-type is deprecated and will be removed in a future release"); } else { (*rsc)->restart_type = pe_restart_ignore; pe_rsc_trace((*rsc), "\tDependency restart handling: ignore"); } value = g_hash_table_lookup((*rsc)->meta, XML_RSC_ATTR_MULTIPLE); if (safe_str_eq(value, "stop_only")) { (*rsc)->recovery_type = recovery_stop_only; pe_rsc_trace((*rsc), "\tMultiple running resource recovery: stop only"); } else if (safe_str_eq(value, "block")) { (*rsc)->recovery_type = recovery_block; pe_rsc_trace((*rsc), "\tMultiple running resource recovery: block"); } else { (*rsc)->recovery_type = recovery_stop_start; pe_rsc_trace((*rsc), "\tMultiple running resource recovery: stop/start"); } value = g_hash_table_lookup((*rsc)->meta, XML_RSC_ATTR_STICKINESS); if (value != NULL && safe_str_neq("default", value)) { (*rsc)->stickiness = char2score(value); } value = g_hash_table_lookup((*rsc)->meta, XML_RSC_ATTR_FAIL_STICKINESS); if (value != NULL && safe_str_neq("default", value)) { (*rsc)->migration_threshold = char2score(value); if ((*rsc)->migration_threshold < 0) { /* @TODO We use 1 here to preserve previous behavior, but this * should probably use the default (INFINITY) or 0 (to disable) * instead. */ pe_warn_once(pe_wo_neg_threshold, XML_RSC_ATTR_FAIL_STICKINESS " must be non-negative, using 1 instead"); (*rsc)->migration_threshold = 1; } } if (safe_str_eq(rclass, PCMK_RESOURCE_CLASS_STONITH)) { set_bit(data_set->flags, pe_flag_have_stonith_resource); set_bit((*rsc)->flags, pe_rsc_fence_device); } value = g_hash_table_lookup((*rsc)->meta, XML_RSC_ATTR_REQUIRES); handle_requires_pref: if (safe_str_eq(value, "nothing")) { } else if (safe_str_eq(value, "quorum")) { set_bit((*rsc)->flags, pe_rsc_needs_quorum); } else if (safe_str_eq(value, "unfencing")) { if (is_set((*rsc)->flags, pe_rsc_fence_device)) { crm_config_warn("%s is a fencing device but requires (un)fencing", (*rsc)->id); value = "quorum"; isdefault = TRUE; goto handle_requires_pref; } else if (is_not_set(data_set->flags, pe_flag_stonith_enabled)) { crm_config_warn("%s requires (un)fencing but fencing is disabled", (*rsc)->id); value = "quorum"; isdefault = TRUE; goto handle_requires_pref; } else { set_bit((*rsc)->flags, pe_rsc_needs_fencing); set_bit((*rsc)->flags, pe_rsc_needs_unfencing); } } else if (safe_str_eq(value, "fencing")) { set_bit((*rsc)->flags, pe_rsc_needs_fencing); if (is_not_set(data_set->flags, pe_flag_stonith_enabled)) { crm_config_warn("%s requires fencing but fencing is disabled", (*rsc)->id); } } else { if (value) { crm_config_err("Invalid value for %s->requires: %s%s", (*rsc)->id, value, is_set(data_set->flags, pe_flag_stonith_enabled) ? "" : " (stonith-enabled=false)"); } isdefault = TRUE; if(is_set((*rsc)->flags, pe_rsc_fence_device)) { value = "quorum"; } else if (((*rsc)->variant == pe_native) && safe_str_eq(crm_element_value((*rsc)->xml, XML_AGENT_ATTR_CLASS), PCMK_RESOURCE_CLASS_OCF) && safe_str_eq(crm_element_value((*rsc)->xml, XML_AGENT_ATTR_PROVIDER), "pacemaker") && safe_str_eq(crm_element_value((*rsc)->xml, XML_ATTR_TYPE), "remote") ) { value = "quorum"; } else if (is_set(data_set->flags, pe_flag_enable_unfencing)) { value = "unfencing"; } else if (is_set(data_set->flags, pe_flag_stonith_enabled)) { value = "fencing"; } else if (data_set->no_quorum_policy == no_quorum_ignore) { value = "nothing"; } else { value = "quorum"; } goto handle_requires_pref; } pe_rsc_trace((*rsc), "\tRequired to start: %s%s", value, isdefault?" (default)":""); value = g_hash_table_lookup((*rsc)->meta, XML_RSC_ATTR_FAIL_TIMEOUT); if (value != NULL) { // Stored as seconds (*rsc)->failure_timeout = (int) (crm_parse_interval_spec(value) / 1000); } if (remote_node) { value = g_hash_table_lookup((*rsc)->parameters, XML_REMOTE_ATTR_RECONNECT_INTERVAL); if (value) { /* reconnect delay works by setting failure_timeout and preventing the * connection from starting until the failure is cleared. */ (*rsc)->remote_reconnect_ms = crm_parse_interval_spec(value); /* we want to override any default failure_timeout in use when remote * reconnect_interval is in use. */ (*rsc)->failure_timeout = (*rsc)->remote_reconnect_ms / 1000; } } get_target_role(*rsc, &((*rsc)->next_role)); pe_rsc_trace((*rsc), "\tDesired next state: %s", (*rsc)->next_role != RSC_ROLE_UNKNOWN ? role2text((*rsc)->next_role) : "default"); if ((*rsc)->fns->unpack(*rsc, data_set) == FALSE) { return FALSE; } if (is_set(data_set->flags, pe_flag_symmetric_cluster)) { // This tag must stay exactly the same because it is tested elsewhere resource_location(*rsc, NULL, 0, "symmetric_default", data_set); } else if (guest_node) { /* remote resources tied to a container resource must always be allowed * to opt-in to the cluster. Whether the connection resource is actually * allowed to be placed on a node is dependent on the container resource */ resource_location(*rsc, NULL, 0, "remote_connection_default", data_set); } pe_rsc_trace((*rsc), "\tAction notification: %s", is_set((*rsc)->flags, pe_rsc_notify) ? "required" : "not required"); (*rsc)->utilization = crm_str_table_new(); pe__unpack_dataset_nvpairs((*rsc)->xml, XML_TAG_UTILIZATION, NULL, (*rsc)->utilization, NULL, FALSE, data_set); /* data_set->resources = g_list_append(data_set->resources, (*rsc)); */ if (expanded_xml) { if (add_template_rsc(xml_obj, data_set) == FALSE) { return FALSE; } } return TRUE; } void common_update_score(resource_t * rsc, const char *id, int score) { node_t *node = NULL; node = pe_hash_table_lookup(rsc->allowed_nodes, id); if (node != NULL) { pe_rsc_trace(rsc, "Updating score for %s on %s: %d + %d", rsc->id, id, node->weight, score); node->weight = merge_weights(node->weight, score); } if (rsc->children) { GListPtr gIter = rsc->children; for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; common_update_score(child_rsc, id, score); } } } gboolean is_parent(resource_t *child, resource_t *rsc) { resource_t *parent = child; if (parent == NULL || rsc == NULL) { return FALSE; } while (parent->parent != NULL) { if (parent->parent == rsc) { return TRUE; } parent = parent->parent; } return FALSE; } resource_t * uber_parent(resource_t * rsc) { resource_t *parent = rsc; if (parent == NULL) { return NULL; } while (parent->parent != NULL && parent->parent->variant != pe_container) { parent = parent->parent; } return parent; } void common_free(resource_t * rsc) { if (rsc == NULL) { return; } pe_rsc_trace(rsc, "Freeing %s %d", rsc->id, rsc->variant); g_list_free(rsc->rsc_cons); g_list_free(rsc->rsc_cons_lhs); g_list_free(rsc->rsc_tickets); g_list_free(rsc->dangling_migrations); if (rsc->parameters != NULL) { g_hash_table_destroy(rsc->parameters); } #if ENABLE_VERSIONED_ATTRS if (rsc->versioned_parameters != NULL) { free_xml(rsc->versioned_parameters); } #endif if (rsc->meta != NULL) { g_hash_table_destroy(rsc->meta); } if (rsc->utilization != NULL) { g_hash_table_destroy(rsc->utilization); } if (rsc->parent == NULL && is_set(rsc->flags, pe_rsc_orphan)) { free_xml(rsc->xml); rsc->xml = NULL; free_xml(rsc->orig_xml); rsc->orig_xml = NULL; /* if rsc->orig_xml, then rsc->xml is an expanded xml from a template */ } else if (rsc->orig_xml) { free_xml(rsc->xml); rsc->xml = NULL; } if (rsc->running_on) { g_list_free(rsc->running_on); rsc->running_on = NULL; } if (rsc->known_on) { g_hash_table_destroy(rsc->known_on); rsc->known_on = NULL; } if (rsc->actions) { g_list_free(rsc->actions); rsc->actions = NULL; } if (rsc->allowed_nodes) { g_hash_table_destroy(rsc->allowed_nodes); rsc->allowed_nodes = NULL; } g_list_free(rsc->fillers); g_list_free(rsc->rsc_location); pe_rsc_trace(rsc, "Resource freed"); free(rsc->id); free(rsc->clone_name); free(rsc->allocated_to); free(rsc->variant_opaque); free(rsc->pending_task); free(rsc); } /*! * \brief * \internal Find a node (and optionally count all) where resource is active * * \param[in] rsc Resource to check * \param[out] count_all If not NULL, will be set to count of active nodes * \param[out] count_clean If not NULL, will be set to count of clean nodes * * \return An active node (or NULL if resource is not active anywhere) * * \note The order of preference is: an active node that is the resource's * partial migration source; if the resource's "requires" is "quorum" or * "nothing", the first active node in the list that is clean and online; * the first active node in the list. */ pe_node_t * pe__find_active_on(const pe_resource_t *rsc, unsigned int *count_all, unsigned int *count_clean) { pe_node_t *active = NULL; pe_node_t *node = NULL; bool keep_looking = FALSE; bool is_happy = FALSE; if (count_all) { *count_all = 0; } if (count_clean) { *count_clean = 0; } if (rsc == NULL) { return NULL; } for (GList *node_iter = rsc->running_on; node_iter != NULL; node_iter = node_iter->next) { node = node_iter->data; keep_looking = FALSE; is_happy = node->details->online && !node->details->unclean; if (count_all) { ++*count_all; } if (count_clean && is_happy) { ++*count_clean; } if (count_all || count_clean) { // If we're counting, we need to go through entire list keep_looking = TRUE; } if (rsc->partial_migration_source != NULL) { if (node->details == rsc->partial_migration_source->details) { // This is the migration source active = node; } else { keep_looking = TRUE; } } else if (is_not_set(rsc->flags, pe_rsc_needs_fencing)) { if (is_happy && (!active || !active->details->online || active->details->unclean)) { // This is the first clean node active = node; } else { keep_looking = TRUE; } } if (active == NULL) { // This is first node in list active = node; } if (keep_looking == FALSE) { // Don't waste time iterating if we don't have to break; } } return active; } /*! * \brief * \internal Find and count active nodes according to "requires" * * \param[in] rsc Resource to check * \param[out] count If not NULL, will be set to count of active nodes * * \return An active node (or NULL if resource is not active anywhere) * * \note This is a convenience wrapper for pe__find_active_on() where the count * of all active nodes or only clean active nodes is desired according to * the "requires" meta-attribute. */ pe_node_t * pe__find_active_requires(const pe_resource_t *rsc, unsigned int *count) { if (rsc && is_not_set(rsc->flags, pe_rsc_needs_fencing)) { return pe__find_active_on(rsc, NULL, count); } return pe__find_active_on(rsc, count, NULL); } + +void +pe__count_common(pe_resource_t *rsc) +{ + if (rsc->children != NULL) { + for (GList *item = rsc->children; item != NULL; item = item->next) { + ((pe_resource_t *) item->data)->fns->count(item->data); + } + + } else if (is_not_set(rsc->flags, pe_rsc_orphan) + || (rsc->role > RSC_ROLE_STOPPED)) { + rsc->cluster->ninstances++; + if (pe__resource_is_disabled(rsc)) { + rsc->cluster->disabled_resources++; + } + if (is_set(rsc->flags, pe_rsc_block)) { + rsc->cluster->blocked_resources++; + } + } +} diff --git a/lib/pengine/group.c b/lib/pengine/group.c index f3297a0172..2a8f220916 100644 --- a/lib/pengine/group.c +++ b/lib/pengine/group.c @@ -1,294 +1,294 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU Lesser General Public License * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #define VARIANT_GROUP 1 #include "./variant.h" gboolean group_unpack(resource_t * rsc, pe_working_set_t * data_set) { xmlNode *xml_obj = rsc->xml; xmlNode *xml_native_rsc = NULL; group_variant_data_t *group_data = NULL; const char *group_ordered = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_ORDERED); const char *group_colocated = g_hash_table_lookup(rsc->meta, "collocated"); const char *clone_id = NULL; pe_rsc_trace(rsc, "Processing resource %s...", rsc->id); group_data = calloc(1, sizeof(group_variant_data_t)); group_data->num_children = 0; group_data->first_child = NULL; group_data->last_child = NULL; rsc->variant_opaque = group_data; group_data->ordered = TRUE; group_data->colocated = TRUE; if (group_ordered != NULL) { crm_str_to_boolean(group_ordered, &(group_data->ordered)); } if (group_colocated != NULL) { crm_str_to_boolean(group_colocated, &(group_data->colocated)); } clone_id = crm_element_value(rsc->xml, XML_RSC_ATTR_INCARNATION); for (xml_native_rsc = __xml_first_child_element(xml_obj); xml_native_rsc != NULL; xml_native_rsc = __xml_next_element(xml_native_rsc)) { if (crm_str_eq((const char *)xml_native_rsc->name, XML_CIB_TAG_RESOURCE, TRUE)) { resource_t *new_rsc = NULL; crm_xml_add(xml_native_rsc, XML_RSC_ATTR_INCARNATION, clone_id); if (common_unpack(xml_native_rsc, &new_rsc, rsc, data_set) == FALSE) { pe_err("Failed unpacking resource %s", crm_element_value(xml_obj, XML_ATTR_ID)); if (new_rsc != NULL && new_rsc->fns != NULL) { new_rsc->fns->free(new_rsc); } } group_data->num_children++; rsc->children = g_list_append(rsc->children, new_rsc); if (group_data->first_child == NULL) { group_data->first_child = new_rsc; } group_data->last_child = new_rsc; - print_resource(LOG_TRACE, "Added ", new_rsc, FALSE); + pe_rsc_trace(rsc, "Added %s member %s", rsc->id, new_rsc->id); } } if (group_data->num_children == 0) { #if 0 /* Bug #1287 */ crm_config_err("Group %s did not have any children", rsc->id); return FALSE; #else crm_config_warn("Group %s did not have any children", rsc->id); return TRUE; #endif } pe_rsc_trace(rsc, "Added %d children to resource %s...", group_data->num_children, rsc->id); return TRUE; } gboolean group_active(resource_t * rsc, gboolean all) { gboolean c_all = TRUE; gboolean c_any = FALSE; GListPtr gIter = rsc->children; for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; if (child_rsc->fns->active(child_rsc, all)) { c_any = TRUE; } else { c_all = FALSE; } } if (c_any == FALSE) { return FALSE; } else if (all && c_all == FALSE) { return FALSE; } return TRUE; } static void group_print_xml(resource_t * rsc, const char *pre_text, long options, void *print_data) { GListPtr gIter = rsc->children; char *child_text = crm_concat(pre_text, " ", ' '); status_print("%sid); status_print("number_resources=\"%d\" ", g_list_length(rsc->children)); status_print(">\n"); for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; child_rsc->fns->print(child_rsc, child_text, options, print_data); } status_print("%s\n", pre_text); free(child_text); } void group_print(resource_t * rsc, const char *pre_text, long options, void *print_data) { char *child_text = NULL; GListPtr gIter = rsc->children; if (pre_text == NULL) { pre_text = " "; } if (options & pe_print_xml) { group_print_xml(rsc, pre_text, options, print_data); return; } child_text = crm_concat(pre_text, " ", ' '); status_print("%sResource Group: %s", pre_text ? pre_text : "", rsc->id); if (options & pe_print_html) { status_print("\n
      \n"); } else if ((options & pe_print_log) == 0) { status_print("\n"); } if (options & pe_print_brief) { print_rscs_brief(rsc->children, child_text, options, print_data, TRUE); } else { for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; if (options & pe_print_html) { status_print("
    • \n"); } child_rsc->fns->print(child_rsc, child_text, options, print_data); if (options & pe_print_html) { status_print("
      • \n"); } } } if (options & pe_print_html) { status_print("
    \n"); } free(child_text); } int pe__group_xml(pcmk__output_t *out, va_list args) { long options = va_arg(args, long); resource_t *rsc = va_arg(args, resource_t *); GListPtr gIter = rsc->children; char *count = crm_itoa(g_list_length(gIter)); int rc = pe__name_and_nvpairs_xml(out, true, "group", 2 , "id", rsc->id , "number_resources", count); free(count); CRM_ASSERT(rc == 0); for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; out->message(out, crm_map_element_name(child_rsc->xml), options, child_rsc); } pcmk__output_xml_pop_parent(out); return rc; } int pe__group_html(pcmk__output_t *out, va_list args) { long options = va_arg(args, long); resource_t *rsc = va_arg(args, resource_t *); out->begin_list(out, NULL, NULL, "Resource Group: %s", rsc->id); if (options & pe_print_brief) { pe__rscs_brief_output(out, rsc->children, options, TRUE); } else { for (GListPtr gIter = rsc->children; gIter; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; out->message(out, crm_map_element_name(child_rsc->xml), options, child_rsc); } } out->end_list(out); return 0; } int pe__group_text(pcmk__output_t *out, va_list args) { long options = va_arg(args, long); resource_t *rsc = va_arg(args, resource_t *); out->begin_list(out, NULL, NULL, "Resource Group: %s", rsc->id); if (options & pe_print_brief) { pe__rscs_brief_output(out, rsc->children, options, TRUE); } else { for (GListPtr gIter = rsc->children; gIter; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; out->message(out, crm_map_element_name(child_rsc->xml), options, child_rsc); } } out->end_list(out); return 0; } void group_free(resource_t * rsc) { CRM_CHECK(rsc != NULL, return); pe_rsc_trace(rsc, "Freeing %s", rsc->id); for (GListPtr gIter = rsc->children; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; CRM_ASSERT(child_rsc); pe_rsc_trace(child_rsc, "Freeing child %s", child_rsc->id); child_rsc->fns->free(child_rsc); } pe_rsc_trace(rsc, "Freeing child list"); g_list_free(rsc->children); common_free(rsc); } enum rsc_role_e group_resource_state(const resource_t * rsc, gboolean current) { enum rsc_role_e group_role = RSC_ROLE_UNKNOWN; GListPtr gIter = rsc->children; for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; enum rsc_role_e role = child_rsc->fns->state(child_rsc, current); if (role > group_role) { group_role = role; } } pe_rsc_trace(rsc, "%s role: %s", rsc->id, role2text(group_role)); return group_role; } diff --git a/lib/pengine/native.c b/lib/pengine/native.c index 43bf5c791d..2f4576863d 100644 --- a/lib/pengine/native.c +++ b/lib/pengine/native.c @@ -1,1478 +1,1472 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU Lesser General Public License * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #define VARIANT_NATIVE 1 #include "./variant.h" /*! * \internal * \brief Check whether a resource is active on multiple nodes */ static bool is_multiply_active(pe_resource_t *rsc) { unsigned int count = 0; if (rsc->variant == pe_native) { pe__find_active_requires(rsc, &count); } return count > 1; } void native_add_running(resource_t * rsc, node_t * node, pe_working_set_t * data_set) { GListPtr gIter = rsc->running_on; CRM_CHECK(node != NULL, return); for (; gIter != NULL; gIter = gIter->next) { node_t *a_node = (node_t *) gIter->data; CRM_CHECK(a_node != NULL, return); if (safe_str_eq(a_node->details->id, node->details->id)) { return; } } pe_rsc_trace(rsc, "Adding %s to %s %s", rsc->id, node->details->uname, is_set(rsc->flags, pe_rsc_managed)?"":"(unmanaged)"); rsc->running_on = g_list_append(rsc->running_on, node); if (rsc->variant == pe_native) { node->details->running_rsc = g_list_append(node->details->running_rsc, rsc); } if (rsc->variant == pe_native && node->details->maintenance) { clear_bit(rsc->flags, pe_rsc_managed); } if (is_not_set(rsc->flags, pe_rsc_managed)) { resource_t *p = rsc->parent; pe_rsc_info(rsc, "resource %s isn't managed", rsc->id); resource_location(rsc, node, INFINITY, "not_managed_default", data_set); while(p && node->details->online) { /* add without the additional location constraint */ p->running_on = g_list_append(p->running_on, node); p = p->parent; } return; } if (is_multiply_active(rsc)) { switch (rsc->recovery_type) { case recovery_stop_only: { GHashTableIter gIter; node_t *local_node = NULL; /* make sure it doesn't come up again */ if (rsc->allowed_nodes != NULL) { g_hash_table_destroy(rsc->allowed_nodes); } rsc->allowed_nodes = node_hash_from_list(data_set->nodes); g_hash_table_iter_init(&gIter, rsc->allowed_nodes); while (g_hash_table_iter_next(&gIter, NULL, (void **)&local_node)) { local_node->weight = -INFINITY; } } break; case recovery_stop_start: break; case recovery_block: clear_bit(rsc->flags, pe_rsc_managed); set_bit(rsc->flags, pe_rsc_block); /* If the resource belongs to a group or bundle configured with * multiple-active=block, block the entire entity. */ if (rsc->parent && (rsc->parent->variant == pe_group || rsc->parent->variant == pe_container) && rsc->parent->recovery_type == recovery_block) { GListPtr gIter = rsc->parent->children; for (; gIter != NULL; gIter = gIter->next) { resource_t *child = (resource_t *) gIter->data; clear_bit(child->flags, pe_rsc_managed); set_bit(child->flags, pe_rsc_block); } } break; } crm_debug("%s is active on multiple nodes including %s: %s", rsc->id, node->details->uname, recovery2text(rsc->recovery_type)); } else { pe_rsc_trace(rsc, "Resource %s is active on: %s", rsc->id, node->details->uname); } if (rsc->parent != NULL) { native_add_running(rsc->parent, node, data_set); } } static void recursive_clear_unique(pe_resource_t *rsc) { clear_bit(rsc->flags, pe_rsc_unique); add_hash_param(rsc->meta, XML_RSC_ATTR_UNIQUE, XML_BOOLEAN_FALSE); for (GList *child = rsc->children; child != NULL; child = child->next) { recursive_clear_unique((pe_resource_t *) child->data); } } gboolean native_unpack(resource_t * rsc, pe_working_set_t * data_set) { resource_t *parent = uber_parent(rsc); native_variant_data_t *native_data = NULL; const char *standard = crm_element_value(rsc->xml, XML_AGENT_ATTR_CLASS); uint32_t ra_caps = pcmk_get_ra_caps(standard); pe_rsc_trace(rsc, "Processing resource %s...", rsc->id); native_data = calloc(1, sizeof(native_variant_data_t)); rsc->variant_opaque = native_data; // Only some agent standards support unique and promotable clones if (is_not_set(ra_caps, pcmk_ra_cap_unique) && is_set(rsc->flags, pe_rsc_unique) && pe_rsc_is_clone(parent)) { /* @COMPAT We should probably reject this situation as an error (as we * do for promotable below) rather than warn and convert, but that would * be a backward-incompatible change that we should probably do with a * transform at a schema major version bump. */ pe__force_anon(standard, parent, rsc->id, data_set); /* Clear globally-unique on the parent and all its descendents unpacked * so far (clearing the parent should make any future children unpacking * correct). We have to clear this resource explicitly because it isn't * hooked into the parent's children yet. */ recursive_clear_unique(parent); recursive_clear_unique(rsc); } if (is_not_set(ra_caps, pcmk_ra_cap_promotable) && is_set(parent->flags, pe_rsc_promotable)) { pe_err("Resource %s is of type %s and therefore " "cannot be used as a promotable clone resource", rsc->id, standard); return FALSE; } return TRUE; } static bool rsc_is_on_node(resource_t *rsc, const node_t *node, int flags) { pe_rsc_trace(rsc, "Checking whether %s is on %s", rsc->id, node->details->uname); if (is_set(flags, pe_find_current) && rsc->running_on) { for (GListPtr iter = rsc->running_on; iter; iter = iter->next) { node_t *loc = (node_t *) iter->data; if (loc->details == node->details) { return TRUE; } } } else if (is_set(flags, pe_find_inactive) && (rsc->running_on == NULL)) { return TRUE; } else if (is_not_set(flags, pe_find_current) && rsc->allocated_to && (rsc->allocated_to->details == node->details)) { return TRUE; } return FALSE; } resource_t * native_find_rsc(resource_t * rsc, const char *id, const node_t *on_node, int flags) { bool match = FALSE; resource_t *result = NULL; CRM_CHECK(id && rsc && rsc->id, return NULL); if (flags & pe_find_clone) { const char *rid = ID(rsc->xml); if (!pe_rsc_is_clone(uber_parent(rsc))) { match = FALSE; } else if (!strcmp(id, rsc->id) || safe_str_eq(id, rid)) { match = TRUE; } } else if (!strcmp(id, rsc->id)) { match = TRUE; } else if (is_set(flags, pe_find_renamed) && rsc->clone_name && strcmp(rsc->clone_name, id) == 0) { match = TRUE; } else if (is_set(flags, pe_find_any) || (is_set(flags, pe_find_anon) && is_not_set(rsc->flags, pe_rsc_unique))) { match = pe_base_name_eq(rsc, id); } if (match && on_node) { bool match_node = rsc_is_on_node(rsc, on_node, flags); if (match_node == FALSE) { match = FALSE; } } if (match) { return rsc; } for (GListPtr gIter = rsc->children; gIter != NULL; gIter = gIter->next) { resource_t *child = (resource_t *) gIter->data; result = rsc->fns->find_rsc(child, id, on_node, flags); if (result) { return result; } } return NULL; } char * native_parameter(resource_t * rsc, node_t * node, gboolean create, const char *name, pe_working_set_t * data_set) { char *value_copy = NULL; const char *value = NULL; GHashTable *hash = NULL; GHashTable *local_hash = NULL; CRM_CHECK(rsc != NULL, return NULL); CRM_CHECK(name != NULL && strlen(name) != 0, return NULL); pe_rsc_trace(rsc, "Looking up %s in %s", name, rsc->id); if (create || g_hash_table_size(rsc->parameters) == 0) { if (node != NULL) { pe_rsc_trace(rsc, "Creating hash with node %s", node->details->uname); } else { pe_rsc_trace(rsc, "Creating default hash"); } local_hash = crm_str_table_new(); get_rsc_attributes(local_hash, rsc, node, data_set); hash = local_hash; } else { hash = rsc->parameters; } value = g_hash_table_lookup(hash, name); if (value == NULL) { /* try meta attributes instead */ value = g_hash_table_lookup(rsc->meta, name); } if (value != NULL) { value_copy = strdup(value); } if (local_hash != NULL) { g_hash_table_destroy(local_hash); } return value_copy; } gboolean native_active(resource_t * rsc, gboolean all) { GListPtr gIter = rsc->running_on; for (; gIter != NULL; gIter = gIter->next) { node_t *a_node = (node_t *) gIter->data; if (a_node->details->unclean) { crm_debug("Resource %s: node %s is unclean", rsc->id, a_node->details->uname); return TRUE; } else if (a_node->details->online == FALSE) { crm_debug("Resource %s: node %s is offline", rsc->id, a_node->details->uname); } else { crm_debug("Resource %s active on %s", rsc->id, a_node->details->uname); return TRUE; } } return FALSE; } struct print_data_s { long options; void *print_data; }; static void native_print_attr(gpointer key, gpointer value, gpointer user_data) { long options = ((struct print_data_s *)user_data)->options; void *print_data = ((struct print_data_s *)user_data)->print_data; status_print("Option: %s = %s\n", (char *)key, (char *)value); } static const char * native_pending_state(resource_t * rsc) { const char *pending_state = NULL; if (safe_str_eq(rsc->pending_task, CRMD_ACTION_START)) { pending_state = "Starting"; } else if (safe_str_eq(rsc->pending_task, CRMD_ACTION_STOP)) { pending_state = "Stopping"; } else if (safe_str_eq(rsc->pending_task, CRMD_ACTION_MIGRATE)) { pending_state = "Migrating"; } else if (safe_str_eq(rsc->pending_task, CRMD_ACTION_MIGRATED)) { /* Work might be done in here. */ pending_state = "Migrating"; } else if (safe_str_eq(rsc->pending_task, CRMD_ACTION_PROMOTE)) { pending_state = "Promoting"; } else if (safe_str_eq(rsc->pending_task, CRMD_ACTION_DEMOTE)) { pending_state = "Demoting"; } return pending_state; } static const char * native_pending_task(resource_t * rsc) { const char *pending_task = NULL; if (safe_str_eq(rsc->pending_task, CRMD_ACTION_STATUS)) { pending_task = "Monitoring"; /* Pending probes are not printed, even if pending * operations are requested. If someone ever requests that * behavior, uncomment this and the corresponding part of * unpack.c:unpack_rsc_op(). */ /* } else if (safe_str_eq(rsc->pending_task, "probe")) { pending_task = "Checking"; */ } return pending_task; } static enum rsc_role_e native_displayable_role(resource_t *rsc) { enum rsc_role_e role = rsc->role; if ((role == RSC_ROLE_STARTED) && is_set(uber_parent(rsc)->flags, pe_rsc_promotable)) { role = RSC_ROLE_SLAVE; } return role; } static const char * native_displayable_state(resource_t *rsc, long options) { const char *rsc_state = NULL; if (options & pe_print_pending) { rsc_state = native_pending_state(rsc); } if (rsc_state == NULL) { rsc_state = role2text(native_displayable_role(rsc)); } return rsc_state; } static void native_print_xml(resource_t * rsc, const char *pre_text, long options, void *print_data) { const char *class = crm_element_value(rsc->xml, XML_AGENT_ATTR_CLASS); const char *prov = crm_element_value(rsc->xml, XML_AGENT_ATTR_PROVIDER); const char *rsc_state = native_displayable_state(rsc, options); const char *target_role = NULL; /* resource information. */ status_print("%sxml, XML_ATTR_TYPE)); status_print("role=\"%s\" ", rsc_state); if (rsc->meta) { target_role = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_TARGET_ROLE); } if (target_role) { status_print("target_role=\"%s\" ", target_role); } status_print("active=\"%s\" ", rsc->fns->active(rsc, TRUE) ? "true" : "false"); status_print("orphaned=\"%s\" ", is_set(rsc->flags, pe_rsc_orphan) ? "true" : "false"); status_print("blocked=\"%s\" ", is_set(rsc->flags, pe_rsc_block) ? "true" : "false"); status_print("managed=\"%s\" ", is_set(rsc->flags, pe_rsc_managed) ? "true" : "false"); status_print("failed=\"%s\" ", is_set(rsc->flags, pe_rsc_failed) ? "true" : "false"); status_print("failure_ignored=\"%s\" ", is_set(rsc->flags, pe_rsc_failure_ignored) ? "true" : "false"); status_print("nodes_running_on=\"%d\" ", g_list_length(rsc->running_on)); if (options & pe_print_pending) { const char *pending_task = native_pending_task(rsc); if (pending_task) { status_print("pending=\"%s\" ", pending_task); } } if (options & pe_print_dev) { status_print("provisional=\"%s\" ", is_set(rsc->flags, pe_rsc_provisional) ? "true" : "false"); status_print("runnable=\"%s\" ", is_set(rsc->flags, pe_rsc_runnable) ? "true" : "false"); status_print("priority=\"%f\" ", (double)rsc->priority); status_print("variant=\"%s\" ", crm_element_name(rsc->xml)); } /* print out the nodes this resource is running on */ if (options & pe_print_rsconly) { status_print("/>\n"); /* do nothing */ } else if (rsc->running_on != NULL) { GListPtr gIter = rsc->running_on; status_print(">\n"); for (; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; status_print("%s \n", pre_text, node->details->uname, node->details->id, node->details->online ? "false" : "true"); } status_print("%s\n", pre_text); } else { status_print("/>\n"); } } /* making this inline rather than a macro prevents a coverity "unreachable" * warning on the first usage */ static inline const char * comma_if(int i) { return i? ", " : ""; } static char * flags_string(pe_resource_t *rsc, pe_node_t *node, long options, const char *target_role) { char *flags[6] = { NULL, }; char *result = NULL; int ndx = 0; if (node && node->details->online == FALSE && node->details->unclean) { flags[ndx++] = strdup("UNCLEAN"); } if (is_set(options, pe_print_pending)) { const char *pending_task = native_pending_task(rsc); if (pending_task) { flags[ndx++] = strdup(pending_task); } } if (target_role) { enum rsc_role_e target_role_e = text2role(target_role); /* Ignore target role Started, as it is the default anyways * (and would also allow a Master to be Master). * Show if target role limits our abilities. */ if (target_role_e == RSC_ROLE_STOPPED) { flags[ndx++] = strdup("disabled"); - rsc->cluster->disabled_resources++; } else if (is_set(uber_parent(rsc)->flags, pe_rsc_promotable) && target_role_e == RSC_ROLE_SLAVE) { flags[ndx++] = crm_strdup_printf("target-role:%s", target_role); - rsc->cluster->disabled_resources++; } } if (is_set(rsc->flags, pe_rsc_block)) { flags[ndx++] = strdup("blocked"); - rsc->cluster->blocked_resources++; } else if (is_not_set(rsc->flags, pe_rsc_managed)) { flags[ndx++] = strdup("unmanaged"); } if (is_set(rsc->flags, pe_rsc_failure_ignored)) { flags[ndx++] = strdup("failure ignored"); } if (ndx > 0) { char *total = g_strjoinv(" ", flags); result = crm_strdup_printf(" (%s)", total); g_free(total); } while (--ndx >= 0) { free(flags[ndx]); } return result; } static char * native_output_string(resource_t *rsc, const char *name, node_t *node, long options, const char *target_role) { const char *desc = NULL; const char *class = crm_element_value(rsc->xml, XML_AGENT_ATTR_CLASS); const char *kind = crm_element_value(rsc->xml, XML_ATTR_TYPE); enum rsc_role_e role = native_displayable_role(rsc); char *retval = NULL; char *unames = NULL; char *provider = NULL; const char *orphan = NULL; char *role_s = NULL; char *node_s = NULL; char *print_dev_s = NULL; char *flags_s = NULL; CRM_ASSERT(kind != NULL); if (is_set(pcmk_get_ra_caps(class), pcmk_ra_cap_provider)) { provider = crm_strdup_printf("::%s", crm_element_value(rsc->xml, XML_AGENT_ATTR_PROVIDER)); } if (is_set(rsc->flags, pe_rsc_orphan)) { orphan = " ORPHANED"; } if (role > RSC_ROLE_SLAVE && is_set(rsc->flags, pe_rsc_failed)) { role_s = crm_strdup_printf(" FAILED %s", role2text(role)); } else if (is_set(rsc->flags, pe_rsc_failed)) { role_s = crm_strdup_printf(" FAILED"); } else { role_s = crm_strdup_printf(" %s", native_displayable_state(rsc, options)); } if (node) { node_s = crm_strdup_printf(" %s", node->details->uname); } if (is_set(options, pe_print_rsconly) || g_list_length(rsc->running_on) > 1) { desc = crm_element_value(rsc->xml, XML_ATTR_DESC); } if (is_not_set(options, pe_print_rsconly) && g_list_length(rsc->running_on) > 1) { GListPtr gIter = rsc->running_on; gchar **arr = calloc(g_list_length(rsc->running_on)+1, sizeof(gchar *)); int i = 0; char *total = NULL; for (; gIter != NULL; gIter = gIter->next) { node_t *n = (node_t *) gIter->data; arr[i] = (gchar *) strdup(n->details->uname); i++; } total = g_strjoinv(" ", arr); unames = crm_strdup_printf(" [ %s ]", total); g_free(total); g_strfreev(arr); } if (is_set(options, pe_print_dev)) { print_dev_s = crm_strdup_printf(" (%s%svariant=%s, priority=%f)", is_set(rsc->flags, pe_rsc_provisional) ? "provisional, " : "", is_set(rsc->flags, pe_rsc_runnable) ? "" : "non-startable, ", crm_element_name(rsc->xml), (double)rsc->priority); } flags_s = flags_string(rsc, node, options, target_role); retval = crm_strdup_printf("%s\t(%s%s:%s):\t%s%s%s%s%s%s%s%s", name, class, provider ? provider : "", kind, orphan ? orphan : "", role_s, node_s ? node_s : "", print_dev_s ? print_dev_s : "", flags_s ? flags_s : "", desc ? " " : "", desc ? desc : "", unames ? unames : ""); free(provider); free(role_s); free(node_s); free(unames); free(print_dev_s); free(flags_s); return retval; } void pe__common_output_html(pcmk__output_t *out, resource_t * rsc, const char *name, node_t *node, long options) { char *s = NULL; const char *kind = crm_element_value(rsc->xml, XML_ATTR_TYPE); const char *target_role = NULL; xmlNodePtr list_node = NULL; const char *cl = NULL; CRM_ASSERT(rsc->variant == pe_native); CRM_ASSERT(kind != NULL); if (rsc->meta) { const char *is_internal = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_INTERNAL_RSC); if (crm_is_true(is_internal) && is_not_set(options, pe_print_implicit)) { crm_trace("skipping print of internal resource %s", rsc->id); return; } target_role = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_TARGET_ROLE); } if ((options & pe_print_rsconly) || g_list_length(rsc->running_on) > 1) { node = NULL; } if (is_not_set(rsc->flags, pe_rsc_managed)) { cl = "rsc-managed"; } else if (is_set(rsc->flags, pe_rsc_failed)) { cl = "rsc-failed"; } else if (rsc->variant == pe_native && (rsc->running_on == NULL)) { cl = "rsc-failed"; } else if (g_list_length(rsc->running_on) > 1) { cl = "rsc-multiple"; } else if (is_set(rsc->flags, pe_rsc_failure_ignored)) { cl = "rsc-failure-ignored"; } else { cl = "rsc-ok"; } s = native_output_string(rsc, name, node, options, target_role); list_node = pcmk__output_create_html_node(out, "li", NULL, NULL, NULL); pcmk_create_html_node(list_node, "span", NULL, cl, s); free(s); if (is_set(options, pe_print_details)) { GHashTableIter iter; gpointer key, value; out->begin_list(out, NULL, NULL, "Options"); g_hash_table_iter_init(&iter, rsc->parameters); while (g_hash_table_iter_next(&iter, &key, &value)) { out->list_item(out, NULL, "Option: %s = %s", (char *) key, (char *) value); } out->end_list(out); } if (is_set(options, pe_print_dev)) { GHashTableIter iter; node_t *n = NULL; out->begin_list(out, NULL, NULL, "Allowed Nodes"); g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&n)) { out->list_item(out, NULL, "%s %d", n->details->uname, n->weight); } out->end_list(out); } if (is_set(options, pe_print_max_details)) { GHashTableIter iter; node_t *n = NULL; out->begin_list(out, NULL, NULL, "=== Allowed Nodes"); g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&n)) { pe__output_node(n, FALSE, out); } out->end_list(out); } } void pe__common_output_text(pcmk__output_t *out, resource_t * rsc, const char *name, node_t *node, long options) { char *s = NULL; const char *target_role = NULL; CRM_ASSERT(rsc->variant == pe_native); if (rsc->meta) { const char *is_internal = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_INTERNAL_RSC); if (crm_is_true(is_internal) && is_not_set(options, pe_print_implicit)) { crm_trace("skipping print of internal resource %s", rsc->id); return; } target_role = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_TARGET_ROLE); } if (is_set(options, pe_print_rsconly) || g_list_length(rsc->running_on) > 1) { node = NULL; } s = native_output_string(rsc, name, node, options, target_role); out->list_item(out, NULL, "%s", s); free(s); if (is_set(options, pe_print_details)) { GHashTableIter iter; gpointer key, value; out->begin_list(out, NULL, NULL, "Options"); g_hash_table_iter_init(&iter, rsc->parameters); while (g_hash_table_iter_next(&iter, &key, &value)) { out->list_item(out, NULL, "Option: %s = %s", (char *) key, (char *) value); } out->end_list(out); } if (is_set(options, pe_print_dev)) { GHashTableIter iter; node_t *n = NULL; out->begin_list(out, NULL, NULL, "Allowed Nodes"); g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&n)) { out->list_item(out, NULL, "%s %d", n->details->uname, n->weight); } out->end_list(out); } if (is_set(options, pe_print_max_details)) { GHashTableIter iter; node_t *n = NULL; out->begin_list(out, NULL, NULL, "=== Allowed Nodes"); g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&n)) { pe__output_node(n, FALSE, out); } out->end_list(out); } } void common_print(resource_t * rsc, const char *pre_text, const char *name, node_t *node, long options, void *print_data) { const char *desc = NULL; const char *class = crm_element_value(rsc->xml, XML_AGENT_ATTR_CLASS); const char *kind = crm_element_value(rsc->xml, XML_ATTR_TYPE); const char *target_role = NULL; enum rsc_role_e role = native_displayable_role(rsc); int offset = 0; int flagOffset = 0; char buffer[LINE_MAX]; char flagBuffer[LINE_MAX]; CRM_ASSERT(rsc->variant == pe_native); CRM_ASSERT(kind != NULL); if (rsc->meta) { const char *is_internal = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_INTERNAL_RSC); if (crm_is_true(is_internal) && is_not_set(options, pe_print_implicit)) { crm_trace("skipping print of internal resource %s", rsc->id); return; } target_role = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_TARGET_ROLE); } if (pre_text == NULL && (options & pe_print_printf)) { pre_text = " "; } if (options & pe_print_xml) { native_print_xml(rsc, pre_text, options, print_data); return; } if ((options & pe_print_rsconly) || g_list_length(rsc->running_on) > 1) { node = NULL; } if (options & pe_print_html) { if (is_not_set(rsc->flags, pe_rsc_managed)) { status_print(""); } else if (is_set(rsc->flags, pe_rsc_failed)) { status_print(""); } else if (rsc->variant == pe_native && (rsc->running_on == NULL)) { status_print(""); } else if (g_list_length(rsc->running_on) > 1) { status_print(""); } else if (is_set(rsc->flags, pe_rsc_failure_ignored)) { status_print(""); } else { status_print(""); } } if(pre_text) { offset += snprintf(buffer + offset, LINE_MAX - offset, "%s", pre_text); } offset += snprintf(buffer + offset, LINE_MAX - offset, "%s", name); offset += snprintf(buffer + offset, LINE_MAX - offset, "\t(%s", class); if (is_set(pcmk_get_ra_caps(class), pcmk_ra_cap_provider)) { const char *prov = crm_element_value(rsc->xml, XML_AGENT_ATTR_PROVIDER); offset += snprintf(buffer + offset, LINE_MAX - offset, "::%s", prov); } offset += snprintf(buffer + offset, LINE_MAX - offset, ":%s):\t", kind); if(is_set(rsc->flags, pe_rsc_orphan)) { offset += snprintf(buffer + offset, LINE_MAX - offset, " ORPHANED "); } if(role > RSC_ROLE_SLAVE && is_set(rsc->flags, pe_rsc_failed)) { offset += snprintf(buffer + offset, LINE_MAX - offset, "FAILED %s", role2text(role)); } else if(is_set(rsc->flags, pe_rsc_failed)) { offset += snprintf(buffer + offset, LINE_MAX - offset, "FAILED"); } else { const char *rsc_state = native_displayable_state(rsc, options); offset += snprintf(buffer + offset, LINE_MAX - offset, "%s", rsc_state); } if(node) { offset += snprintf(buffer + offset, LINE_MAX - offset, " %s", node->details->uname); if (node->details->online == FALSE && node->details->unclean) { flagOffset += snprintf(flagBuffer + flagOffset, LINE_MAX - flagOffset, "%sUNCLEAN", comma_if(flagOffset)); } } if (options & pe_print_pending) { const char *pending_task = native_pending_task(rsc); if (pending_task) { flagOffset += snprintf(flagBuffer + flagOffset, LINE_MAX - flagOffset, "%s%s", comma_if(flagOffset), pending_task); } } if (target_role) { enum rsc_role_e target_role_e = text2role(target_role); /* Ignore target role Started, as it is the default anyways * (and would also allow a Master to be Master). * Show if target role limits our abilities. */ if (target_role_e == RSC_ROLE_STOPPED) { flagOffset += snprintf(flagBuffer + flagOffset, LINE_MAX - flagOffset, "%sdisabled", comma_if(flagOffset)); - rsc->cluster->disabled_resources++; } else if (is_set(uber_parent(rsc)->flags, pe_rsc_promotable) && target_role_e == RSC_ROLE_SLAVE) { flagOffset += snprintf(flagBuffer + flagOffset, LINE_MAX - flagOffset, "%starget-role:%s", comma_if(flagOffset), target_role); - rsc->cluster->disabled_resources++; } } if (is_set(rsc->flags, pe_rsc_block)) { flagOffset += snprintf(flagBuffer + flagOffset, LINE_MAX - flagOffset, "%sblocked", comma_if(flagOffset)); - rsc->cluster->blocked_resources++; } else if (is_not_set(rsc->flags, pe_rsc_managed)) { flagOffset += snprintf(flagBuffer + flagOffset, LINE_MAX - flagOffset, "%sunmanaged", comma_if(flagOffset)); } if(is_set(rsc->flags, pe_rsc_failure_ignored)) { flagOffset += snprintf(flagBuffer + flagOffset, LINE_MAX - flagOffset, "%sfailure ignored", comma_if(flagOffset)); } if ((options & pe_print_rsconly) || g_list_length(rsc->running_on) > 1) { desc = crm_element_value(rsc->xml, XML_ATTR_DESC); } CRM_LOG_ASSERT(offset > 0); if(flagOffset > 0) { status_print("%s (%s)%s%s", buffer, flagBuffer, desc?" ":"", desc?desc:""); } else { status_print("%s%s%s", buffer, desc?" ":"", desc?desc:""); } #if CURSES_ENABLED if ((options & pe_print_rsconly) || g_list_length(rsc->running_on) > 1) { /* Done */ } else if (options & pe_print_ncurses) { /* coverity[negative_returns] False positive */ move(-1, 0); } #endif if (options & pe_print_html) { status_print(" "); } if ((options & pe_print_rsconly)) { } else if (g_list_length(rsc->running_on) > 1) { GListPtr gIter = rsc->running_on; int counter = 0; if (options & pe_print_html) { status_print("
      \n"); } else if ((options & pe_print_printf) || (options & pe_print_ncurses)) { status_print("["); } for (; gIter != NULL; gIter = gIter->next) { node_t *n = (node_t *) gIter->data; counter++; if (options & pe_print_html) { status_print("
    • \n%s", n->details->uname); } else if ((options & pe_print_printf) || (options & pe_print_ncurses)) { status_print(" %s", n->details->uname); } else if ((options & pe_print_log)) { status_print("\t%d : %s", counter, n->details->uname); } else { status_print("%s", n->details->uname); } if (options & pe_print_html) { status_print("
      • \n"); } } if (options & pe_print_html) { status_print("
    \n"); } else if ((options & pe_print_printf) || (options & pe_print_ncurses)) { status_print(" ]"); } } if (options & pe_print_html) { status_print("
    \n"); } else if (options & pe_print_suppres_nl) { /* nothing */ } else if ((options & pe_print_printf) || (options & pe_print_ncurses)) { status_print("\n"); } if (options & pe_print_details) { struct print_data_s pdata; pdata.options = options; pdata.print_data = print_data; g_hash_table_foreach(rsc->parameters, native_print_attr, &pdata); } if (options & pe_print_dev) { GHashTableIter iter; node_t *n = NULL; status_print("%s\t(%s%svariant=%s, priority=%f)", pre_text, is_set(rsc->flags, pe_rsc_provisional) ? "provisional, " : "", is_set(rsc->flags, pe_rsc_runnable) ? "" : "non-startable, ", crm_element_name(rsc->xml), (double)rsc->priority); status_print("%s\tAllowed Nodes", pre_text); g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&n)) { status_print("%s\t * %s %d", pre_text, n->details->uname, n->weight); } } if (options & pe_print_max_details) { GHashTableIter iter; node_t *n = NULL; status_print("%s\t=== Allowed Nodes\n", pre_text); g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&n)) { print_node("\t", n, FALSE); } } } void native_print(resource_t * rsc, const char *pre_text, long options, void *print_data) { node_t *node = NULL; CRM_ASSERT(rsc->variant == pe_native); if (options & pe_print_xml) { native_print_xml(rsc, pre_text, options, print_data); return; } node = pe__current_node(rsc); if (node == NULL) { // This is set only if a non-probe action is pending on this node node = rsc->pending_node; } common_print(rsc, pre_text, rsc_printable_id(rsc), node, options, print_data); } int pe__resource_xml(pcmk__output_t *out, va_list args) { long options = va_arg(args, int); resource_t *rsc = va_arg(args, resource_t *); const char *class = crm_element_value(rsc->xml, XML_AGENT_ATTR_CLASS); const char *prov = crm_element_value(rsc->xml, XML_AGENT_ATTR_PROVIDER); const char *rsc_state = native_displayable_state(rsc, options); long is_print_pending = options & pe_print_pending; long is_print_dev = options & pe_print_dev; char ra_name[LINE_MAX]; char *nodes_running_on = NULL; char *priority = NULL; int rc = 0; CRM_ASSERT(rsc->variant == pe_native); /* resource information. */ sprintf(ra_name, "%s%s%s:%s", class, prov ? "::" : "", prov ? prov : "" , crm_element_value(rsc->xml, XML_ATTR_TYPE)); nodes_running_on = crm_itoa(g_list_length(rsc->running_on)); priority = crm_ftoa(rsc->priority); rc = pe__name_and_nvpairs_xml(out, true, "resource", 16 , "id", rsc_printable_id(rsc) , "resource_agent", ra_name , "role", rsc_state , "target_role", (rsc->meta ? g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_TARGET_ROLE) : NULL) , "active", BOOL2STR(rsc->fns->active(rsc, TRUE)) , "orphaned", BOOL2STR(is_set(rsc->flags, pe_rsc_orphan)) , "blocked", BOOL2STR(is_set(rsc->flags, pe_rsc_block)) , "managed", BOOL2STR(is_set(rsc->flags, pe_rsc_managed)) , "failed", BOOL2STR(is_set(rsc->flags, pe_rsc_failed)) , "failure_ignored", BOOL2STR(is_set(rsc->flags, pe_rsc_failure_ignored)) , "nodes_running_on", nodes_running_on , "pending", (is_print_pending ? native_pending_task(rsc) : NULL) , "provisional", (is_print_dev ? BOOL2STR(is_set(rsc->flags, pe_rsc_provisional)) : NULL) , "runnable", (is_print_dev ? BOOL2STR(is_set(rsc->flags, pe_rsc_runnable)) : NULL) , "priority", (is_print_dev ? priority : NULL) , "variant", (is_print_dev ? crm_element_name(rsc->xml) : NULL)); free(priority); free(nodes_running_on); CRM_ASSERT(rc == 0); if (rsc->running_on != NULL) { GListPtr gIter = rsc->running_on; for (; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; rc = pe__name_and_nvpairs_xml(out, false, "node", 3 , "name", node->details->uname , "id", node->details->id , "cached", BOOL2STR(node->details->online)); CRM_ASSERT(rc == 0); } } pcmk__output_xml_pop_parent(out); return rc; } int pe__resource_html(pcmk__output_t *out, va_list args) { long options = va_arg(args, int); resource_t *rsc = va_arg(args, resource_t *); node_t *node = pe__current_node(rsc); CRM_ASSERT(rsc->variant == pe_native); if (node == NULL) { // This is set only if a non-probe action is pending on this node node = rsc->pending_node; } pe__common_output_html(out, rsc, rsc_printable_id(rsc), node, options); return 0; } int pe__resource_text(pcmk__output_t *out, va_list args) { long options = va_arg(args, int); resource_t *rsc = va_arg(args, resource_t *); node_t *node = pe__current_node(rsc); CRM_ASSERT(rsc->variant == pe_native); if (node == NULL) { // This is set only if a non-probe action is pending on this node node = rsc->pending_node; } pe__common_output_text(out, rsc, rsc_printable_id(rsc), node, options); return 0; } void native_free(resource_t * rsc) { pe_rsc_trace(rsc, "Freeing resource action list (not the data)"); common_free(rsc); } enum rsc_role_e native_resource_state(const resource_t * rsc, gboolean current) { enum rsc_role_e role = rsc->next_role; if (current) { role = rsc->role; } pe_rsc_trace(rsc, "%s state: %s", rsc->id, role2text(role)); return role; } /*! * \internal * \brief List nodes where a resource (or any of its children) is * * \param[in] rsc Resource to check * \param[out] list List to add result to * \param[in] current 0 = where known, 1 = running, 2 = running or pending * * \return If list contains only one node, that node */ pe_node_t * native_location(const pe_resource_t *rsc, GList **list, int current) { node_t *one = NULL; GListPtr result = NULL; if (rsc->children) { GListPtr gIter = rsc->children; for (; gIter != NULL; gIter = gIter->next) { resource_t *child = (resource_t *) gIter->data; child->fns->location(child, &result, current); } } else if (current) { if (rsc->running_on) { result = g_list_copy(rsc->running_on); } if ((current == 2) && rsc->pending_node && !pe_find_node_id(result, rsc->pending_node->details->id)) { result = g_list_append(result, rsc->pending_node); } } else if (current == FALSE && rsc->allocated_to) { result = g_list_append(NULL, rsc->allocated_to); } if (result && (result->next == NULL)) { one = result->data; } if (list) { GListPtr gIter = result; for (; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; if (*list == NULL || pe_find_node_id(*list, node->details->id) == NULL) { *list = g_list_append(*list, node); } } } g_list_free(result); return one; } static void get_rscs_brief(GListPtr rsc_list, GHashTable * rsc_table, GHashTable * active_table) { GListPtr gIter = rsc_list; for (; gIter != NULL; gIter = gIter->next) { resource_t *rsc = (resource_t *) gIter->data; const char *class = crm_element_value(rsc->xml, XML_AGENT_ATTR_CLASS); const char *kind = crm_element_value(rsc->xml, XML_ATTR_TYPE); int offset = 0; char buffer[LINE_MAX]; int *rsc_counter = NULL; int *active_counter = NULL; if (rsc->variant != pe_native) { continue; } offset += snprintf(buffer + offset, LINE_MAX - offset, "%s", class); if (is_set(pcmk_get_ra_caps(class), pcmk_ra_cap_provider)) { const char *prov = crm_element_value(rsc->xml, XML_AGENT_ATTR_PROVIDER); offset += snprintf(buffer + offset, LINE_MAX - offset, "::%s", prov); } offset += snprintf(buffer + offset, LINE_MAX - offset, ":%s", kind); CRM_LOG_ASSERT(offset > 0); if (rsc_table) { rsc_counter = g_hash_table_lookup(rsc_table, buffer); if (rsc_counter == NULL) { rsc_counter = calloc(1, sizeof(int)); *rsc_counter = 0; g_hash_table_insert(rsc_table, strdup(buffer), rsc_counter); } (*rsc_counter)++; } if (active_table) { GListPtr gIter2 = rsc->running_on; for (; gIter2 != NULL; gIter2 = gIter2->next) { node_t *node = (node_t *) gIter2->data; GHashTable *node_table = NULL; if (node->details->unclean == FALSE && node->details->online == FALSE) { continue; } node_table = g_hash_table_lookup(active_table, node->details->uname); if (node_table == NULL) { node_table = crm_str_table_new(); g_hash_table_insert(active_table, strdup(node->details->uname), node_table); } active_counter = g_hash_table_lookup(node_table, buffer); if (active_counter == NULL) { active_counter = calloc(1, sizeof(int)); *active_counter = 0; g_hash_table_insert(node_table, strdup(buffer), active_counter); } (*active_counter)++; } } } } static void destroy_node_table(gpointer data) { GHashTable *node_table = data; if (node_table) { g_hash_table_destroy(node_table); } } void print_rscs_brief(GListPtr rsc_list, const char *pre_text, long options, void *print_data, gboolean print_all) { GHashTable *rsc_table = crm_str_table_new(); GHashTable *active_table = g_hash_table_new_full(crm_str_hash, g_str_equal, free, destroy_node_table); GHashTableIter hash_iter; char *type = NULL; int *rsc_counter = NULL; get_rscs_brief(rsc_list, rsc_table, active_table); g_hash_table_iter_init(&hash_iter, rsc_table); while (g_hash_table_iter_next(&hash_iter, (gpointer *)&type, (gpointer *)&rsc_counter)) { GHashTableIter hash_iter2; char *node_name = NULL; GHashTable *node_table = NULL; int active_counter_all = 0; g_hash_table_iter_init(&hash_iter2, active_table); while (g_hash_table_iter_next(&hash_iter2, (gpointer *)&node_name, (gpointer *)&node_table)) { int *active_counter = g_hash_table_lookup(node_table, type); if (active_counter == NULL || *active_counter == 0) { continue; } else { active_counter_all += *active_counter; } if (options & pe_print_rsconly) { node_name = NULL; } if (options & pe_print_html) { status_print("
  • \n"); } if (print_all) { status_print("%s%d/%d\t(%s):\tActive %s\n", pre_text ? pre_text : "", active_counter ? *active_counter : 0, rsc_counter ? *rsc_counter : 0, type, active_counter && (*active_counter > 0) && node_name ? node_name : ""); } else { status_print("%s%d\t(%s):\tActive %s\n", pre_text ? pre_text : "", active_counter ? *active_counter : 0, type, active_counter && (*active_counter > 0) && node_name ? node_name : ""); } if (options & pe_print_html) { status_print("
    • \n"); } } if (print_all && active_counter_all == 0) { if (options & pe_print_html) { status_print("
  • \n"); } status_print("%s%d/%d\t(%s):\tActive\n", pre_text ? pre_text : "", active_counter_all, rsc_counter ? *rsc_counter : 0, type); if (options & pe_print_html) { status_print("
    • \n"); } } } if (rsc_table) { g_hash_table_destroy(rsc_table); rsc_table = NULL; } if (active_table) { g_hash_table_destroy(active_table); active_table = NULL; } } void pe__rscs_brief_output(pcmk__output_t *out, GListPtr rsc_list, long options, gboolean print_all) { GHashTable *rsc_table = crm_str_table_new(); GHashTable *active_table = g_hash_table_new_full(crm_str_hash, g_str_equal, free, destroy_node_table); GHashTableIter hash_iter; char *type = NULL; int *rsc_counter = NULL; get_rscs_brief(rsc_list, rsc_table, active_table); g_hash_table_iter_init(&hash_iter, rsc_table); while (g_hash_table_iter_next(&hash_iter, (gpointer *)&type, (gpointer *)&rsc_counter)) { GHashTableIter hash_iter2; char *node_name = NULL; GHashTable *node_table = NULL; int active_counter_all = 0; g_hash_table_iter_init(&hash_iter2, active_table); while (g_hash_table_iter_next(&hash_iter2, (gpointer *)&node_name, (gpointer *)&node_table)) { int *active_counter = g_hash_table_lookup(node_table, type); if (active_counter == NULL || *active_counter == 0) { continue; } else { active_counter_all += *active_counter; } if (options & pe_print_rsconly) { node_name = NULL; } if (print_all) { out->list_item(out, NULL, " %d/%d\t(%s):\tActive %s", *active_counter, rsc_counter ? *rsc_counter : 0, type, (*active_counter > 0) && node_name ? node_name : ""); } else { out->list_item(out, NULL, " %d\t(%s):\tActive %s", *active_counter, type, (*active_counter > 0) && node_name ? node_name : ""); } } if (print_all && active_counter_all == 0) { out->list_item(out, NULL, " %d/%d\t(%s):\tActive", active_counter_all, rsc_counter ? *rsc_counter : 0, type); } } if (rsc_table) { g_hash_table_destroy(rsc_table); rsc_table = NULL; } if (active_table) { g_hash_table_destroy(active_table); active_table = NULL; } } diff --git a/lib/pengine/pe_output.c b/lib/pengine/pe_output.c index 76a143f135..1dc349dd4c 100644 --- a/lib/pengine/pe_output.c +++ b/lib/pengine/pe_output.c @@ -1,173 +1,160 @@ /* * Copyright 2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU Lesser General Public License * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY. */ #include #include #include int pe__name_and_nvpairs_xml(pcmk__output_t *out, bool is_list, const char *tag_name , size_t pairs_count, ...) { xmlNodePtr xml_node = NULL; va_list args; CRM_ASSERT(tag_name != NULL); xml_node = pcmk__output_xml_peek_parent(out); CRM_ASSERT(xml_node != NULL); xml_node = is_list ? create_xml_node(xml_node, tag_name) : xmlNewChild(xml_node, NULL, (pcmkXmlStr) tag_name, NULL); va_start(args, pairs_count); while(pairs_count--) { const char *param_name = va_arg(args, const char *); const char *param_value = va_arg(args, const char *); if (param_name && param_value) { xmlSetProp(xml_node, (pcmkXmlStr)param_name, (pcmkXmlStr)param_value); } }; va_end(args); if (is_list) { pcmk__output_xml_push_parent(out, xml_node); } return 0; } static int pe__ticket_html(pcmk__output_t *out, va_list args) { ticket_t *ticket = va_arg(args, ticket_t *); if (ticket->last_granted > -1) { char *time = pcmk_format_named_time("last-granted", ticket->last_granted); out->list_item(out, NULL, "%s:\t%s%s %s", ticket->id, ticket->granted ? "granted" : "revoked", ticket->standby ? " [standby]" : "", time); free(time); } else { out->list_item(out, NULL, "%s:\t%s%s", ticket->id, ticket->granted ? "granted" : "revoked", ticket->standby ? " [standby]" : ""); } return 0; } static int pe__ticket_text(pcmk__output_t *out, va_list args) { ticket_t *ticket = va_arg(args, ticket_t *); if (ticket->last_granted > -1) { char *time = pcmk_format_named_time("last-granted", ticket->last_granted); out->list_item(out, ticket->id, "\t%s%s %s", ticket->granted ? "granted" : "revoked", ticket->standby ? " [standby]" : "", time); free(time); } else { out->list_item(out, ticket->id, "\t%s%s", ticket->granted ? "granted" : "revoked", ticket->standby ? " [standby]" : ""); } return 0; } static int pe__ticket_xml(pcmk__output_t *out, va_list args) { xmlNodePtr node = NULL; ticket_t *ticket = va_arg(args, ticket_t *); node = pcmk__output_create_xml_node(out, "ticket"); xmlSetProp(node, (pcmkXmlStr) "id", (pcmkXmlStr) ticket->id); xmlSetProp(node, (pcmkXmlStr) "status", (pcmkXmlStr) (ticket->granted ? "granted" : "revoked")); xmlSetProp(node, (pcmkXmlStr) "standby", (pcmkXmlStr) (ticket->standby ? "true" : "false")); if (ticket->last_granted > -1) { xmlSetProp(node, (pcmkXmlStr) "last-granted", (pcmkXmlStr) crm_now_string(&ticket->last_granted)); } return 0; } static pcmk__message_entry_t fmt_functions[] = { { "bundle", "xml", pe__bundle_xml }, { "bundle", "html", pe__bundle_html }, { "bundle", "text", pe__bundle_text }, { "clone", "xml", pe__clone_xml }, { "clone", "html", pe__clone_html }, { "clone", "text", pe__clone_text }, { "group", "xml", pe__group_xml }, { "group", "html", pe__group_html }, { "group", "text", pe__group_text }, { "primitive", "xml", pe__resource_xml }, { "primitive", "html", pe__resource_html }, { "primitive", "text", pe__resource_text }, { "ticket", "html", pe__ticket_html }, { "ticket", "text", pe__ticket_text }, { "ticket", "xml", pe__ticket_xml }, { NULL, NULL, NULL } }; void pe__register_messages(pcmk__output_t *out) { pcmk__register_messages(out, fmt_functions); } void pe__output_node(node_t *node, gboolean details, pcmk__output_t *out) { if (node == NULL) { crm_trace(""); return; } CRM_ASSERT(node->details); crm_trace("%sNode %s: (weight=%d, fixed=%s)", node->details->online ? "" : "Unavailable/Unclean ", node->details->uname, node->weight, node->fixed ? "True" : "False"); if (details) { char *pe_mutable = strdup("\t\t"); GListPtr gIter = node->details->running_rsc; crm_trace("\t\t===Node Attributes"); g_hash_table_foreach(node->details->attrs, print_str_str, pe_mutable); free(pe_mutable); crm_trace("\t\t=== Resources"); for (; gIter != NULL; gIter = gIter->next) { resource_t *rsc = (resource_t *) gIter->data; - pe__output_resource(LOG_TRACE, rsc, FALSE, out); + // @TODO pe_print_log probably doesn't belong here + out->message(out, crm_map_element_name(rsc->xml), + pe_print_log|pe_print_pending, rsc); } } } - -void -pe__output_resource(int log_level, resource_t *rsc, gboolean details, pcmk__output_t *out) -{ - long options = pe_print_log | pe_print_pending; - - if (rsc == NULL) { - do_crm_log(log_level - 1, ""); - return; - } - if (details) { - options |= pe_print_details; - } - out->message(out, crm_map_element_name(rsc->xml), options, rsc); -} diff --git a/lib/pengine/status.c b/lib/pengine/status.c index a8b094775f..1ef1488e9b 100644 --- a/lib/pengine/status.c +++ b/lib/pengine/status.c @@ -1,428 +1,435 @@ /* * Copyright 2004-2018 Andrew Beekhof * * This source code is licensed under the GNU Lesser General Public License * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include /*! * \brief Create a new working set * * \return New, initialized working set on success, else NULL (and set errno) * \note Only pe_working_set_t objects created with this function (as opposed * to statically declared or directly allocated) should be used with the * functions in this library, to allow for future extensions to the * data type. The caller is responsible for freeing the memory with * pe_free_working_set() when the instance is no longer needed. */ pe_working_set_t * pe_new_working_set() { pe_working_set_t *data_set = calloc(1, sizeof(pe_working_set_t)); if (data_set != NULL) { set_working_set_defaults(data_set); } return data_set; } /*! * \brief Free a working set * * \param[in] data_set Working set to free */ void pe_free_working_set(pe_working_set_t *data_set) { if (data_set != NULL) { pe_reset_working_set(data_set); free(data_set); } } /* * Unpack everything * At the end you'll have: * - A list of nodes * - A list of resources (each with any dependencies on other resources) * - A list of constraints between resources and nodes * - A list of constraints between start/stop actions * - A list of nodes that need to be stonith'd * - A list of nodes that need to be shutdown * - A list of the possible stop/start actions (without dependencies) */ gboolean cluster_status(pe_working_set_t * data_set) { xmlNode *config = get_xpath_object("//"XML_CIB_TAG_CRMCONFIG, data_set->input, LOG_TRACE); xmlNode *cib_nodes = get_xpath_object("//"XML_CIB_TAG_NODES, data_set->input, LOG_TRACE); xmlNode *cib_resources = get_xpath_object("//"XML_CIB_TAG_RESOURCES, data_set->input, LOG_TRACE); xmlNode *cib_status = get_xpath_object("//"XML_CIB_TAG_STATUS, data_set->input, LOG_TRACE); xmlNode *cib_tags = get_xpath_object("//"XML_CIB_TAG_TAGS, data_set->input, LOG_TRACE); const char *value = crm_element_value(data_set->input, XML_ATTR_HAVE_QUORUM); crm_trace("Beginning unpack"); /* reset remaining global variables */ data_set->failed = create_xml_node(NULL, "failed-ops"); if (data_set->input == NULL) { return FALSE; } if (data_set->now == NULL) { data_set->now = crm_time_new(NULL); } if (data_set->dc_uuid == NULL) { data_set->dc_uuid = crm_element_value_copy(data_set->input, XML_ATTR_DC_UUID); } clear_bit(data_set->flags, pe_flag_have_quorum); if (crm_is_true(value)) { set_bit(data_set->flags, pe_flag_have_quorum); } data_set->op_defaults = get_xpath_object("//"XML_CIB_TAG_OPCONFIG, data_set->input, LOG_TRACE); data_set->rsc_defaults = get_xpath_object("//"XML_CIB_TAG_RSCCONFIG, data_set->input, LOG_TRACE); unpack_config(config, data_set); if (is_not_set(data_set->flags, pe_flag_quick_location) && is_not_set(data_set->flags, pe_flag_have_quorum) && data_set->no_quorum_policy != no_quorum_ignore) { crm_warn("Fencing and resource management disabled due to lack of quorum"); } unpack_nodes(cib_nodes, data_set); if(is_not_set(data_set->flags, pe_flag_quick_location)) { unpack_remote_nodes(cib_resources, data_set); } unpack_resources(cib_resources, data_set); unpack_tags(cib_tags, data_set); if(is_not_set(data_set->flags, pe_flag_quick_location)) { unpack_status(cib_status, data_set); } + if (is_not_set(data_set->flags, pe_flag_no_counts)) { + for (GList *item = data_set->resources; item != NULL; + item = item->next) { + ((pe_resource_t *) (item->data))->fns->count(item->data); + } + } + set_bit(data_set->flags, pe_flag_have_status); return TRUE; } /*! * \internal * \brief Free a list of pe_resource_t * * \param[in] resources List to free * * \note When a working set's resource list is freed, that includes the original * storage for the uname and id of any Pacemaker Remote nodes in the * working set's node list, so take care not to use those afterward. * \todo Refactor pe_node_t to strdup() the node name. */ static void pe_free_resources(GListPtr resources) { resource_t *rsc = NULL; GListPtr iterator = resources; while (iterator != NULL) { rsc = (resource_t *) iterator->data; iterator = iterator->next; rsc->fns->free(rsc); } if (resources != NULL) { g_list_free(resources); } } static void pe_free_actions(GListPtr actions) { GListPtr iterator = actions; while (iterator != NULL) { pe_free_action(iterator->data); iterator = iterator->next; } if (actions != NULL) { g_list_free(actions); } } static void pe_free_nodes(GListPtr nodes) { for (GList *iterator = nodes; iterator != NULL; iterator = iterator->next) { pe_node_t *node = (pe_node_t *) iterator->data; // Shouldn't be possible, but to be safe ... if (node == NULL) { continue; } if (node->details == NULL) { free(node); continue; } /* This is called after pe_free_resources(), which means that we can't * use node->details->uname for Pacemaker Remote nodes. */ crm_trace("Freeing node %s", (pe__is_guest_or_remote_node(node)? "(guest or remote)" : node->details->uname)); if (node->details->attrs != NULL) { g_hash_table_destroy(node->details->attrs); } if (node->details->utilization != NULL) { g_hash_table_destroy(node->details->utilization); } if (node->details->digest_cache != NULL) { g_hash_table_destroy(node->details->digest_cache); } g_list_free(node->details->running_rsc); g_list_free(node->details->allocated_rsc); free(node->details); free(node); } if (nodes != NULL) { g_list_free(nodes); } } static void pe__free_ordering(GListPtr constraints) { GListPtr iterator = constraints; while (iterator != NULL) { pe__ordering_t *order = iterator->data; iterator = iterator->next; free(order->lh_action_task); free(order->rh_action_task); free(order); } if (constraints != NULL) { g_list_free(constraints); } } static void pe__free_location(GListPtr constraints) { GListPtr iterator = constraints; while (iterator != NULL) { pe__location_t *cons = iterator->data; iterator = iterator->next; g_list_free_full(cons->node_list_rh, free); free(cons->id); free(cons); } if (constraints != NULL) { g_list_free(constraints); } } /*! * \brief Reset working set to default state without freeing it or constraints * * \param[in,out] data_set Working set to reset * * \deprecated This function is deprecated as part of the API; * pe_reset_working_set() should be used instead. */ void cleanup_calculations(pe_working_set_t * data_set) { if (data_set == NULL) { return; } clear_bit(data_set->flags, pe_flag_have_status); if (data_set->config_hash != NULL) { g_hash_table_destroy(data_set->config_hash); } if (data_set->singletons != NULL) { g_hash_table_destroy(data_set->singletons); } if (data_set->tickets) { g_hash_table_destroy(data_set->tickets); } if (data_set->template_rsc_sets) { g_hash_table_destroy(data_set->template_rsc_sets); } if (data_set->tags) { g_hash_table_destroy(data_set->tags); } free(data_set->dc_uuid); crm_trace("deleting resources"); pe_free_resources(data_set->resources); crm_trace("deleting actions"); pe_free_actions(data_set->actions); crm_trace("deleting nodes"); pe_free_nodes(data_set->nodes); pe__free_param_checks(data_set); g_list_free(data_set->stop_needed); free_xml(data_set->graph); crm_time_free(data_set->now); free_xml(data_set->input); free_xml(data_set->failed); set_working_set_defaults(data_set); CRM_CHECK(data_set->ordering_constraints == NULL,; ); CRM_CHECK(data_set->placement_constraints == NULL,; ); } /*! * \brief Reset a working set to default state without freeing it * * \param[in,out] data_set Working set to reset */ void pe_reset_working_set(pe_working_set_t *data_set) { if (data_set == NULL) { return; } crm_trace("Deleting %d ordering constraints", g_list_length(data_set->ordering_constraints)); pe__free_ordering(data_set->ordering_constraints); data_set->ordering_constraints = NULL; crm_trace("Deleting %d location constraints", g_list_length(data_set->placement_constraints)); pe__free_location(data_set->placement_constraints); data_set->placement_constraints = NULL; crm_trace("Deleting %d colocation constraints", g_list_length(data_set->colocation_constraints)); g_list_free_full(data_set->colocation_constraints, free); data_set->colocation_constraints = NULL; crm_trace("Deleting %d ticket constraints", g_list_length(data_set->ticket_constraints)); g_list_free_full(data_set->ticket_constraints, free); data_set->ticket_constraints = NULL; cleanup_calculations(data_set); } void set_working_set_defaults(pe_working_set_t * data_set) { memset(data_set, 0, sizeof(pe_working_set_t)); data_set->order_id = 1; data_set->action_id = 1; data_set->no_quorum_policy = no_quorum_freeze; data_set->flags = 0x0ULL; set_bit(data_set->flags, pe_flag_stop_rsc_orphans); set_bit(data_set->flags, pe_flag_symmetric_cluster); set_bit(data_set->flags, pe_flag_stop_action_orphans); #ifdef DEFAULT_CONCURRENT_FENCING_TRUE set_bit(data_set->flags, pe_flag_concurrent_fencing); #endif } resource_t * pe_find_resource(GListPtr rsc_list, const char *id) { return pe_find_resource_with_flags(rsc_list, id, pe_find_renamed); } resource_t * pe_find_resource_with_flags(GListPtr rsc_list, const char *id, enum pe_find flags) { GListPtr rIter = NULL; for (rIter = rsc_list; id && rIter; rIter = rIter->next) { resource_t *parent = rIter->data; resource_t *match = parent->fns->find_rsc(parent, id, NULL, flags); if (match != NULL) { return match; } } crm_trace("No match for %s", id); return NULL; } node_t * pe_find_node_any(GListPtr nodes, const char *id, const char *uname) { node_t *match = pe_find_node_id(nodes, id); if (match) { return match; } crm_trace("Looking up %s via its uname instead", uname); return pe_find_node(nodes, uname); } node_t * pe_find_node_id(GListPtr nodes, const char *id) { GListPtr gIter = nodes; for (; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; if (node && safe_str_eq(node->details->id, id)) { return node; } } /* error */ return NULL; } node_t * pe_find_node(GListPtr nodes, const char *uname) { GListPtr gIter = nodes; for (; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; if (node && safe_str_eq(node->details->uname, uname)) { return node; } } /* error */ return NULL; } diff --git a/lib/pengine/unpack.c b/lib/pengine/unpack.c index 1b7b8eb4c4..64c1979bf8 100644 --- a/lib/pengine/unpack.c +++ b/lib/pengine/unpack.c @@ -1,3775 +1,3774 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU Lesser General Public License * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include CRM_TRACE_INIT_DATA(pe_status); #define set_config_flag(data_set, option, flag) do { \ const char *tmp = pe_pref(data_set->config_hash, option); \ if(tmp) { \ if(crm_is_true(tmp)) { \ set_bit(data_set->flags, flag); \ } else { \ clear_bit(data_set->flags, flag); \ } \ } \ } while(0) static void unpack_rsc_op(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op, xmlNode **last_failure, enum action_fail_response *failed, pe_working_set_t *data_set); static gboolean determine_remote_online_status(pe_working_set_t * data_set, node_t * this_node); static void add_node_attrs(xmlNode *attrs, pe_node_t *node, bool overwrite, pe_working_set_t *data_set); // Bitmask for warnings we only want to print once uint32_t pe_wo = 0; static gboolean is_dangling_guest_node(node_t *node) { /* we are looking for a remote-node that was supposed to be mapped to a * container resource, but all traces of that container have disappeared * from both the config and the status section. */ if (pe__is_guest_or_remote_node(node) && node->details->remote_rsc && node->details->remote_rsc->container == NULL && is_set(node->details->remote_rsc->flags, pe_rsc_orphan_container_filler)) { return TRUE; } return FALSE; } /*! * \brief Schedule a fence action for a node * * \param[in,out] data_set Current working set of cluster * \param[in,out] node Node to fence * \param[in] reason Text description of why fencing is needed */ void pe_fence_node(pe_working_set_t * data_set, node_t * node, const char *reason) { CRM_CHECK(node, return); /* A guest node is fenced by marking its container as failed */ if (pe__is_guest_node(node)) { resource_t *rsc = node->details->remote_rsc->container; if (is_set(rsc->flags, pe_rsc_failed) == FALSE) { if (!is_set(rsc->flags, pe_rsc_managed)) { crm_notice("Not fencing guest node %s " "(otherwise would because %s): " "its guest resource %s is unmanaged", node->details->uname, reason, rsc->id); } else { crm_warn("Guest node %s will be fenced " "(by recovering its guest resource %s): %s", node->details->uname, rsc->id, reason); /* We don't mark the node as unclean because that would prevent the * node from running resources. We want to allow it to run resources * in this transition if the recovery succeeds. */ node->details->remote_requires_reset = TRUE; set_bit(rsc->flags, pe_rsc_failed); } } } else if (is_dangling_guest_node(node)) { crm_info("Cleaning up dangling connection for guest node %s: " "fencing was already done because %s, " "and guest resource no longer exists", node->details->uname, reason); set_bit(node->details->remote_rsc->flags, pe_rsc_failed); } else if (pe__is_remote_node(node)) { resource_t *rsc = node->details->remote_rsc; if (rsc && (!is_set(rsc->flags, pe_rsc_managed))) { crm_notice("Not fencing remote node %s " "(otherwise would because %s): connection is unmanaged", node->details->uname, reason); } else if(node->details->remote_requires_reset == FALSE) { node->details->remote_requires_reset = TRUE; crm_warn("Remote node %s %s: %s", node->details->uname, pe_can_fence(data_set, node)? "will be fenced" : "is unclean", reason); } node->details->unclean = TRUE; pe_fence_op(node, NULL, TRUE, reason, data_set); } else if (node->details->unclean) { crm_trace("Cluster node %s %s because %s", node->details->uname, pe_can_fence(data_set, node)? "would also be fenced" : "also is unclean", reason); } else { crm_warn("Cluster node %s %s: %s", node->details->uname, pe_can_fence(data_set, node)? "will be fenced" : "is unclean", reason); node->details->unclean = TRUE; pe_fence_op(node, NULL, TRUE, reason, data_set); } } // @TODO xpaths can't handle templates, rules, or id-refs // nvpair with provides or requires set to unfencing #define XPATH_UNFENCING_NVPAIR XML_CIB_TAG_NVPAIR \ "[(@" XML_NVPAIR_ATTR_NAME "='" XML_RSC_ATTR_PROVIDES "'" \ "or @" XML_NVPAIR_ATTR_NAME "='" XML_RSC_ATTR_REQUIRES "') " \ "and @" XML_NVPAIR_ATTR_VALUE "='unfencing']" // unfencing in rsc_defaults or any resource #define XPATH_ENABLE_UNFENCING \ "/" XML_TAG_CIB "/" XML_CIB_TAG_CONFIGURATION "/" XML_CIB_TAG_RESOURCES \ "//" XML_TAG_META_SETS "/" XPATH_UNFENCING_NVPAIR \ "|/" XML_TAG_CIB "/" XML_CIB_TAG_CONFIGURATION "/" XML_CIB_TAG_RSCCONFIG \ "/" XML_TAG_META_SETS "/" XPATH_UNFENCING_NVPAIR static void set_if_xpath(unsigned long long flag, const char *xpath, pe_working_set_t *data_set) { xmlXPathObjectPtr result = NULL; if (is_not_set(data_set->flags, flag)) { result = xpath_search(data_set->input, xpath); if (result && (numXpathResults(result) > 0)) { set_bit(data_set->flags, flag); } freeXpathObject(result); } } gboolean unpack_config(xmlNode * config, pe_working_set_t * data_set) { const char *value = NULL; GHashTable *config_hash = crm_str_table_new(); data_set->config_hash = config_hash; pe__unpack_dataset_nvpairs(config, XML_CIB_TAG_PROPSET, NULL, config_hash, CIB_OPTIONS_FIRST, FALSE, data_set); verify_pe_options(data_set->config_hash); set_config_flag(data_set, "enable-startup-probes", pe_flag_startup_probes); if(is_not_set(data_set->flags, pe_flag_startup_probes)) { crm_info("Startup probes: disabled (dangerous)"); } value = pe_pref(data_set->config_hash, XML_ATTR_HAVE_WATCHDOG); if (value && crm_is_true(value)) { crm_notice("Watchdog will be used via SBD if fencing is required " "and stonith-watchdog-timeout is nonzero"); set_bit(data_set->flags, pe_flag_have_stonith_resource); } /* Set certain flags via xpath here, so they can be used before the relevant * configuration sections are unpacked. */ set_if_xpath(pe_flag_enable_unfencing, XPATH_ENABLE_UNFENCING, data_set); value = pe_pref(data_set->config_hash, "stonith-timeout"); data_set->stonith_timeout = (int) crm_parse_interval_spec(value); crm_debug("STONITH timeout: %d", data_set->stonith_timeout); set_config_flag(data_set, "stonith-enabled", pe_flag_stonith_enabled); crm_debug("STONITH of failed nodes is %s", is_set(data_set->flags, pe_flag_stonith_enabled) ? "enabled" : "disabled"); data_set->stonith_action = pe_pref(data_set->config_hash, "stonith-action"); if (!strcmp(data_set->stonith_action, "poweroff")) { pe_warn_once(pe_wo_poweroff, "Support for stonith-action of 'poweroff' is deprecated " "and will be removed in a future release (use 'off' instead)"); data_set->stonith_action = "off"; } crm_trace("STONITH will %s nodes", data_set->stonith_action); set_config_flag(data_set, "concurrent-fencing", pe_flag_concurrent_fencing); crm_debug("Concurrent fencing is %s", is_set(data_set->flags, pe_flag_concurrent_fencing) ? "enabled" : "disabled"); set_config_flag(data_set, "stop-all-resources", pe_flag_stop_everything); crm_debug("Stop all active resources: %s", is_set(data_set->flags, pe_flag_stop_everything) ? "true" : "false"); set_config_flag(data_set, "symmetric-cluster", pe_flag_symmetric_cluster); if (is_set(data_set->flags, pe_flag_symmetric_cluster)) { crm_debug("Cluster is symmetric" " - resources can run anywhere by default"); } value = pe_pref(data_set->config_hash, "no-quorum-policy"); if (safe_str_eq(value, "ignore")) { data_set->no_quorum_policy = no_quorum_ignore; } else if (safe_str_eq(value, "freeze")) { data_set->no_quorum_policy = no_quorum_freeze; } else if (safe_str_eq(value, "suicide")) { if (is_set(data_set->flags, pe_flag_stonith_enabled)) { int do_panic = 0; crm_element_value_int(data_set->input, XML_ATTR_QUORUM_PANIC, &do_panic); if (do_panic || is_set(data_set->flags, pe_flag_have_quorum)) { data_set->no_quorum_policy = no_quorum_suicide; } else { crm_notice("Resetting no-quorum-policy to 'stop': cluster has never had quorum"); data_set->no_quorum_policy = no_quorum_stop; } } else { crm_config_err("Resetting no-quorum-policy to 'stop': stonith is not configured"); data_set->no_quorum_policy = no_quorum_stop; } } else { data_set->no_quorum_policy = no_quorum_stop; } switch (data_set->no_quorum_policy) { case no_quorum_freeze: crm_debug("On loss of quorum: Freeze resources"); break; case no_quorum_stop: crm_debug("On loss of quorum: Stop ALL resources"); break; case no_quorum_suicide: crm_notice("On loss of quorum: Fence all remaining nodes"); break; case no_quorum_ignore: crm_notice("On loss of quorum: Ignore"); break; } set_config_flag(data_set, "stop-orphan-resources", pe_flag_stop_rsc_orphans); crm_trace("Orphan resources are %s", is_set(data_set->flags, pe_flag_stop_rsc_orphans) ? "stopped" : "ignored"); set_config_flag(data_set, "stop-orphan-actions", pe_flag_stop_action_orphans); crm_trace("Orphan resource actions are %s", is_set(data_set->flags, pe_flag_stop_action_orphans) ? "stopped" : "ignored"); set_config_flag(data_set, "remove-after-stop", pe_flag_remove_after_stop); crm_trace("Stopped resources are removed from the status section: %s", is_set(data_set->flags, pe_flag_remove_after_stop) ? "true" : "false"); set_config_flag(data_set, "maintenance-mode", pe_flag_maintenance_mode); crm_trace("Maintenance mode: %s", is_set(data_set->flags, pe_flag_maintenance_mode) ? "true" : "false"); set_config_flag(data_set, "start-failure-is-fatal", pe_flag_start_failure_fatal); crm_trace("Start failures are %s", is_set(data_set->flags, pe_flag_start_failure_fatal) ? "always fatal" : "handled by failcount"); if (is_set(data_set->flags, pe_flag_stonith_enabled)) { set_config_flag(data_set, "startup-fencing", pe_flag_startup_fencing); } if (is_set(data_set->flags, pe_flag_startup_fencing)) { crm_trace("Unseen nodes will be fenced"); } else { pe_warn_once(pe_wo_blind, "Blind faith: not fencing unseen nodes"); } node_score_red = char2score(pe_pref(data_set->config_hash, "node-health-red")); node_score_green = char2score(pe_pref(data_set->config_hash, "node-health-green")); node_score_yellow = char2score(pe_pref(data_set->config_hash, "node-health-yellow")); crm_debug("Node scores: 'red' = %s, 'yellow' = %s, 'green' = %s", pe_pref(data_set->config_hash, "node-health-red"), pe_pref(data_set->config_hash, "node-health-yellow"), pe_pref(data_set->config_hash, "node-health-green")); data_set->placement_strategy = pe_pref(data_set->config_hash, "placement-strategy"); crm_trace("Placement strategy: %s", data_set->placement_strategy); return TRUE; } static void destroy_digest_cache(gpointer ptr) { op_digest_cache_t *data = ptr; free_xml(data->params_all); free_xml(data->params_secure); free_xml(data->params_restart); free(data->digest_all_calc); free(data->digest_restart_calc); free(data->digest_secure_calc); free(data); } node_t * pe_create_node(const char *id, const char *uname, const char *type, const char *score, pe_working_set_t * data_set) { node_t *new_node = NULL; if (pe_find_node(data_set->nodes, uname) != NULL) { crm_config_warn("Detected multiple node entries with uname=%s" " - this is rarely intended", uname); } new_node = calloc(1, sizeof(node_t)); if (new_node == NULL) { return NULL; } new_node->weight = char2score(score); new_node->fixed = FALSE; new_node->details = calloc(1, sizeof(struct pe_node_shared_s)); if (new_node->details == NULL) { free(new_node); return NULL; } crm_trace("Creating node for entry %s/%s", uname, id); new_node->details->id = id; new_node->details->uname = uname; new_node->details->online = FALSE; new_node->details->shutdown = FALSE; new_node->details->rsc_discovery_enabled = TRUE; new_node->details->running_rsc = NULL; new_node->details->type = node_ping; if (safe_str_eq(type, "remote")) { new_node->details->type = node_remote; set_bit(data_set->flags, pe_flag_have_remote_nodes); } else if ((type == NULL) || safe_str_eq(type, "member")) { new_node->details->type = node_member; } new_node->details->attrs = crm_str_table_new(); if (pe__is_guest_or_remote_node(new_node)) { g_hash_table_insert(new_node->details->attrs, strdup(CRM_ATTR_KIND), strdup("remote")); } else { g_hash_table_insert(new_node->details->attrs, strdup(CRM_ATTR_KIND), strdup("cluster")); } new_node->details->utilization = crm_str_table_new(); new_node->details->digest_cache = g_hash_table_new_full(crm_str_hash, g_str_equal, free, destroy_digest_cache); data_set->nodes = g_list_insert_sorted(data_set->nodes, new_node, sort_node_uname); return new_node; } bool remote_id_conflict(const char *remote_name, pe_working_set_t *data) { bool match = FALSE; #if 1 pe_find_resource(data->resources, remote_name); #else if (data->name_check == NULL) { data->name_check = g_hash_table_new(crm_str_hash, g_str_equal); for (xml_rsc = __xml_first_child_element(parent); xml_rsc != NULL; xml_rsc = __xml_next_element(xml_rsc)) { const char *id = ID(xml_rsc); /* avoiding heap allocation here because we know the duration of this hashtable allows us to */ g_hash_table_insert(data->name_check, (char *) id, (char *) id); } } if (g_hash_table_lookup(data->name_check, remote_name)) { match = TRUE; } #endif if (match) { crm_err("Invalid remote-node name, a resource called '%s' already exists.", remote_name); return NULL; } return match; } static const char * expand_remote_rsc_meta(xmlNode *xml_obj, xmlNode *parent, pe_working_set_t *data) { xmlNode *attr_set = NULL; xmlNode *attr = NULL; const char *container_id = ID(xml_obj); const char *remote_name = NULL; const char *remote_server = NULL; const char *remote_port = NULL; const char *connect_timeout = "60s"; const char *remote_allow_migrate=NULL; const char *is_managed = NULL; for (attr_set = __xml_first_child_element(xml_obj); attr_set != NULL; attr_set = __xml_next_element(attr_set)) { if (safe_str_neq((const char *)attr_set->name, XML_TAG_META_SETS)) { continue; } for (attr = __xml_first_child_element(attr_set); attr != NULL; attr = __xml_next_element(attr)) { const char *value = crm_element_value(attr, XML_NVPAIR_ATTR_VALUE); const char *name = crm_element_value(attr, XML_NVPAIR_ATTR_NAME); if (safe_str_eq(name, XML_RSC_ATTR_REMOTE_NODE)) { remote_name = value; } else if (safe_str_eq(name, "remote-addr")) { remote_server = value; } else if (safe_str_eq(name, "remote-port")) { remote_port = value; } else if (safe_str_eq(name, "remote-connect-timeout")) { connect_timeout = value; } else if (safe_str_eq(name, "remote-allow-migrate")) { remote_allow_migrate=value; } else if (safe_str_eq(name, XML_RSC_ATTR_MANAGED)) { is_managed = value; } } } if (remote_name == NULL) { return NULL; } if (remote_id_conflict(remote_name, data)) { return NULL; } pe_create_remote_xml(parent, remote_name, container_id, remote_allow_migrate, is_managed, connect_timeout, remote_server, remote_port); return remote_name; } static void handle_startup_fencing(pe_working_set_t *data_set, node_t *new_node) { if ((new_node->details->type == node_remote) && (new_node->details->remote_rsc == NULL)) { /* Ignore fencing for remote nodes that don't have a connection resource * associated with them. This happens when remote node entries get left * in the nodes section after the connection resource is removed. */ return; } if (is_set(data_set->flags, pe_flag_startup_fencing)) { // All nodes are unclean until we've seen their status entry new_node->details->unclean = TRUE; } else { // Blind faith ... new_node->details->unclean = FALSE; } /* We need to be able to determine if a node's status section * exists or not separate from whether the node is unclean. */ new_node->details->unseen = TRUE; } gboolean unpack_nodes(xmlNode * xml_nodes, pe_working_set_t * data_set) { xmlNode *xml_obj = NULL; node_t *new_node = NULL; const char *id = NULL; const char *uname = NULL; const char *type = NULL; const char *score = NULL; for (xml_obj = __xml_first_child_element(xml_nodes); xml_obj != NULL; xml_obj = __xml_next_element(xml_obj)) { if (crm_str_eq((const char *)xml_obj->name, XML_CIB_TAG_NODE, TRUE)) { new_node = NULL; id = crm_element_value(xml_obj, XML_ATTR_ID); uname = crm_element_value(xml_obj, XML_ATTR_UNAME); type = crm_element_value(xml_obj, XML_ATTR_TYPE); score = crm_element_value(xml_obj, XML_RULE_ATTR_SCORE); crm_trace("Processing node %s/%s", uname, id); if (id == NULL) { crm_config_err("Must specify id tag in "); continue; } new_node = pe_create_node(id, uname, type, score, data_set); if (new_node == NULL) { return FALSE; } /* if(data_set->have_quorum == FALSE */ /* && data_set->no_quorum_policy == no_quorum_stop) { */ /* /\* start shutting resources down *\/ */ /* new_node->weight = -INFINITY; */ /* } */ handle_startup_fencing(data_set, new_node); add_node_attrs(xml_obj, new_node, FALSE, data_set); pe__unpack_dataset_nvpairs(xml_obj, XML_TAG_UTILIZATION, NULL, new_node->details->utilization, NULL, FALSE, data_set); crm_trace("Done with node %s", crm_element_value(xml_obj, XML_ATTR_UNAME)); } } if (data_set->localhost && pe_find_node(data_set->nodes, data_set->localhost) == NULL) { crm_info("Creating a fake local node"); pe_create_node(data_set->localhost, data_set->localhost, NULL, 0, data_set); } return TRUE; } static void setup_container(resource_t * rsc, pe_working_set_t * data_set) { const char *container_id = NULL; if (rsc->children) { GListPtr gIter = rsc->children; for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; setup_container(child_rsc, data_set); } return; } container_id = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_CONTAINER); if (container_id && safe_str_neq(container_id, rsc->id)) { resource_t *container = pe_find_resource(data_set->resources, container_id); if (container) { rsc->container = container; set_bit(container->flags, pe_rsc_is_container); container->fillers = g_list_append(container->fillers, rsc); pe_rsc_trace(rsc, "Resource %s's container is %s", rsc->id, container_id); } else { pe_err("Resource %s: Unknown resource container (%s)", rsc->id, container_id); } } } gboolean unpack_remote_nodes(xmlNode * xml_resources, pe_working_set_t * data_set) { xmlNode *xml_obj = NULL; /* Create remote nodes and guest nodes from the resource configuration * before unpacking resources. */ for (xml_obj = __xml_first_child_element(xml_resources); xml_obj != NULL; xml_obj = __xml_next_element(xml_obj)) { const char *new_node_id = NULL; /* Check for remote nodes, which are defined by ocf:pacemaker:remote * primitives. */ if (xml_contains_remote_node(xml_obj)) { new_node_id = ID(xml_obj); /* The "pe_find_node" check is here to make sure we don't iterate over * an expanded node that has already been added to the node list. */ if (new_node_id && pe_find_node(data_set->nodes, new_node_id) == NULL) { crm_trace("Found remote node %s defined by resource %s", new_node_id, ID(xml_obj)); pe_create_node(new_node_id, new_node_id, "remote", NULL, data_set); } continue; } /* Check for guest nodes, which are defined by special meta-attributes * of a primitive of any type (for example, VirtualDomain or Xen). */ if (crm_str_eq((const char *)xml_obj->name, XML_CIB_TAG_RESOURCE, TRUE)) { /* This will add an ocf:pacemaker:remote primitive to the * configuration for the guest node's connection, to be unpacked * later. */ new_node_id = expand_remote_rsc_meta(xml_obj, xml_resources, data_set); if (new_node_id && pe_find_node(data_set->nodes, new_node_id) == NULL) { crm_trace("Found guest node %s in resource %s", new_node_id, ID(xml_obj)); pe_create_node(new_node_id, new_node_id, "remote", NULL, data_set); } continue; } /* Check for guest nodes inside a group. Clones are currently not * supported as guest nodes. */ if (crm_str_eq((const char *)xml_obj->name, XML_CIB_TAG_GROUP, TRUE)) { xmlNode *xml_obj2 = NULL; for (xml_obj2 = __xml_first_child_element(xml_obj); xml_obj2 != NULL; xml_obj2 = __xml_next_element(xml_obj2)) { new_node_id = expand_remote_rsc_meta(xml_obj2, xml_resources, data_set); if (new_node_id && pe_find_node(data_set->nodes, new_node_id) == NULL) { crm_trace("Found guest node %s in resource %s inside group %s", new_node_id, ID(xml_obj2), ID(xml_obj)); pe_create_node(new_node_id, new_node_id, "remote", NULL, data_set); } } } } return TRUE; } /* Call this after all the nodes and resources have been * unpacked, but before the status section is read. * * A remote node's online status is reflected by the state * of the remote node's connection resource. We need to link * the remote node to this connection resource so we can have * easy access to the connection resource during the PE calculations. */ static void link_rsc2remotenode(pe_working_set_t *data_set, resource_t *new_rsc) { node_t *remote_node = NULL; if (new_rsc->is_remote_node == FALSE) { return; } if (is_set(data_set->flags, pe_flag_quick_location)) { /* remote_nodes and remote_resources are not linked in quick location calculations */ return; } - print_resource(LOG_TRACE, "Linking remote-node connection resource, ", new_rsc, FALSE); - remote_node = pe_find_node(data_set->nodes, new_rsc->id); CRM_CHECK(remote_node != NULL, return;); + pe_rsc_trace(new_rsc, "Linking remote connection resource %s to node %s", + new_rsc->id, remote_node->details->uname); remote_node->details->remote_rsc = new_rsc; if (new_rsc->container == NULL) { /* Handle start-up fencing for remote nodes (as opposed to guest nodes) * the same as is done for cluster nodes. */ handle_startup_fencing(data_set, remote_node); } else { /* pe_create_node() marks the new node as "remote" or "cluster"; now * that we know the node is a guest node, update it correctly. */ g_hash_table_replace(remote_node->details->attrs, strdup(CRM_ATTR_KIND), strdup("container")); } } static void destroy_tag(gpointer data) { tag_t *tag = data; if (tag) { free(tag->id); g_list_free_full(tag->refs, free); free(tag); } } /*! * \internal * \brief Parse configuration XML for resource information * * \param[in] xml_resources Top of resource configuration XML * \param[in,out] data_set Where to put resource information * * \return TRUE * * \note unpack_remote_nodes() MUST be called before this, so that the nodes can * be used when common_unpack() calls resource_location() */ gboolean unpack_resources(xmlNode * xml_resources, pe_working_set_t * data_set) { xmlNode *xml_obj = NULL; GListPtr gIter = NULL; data_set->template_rsc_sets = g_hash_table_new_full(crm_str_hash, g_str_equal, free, destroy_tag); for (xml_obj = __xml_first_child_element(xml_resources); xml_obj != NULL; xml_obj = __xml_next_element(xml_obj)) { resource_t *new_rsc = NULL; if (crm_str_eq((const char *)xml_obj->name, XML_CIB_TAG_RSC_TEMPLATE, TRUE)) { const char *template_id = ID(xml_obj); if (template_id && g_hash_table_lookup_extended(data_set->template_rsc_sets, template_id, NULL, NULL) == FALSE) { /* Record the template's ID for the knowledge of its existence anyway. */ g_hash_table_insert(data_set->template_rsc_sets, strdup(template_id), NULL); } continue; } crm_trace("Beginning unpack... <%s id=%s... >", crm_element_name(xml_obj), ID(xml_obj)); if (common_unpack(xml_obj, &new_rsc, NULL, data_set)) { data_set->resources = g_list_append(data_set->resources, new_rsc); - print_resource(LOG_TRACE, "Added ", new_rsc, FALSE); + pe_rsc_trace(new_rsc, "Added resource %s", new_rsc->id); } else { crm_config_err("Failed unpacking %s %s", crm_element_name(xml_obj), crm_element_value(xml_obj, XML_ATTR_ID)); if (new_rsc != NULL && new_rsc->fns != NULL) { new_rsc->fns->free(new_rsc); } } } for (gIter = data_set->resources; gIter != NULL; gIter = gIter->next) { resource_t *rsc = (resource_t *) gIter->data; setup_container(rsc, data_set); link_rsc2remotenode(data_set, rsc); } data_set->resources = g_list_sort(data_set->resources, sort_rsc_priority); if (is_set(data_set->flags, pe_flag_quick_location)) { /* Ignore */ } else if (is_set(data_set->flags, pe_flag_stonith_enabled) && is_set(data_set->flags, pe_flag_have_stonith_resource) == FALSE) { crm_config_err("Resource start-up disabled since no STONITH resources have been defined"); crm_config_err("Either configure some or disable STONITH with the stonith-enabled option"); crm_config_err("NOTE: Clusters with shared data need STONITH to ensure data integrity"); } return TRUE; } gboolean unpack_tags(xmlNode * xml_tags, pe_working_set_t * data_set) { xmlNode *xml_tag = NULL; data_set->tags = g_hash_table_new_full(crm_str_hash, g_str_equal, free, destroy_tag); for (xml_tag = __xml_first_child_element(xml_tags); xml_tag != NULL; xml_tag = __xml_next_element(xml_tag)) { xmlNode *xml_obj_ref = NULL; const char *tag_id = ID(xml_tag); if (crm_str_eq((const char *)xml_tag->name, XML_CIB_TAG_TAG, TRUE) == FALSE) { continue; } if (tag_id == NULL) { crm_config_err("Failed unpacking %s: %s should be specified", crm_element_name(xml_tag), XML_ATTR_ID); continue; } for (xml_obj_ref = __xml_first_child_element(xml_tag); xml_obj_ref != NULL; xml_obj_ref = __xml_next_element(xml_obj_ref)) { const char *obj_ref = ID(xml_obj_ref); if (crm_str_eq((const char *)xml_obj_ref->name, XML_CIB_TAG_OBJ_REF, TRUE) == FALSE) { continue; } if (obj_ref == NULL) { crm_config_err("Failed unpacking %s for tag %s: %s should be specified", crm_element_name(xml_obj_ref), tag_id, XML_ATTR_ID); continue; } if (add_tag_ref(data_set->tags, tag_id, obj_ref) == FALSE) { return FALSE; } } } return TRUE; } /* The ticket state section: * "/cib/status/tickets/ticket_state" */ static gboolean unpack_ticket_state(xmlNode * xml_ticket, pe_working_set_t * data_set) { const char *ticket_id = NULL; const char *granted = NULL; const char *last_granted = NULL; const char *standby = NULL; xmlAttrPtr xIter = NULL; ticket_t *ticket = NULL; ticket_id = ID(xml_ticket); if (ticket_id == NULL || strlen(ticket_id) == 0) { return FALSE; } crm_trace("Processing ticket state for %s", ticket_id); ticket = g_hash_table_lookup(data_set->tickets, ticket_id); if (ticket == NULL) { ticket = ticket_new(ticket_id, data_set); if (ticket == NULL) { return FALSE; } } for (xIter = xml_ticket->properties; xIter; xIter = xIter->next) { const char *prop_name = (const char *)xIter->name; const char *prop_value = crm_element_value(xml_ticket, prop_name); if (crm_str_eq(prop_name, XML_ATTR_ID, TRUE)) { continue; } g_hash_table_replace(ticket->state, strdup(prop_name), strdup(prop_value)); } granted = g_hash_table_lookup(ticket->state, "granted"); if (granted && crm_is_true(granted)) { ticket->granted = TRUE; crm_info("We have ticket '%s'", ticket->id); } else { ticket->granted = FALSE; crm_info("We do not have ticket '%s'", ticket->id); } last_granted = g_hash_table_lookup(ticket->state, "last-granted"); if (last_granted) { ticket->last_granted = crm_parse_int(last_granted, 0); } standby = g_hash_table_lookup(ticket->state, "standby"); if (standby && crm_is_true(standby)) { ticket->standby = TRUE; if (ticket->granted) { crm_info("Granted ticket '%s' is in standby-mode", ticket->id); } } else { ticket->standby = FALSE; } crm_trace("Done with ticket state for %s", ticket_id); return TRUE; } static gboolean unpack_tickets_state(xmlNode * xml_tickets, pe_working_set_t * data_set) { xmlNode *xml_obj = NULL; for (xml_obj = __xml_first_child_element(xml_tickets); xml_obj != NULL; xml_obj = __xml_next_element(xml_obj)) { if (crm_str_eq((const char *)xml_obj->name, XML_CIB_TAG_TICKET_STATE, TRUE) == FALSE) { continue; } unpack_ticket_state(xml_obj, data_set); } return TRUE; } static void unpack_handle_remote_attrs(node_t *this_node, xmlNode *state, pe_working_set_t * data_set) { const char *resource_discovery_enabled = NULL; xmlNode *attrs = NULL; resource_t *rsc = NULL; if (crm_str_eq((const char *)state->name, XML_CIB_TAG_STATE, TRUE) == FALSE) { return; } if ((this_node == NULL) || !pe__is_guest_or_remote_node(this_node)) { return; } crm_trace("Processing remote node id=%s, uname=%s", this_node->details->id, this_node->details->uname); this_node->details->remote_maintenance = crm_atoi(crm_element_value(state, XML_NODE_IS_MAINTENANCE), "0"); rsc = this_node->details->remote_rsc; if (this_node->details->remote_requires_reset == FALSE) { this_node->details->unclean = FALSE; this_node->details->unseen = FALSE; } attrs = find_xml_node(state, XML_TAG_TRANSIENT_NODEATTRS, FALSE); add_node_attrs(attrs, this_node, TRUE, data_set); if (pe__shutdown_requested(this_node)) { crm_info("Node %s is shutting down", this_node->details->uname); this_node->details->shutdown = TRUE; if (rsc) { rsc->next_role = RSC_ROLE_STOPPED; } } if (crm_is_true(pe_node_attribute_raw(this_node, "standby"))) { crm_info("Node %s is in standby-mode", this_node->details->uname); this_node->details->standby = TRUE; } if (crm_is_true(pe_node_attribute_raw(this_node, "maintenance")) || (rsc && !is_set(rsc->flags, pe_rsc_managed))) { crm_info("Node %s is in maintenance-mode", this_node->details->uname); this_node->details->maintenance = TRUE; } resource_discovery_enabled = pe_node_attribute_raw(this_node, XML_NODE_ATTR_RSC_DISCOVERY); if (resource_discovery_enabled && !crm_is_true(resource_discovery_enabled)) { if (pe__is_remote_node(this_node) && is_not_set(data_set->flags, pe_flag_stonith_enabled)) { crm_warn("Ignoring %s attribute on remote node %s because stonith is disabled", XML_NODE_ATTR_RSC_DISCOVERY, this_node->details->uname); } else { /* This is either a remote node with fencing enabled, or a guest * node. We don't care whether fencing is enabled when fencing guest * nodes, because they are "fenced" by recovering their containing * resource. */ crm_info("Node %s has resource discovery disabled", this_node->details->uname); this_node->details->rsc_discovery_enabled = FALSE; } } } static bool unpack_node_loop(xmlNode * status, bool fence, pe_working_set_t * data_set) { bool changed = false; xmlNode *lrm_rsc = NULL; for (xmlNode *state = __xml_first_child_element(status); state != NULL; state = __xml_next_element(state)) { const char *id = NULL; const char *uname = NULL; node_t *this_node = NULL; bool process = FALSE; if (crm_str_eq((const char *)state->name, XML_CIB_TAG_STATE, TRUE) == FALSE) { continue; } id = crm_element_value(state, XML_ATTR_ID); uname = crm_element_value(state, XML_ATTR_UNAME); this_node = pe_find_node_any(data_set->nodes, id, uname); if (this_node == NULL) { crm_info("Node %s is unknown", id); continue; } else if (this_node->details->unpacked) { crm_info("Node %s is already processed", id); continue; } else if (!pe__is_guest_or_remote_node(this_node) && is_set(data_set->flags, pe_flag_stonith_enabled)) { // A redundant test, but preserves the order for regression tests process = TRUE; } else if (pe__is_guest_or_remote_node(this_node)) { bool check = FALSE; resource_t *rsc = this_node->details->remote_rsc; if(fence) { check = TRUE; } else if(rsc == NULL) { /* Not ready yet */ } else if (pe__is_guest_node(this_node) && rsc->role == RSC_ROLE_STARTED && rsc->container->role == RSC_ROLE_STARTED) { /* Both the connection and its containing resource need to be * known to be up before we process resources running in it. */ check = TRUE; crm_trace("Checking node %s/%s/%s status %d/%d/%d", id, rsc->id, rsc->container->id, fence, rsc->role, RSC_ROLE_STARTED); } else if (!pe__is_guest_node(this_node) && rsc->role == RSC_ROLE_STARTED) { check = TRUE; crm_trace("Checking node %s/%s status %d/%d/%d", id, rsc->id, fence, rsc->role, RSC_ROLE_STARTED); } if (check) { determine_remote_online_status(data_set, this_node); unpack_handle_remote_attrs(this_node, state, data_set); process = TRUE; } } else if (this_node->details->online) { process = TRUE; } else if (fence) { process = TRUE; } if(process) { crm_trace("Processing lrm resource entries on %shealthy%s node: %s", fence?"un":"", (pe__is_guest_or_remote_node(this_node)? " remote" : ""), this_node->details->uname); changed = TRUE; this_node->details->unpacked = TRUE; lrm_rsc = find_xml_node(state, XML_CIB_TAG_LRM, FALSE); lrm_rsc = find_xml_node(lrm_rsc, XML_LRM_TAG_RESOURCES, FALSE); unpack_lrm_resources(this_node, lrm_rsc, data_set); } } return changed; } /* remove nodes that are down, stopping */ /* create positive rsc_to_node constraints between resources and the nodes they are running on */ /* anything else? */ gboolean unpack_status(xmlNode * status, pe_working_set_t * data_set) { const char *id = NULL; const char *uname = NULL; xmlNode *state = NULL; node_t *this_node = NULL; crm_trace("Beginning unpack"); if (data_set->tickets == NULL) { data_set->tickets = g_hash_table_new_full(crm_str_hash, g_str_equal, free, destroy_ticket); } for (state = __xml_first_child_element(status); state != NULL; state = __xml_next_element(state)) { if (crm_str_eq((const char *)state->name, XML_CIB_TAG_TICKETS, TRUE)) { unpack_tickets_state((xmlNode *) state, data_set); } else if (crm_str_eq((const char *)state->name, XML_CIB_TAG_STATE, TRUE)) { xmlNode *attrs = NULL; const char *resource_discovery_enabled = NULL; id = crm_element_value(state, XML_ATTR_ID); uname = crm_element_value(state, XML_ATTR_UNAME); this_node = pe_find_node_any(data_set->nodes, id, uname); if (uname == NULL) { /* error */ continue; } else if (this_node == NULL) { crm_config_warn("Node %s in status section no longer exists", uname); continue; } else if (pe__is_guest_or_remote_node(this_node)) { /* online state for remote nodes is determined by the * rsc state after all the unpacking is done. we do however * need to mark whether or not the node has been fenced as this plays * a role during unpacking cluster node resource state */ this_node->details->remote_was_fenced = crm_atoi(crm_element_value(state, XML_NODE_IS_FENCED), "0"); continue; } crm_trace("Processing node id=%s, uname=%s", id, uname); /* Mark the node as provisionally clean * - at least we have seen it in the current cluster's lifetime */ this_node->details->unclean = FALSE; this_node->details->unseen = FALSE; attrs = find_xml_node(state, XML_TAG_TRANSIENT_NODEATTRS, FALSE); add_node_attrs(attrs, this_node, TRUE, data_set); if (crm_is_true(pe_node_attribute_raw(this_node, "standby"))) { crm_info("Node %s is in standby-mode", this_node->details->uname); this_node->details->standby = TRUE; } if (crm_is_true(pe_node_attribute_raw(this_node, "maintenance"))) { crm_info("Node %s is in maintenance-mode", this_node->details->uname); this_node->details->maintenance = TRUE; } resource_discovery_enabled = pe_node_attribute_raw(this_node, XML_NODE_ATTR_RSC_DISCOVERY); if (resource_discovery_enabled && !crm_is_true(resource_discovery_enabled)) { crm_warn("ignoring %s attribute on node %s, disabling resource discovery is not allowed on cluster nodes", XML_NODE_ATTR_RSC_DISCOVERY, this_node->details->uname); } crm_trace("determining node state"); determine_online_status(state, this_node, data_set); if (is_not_set(data_set->flags, pe_flag_have_quorum) && this_node->details->online && (data_set->no_quorum_policy == no_quorum_suicide)) { /* Everything else should flow from this automatically * At least until the PE becomes able to migrate off healthy resources */ pe_fence_node(data_set, this_node, "cluster does not have quorum"); } } } while(unpack_node_loop(status, FALSE, data_set)) { crm_trace("Start another loop"); } // Now catch any nodes we didn't see unpack_node_loop(status, is_set(data_set->flags, pe_flag_stonith_enabled), data_set); /* Now that we know where resources are, we can schedule stops of containers * with failed bundle connections */ if (data_set->stop_needed != NULL) { for (GList *item = data_set->stop_needed; item; item = item->next) { pe_resource_t *container = item->data; pe_node_t *node = pe__current_node(container); if (node) { stop_action(container, node, FALSE); } } g_list_free(data_set->stop_needed); data_set->stop_needed = NULL; } for (GListPtr gIter = data_set->nodes; gIter != NULL; gIter = gIter->next) { node_t *this_node = gIter->data; if (this_node == NULL) { continue; } else if (!pe__is_guest_or_remote_node(this_node)) { continue; } else if(this_node->details->unpacked) { continue; } determine_remote_online_status(data_set, this_node); } return TRUE; } static gboolean determine_online_status_no_fencing(pe_working_set_t * data_set, xmlNode * node_state, node_t * this_node) { gboolean online = FALSE; const char *join = crm_element_value(node_state, XML_NODE_JOIN_STATE); const char *is_peer = crm_element_value(node_state, XML_NODE_IS_PEER); const char *in_cluster = crm_element_value(node_state, XML_NODE_IN_CLUSTER); const char *exp_state = crm_element_value(node_state, XML_NODE_EXPECTED); if (!crm_is_true(in_cluster)) { crm_trace("Node is down: in_cluster=%s", crm_str(in_cluster)); } else if (safe_str_eq(is_peer, ONLINESTATUS)) { if (safe_str_eq(join, CRMD_JOINSTATE_MEMBER)) { online = TRUE; } else { crm_debug("Node is not ready to run resources: %s", join); } } else if (this_node->details->expected_up == FALSE) { crm_trace("Controller is down: in_cluster=%s", crm_str(in_cluster)); crm_trace("\tis_peer=%s, join=%s, expected=%s", crm_str(is_peer), crm_str(join), crm_str(exp_state)); } else { /* mark it unclean */ pe_fence_node(data_set, this_node, "peer is unexpectedly down"); crm_info("\tin_cluster=%s, is_peer=%s, join=%s, expected=%s", crm_str(in_cluster), crm_str(is_peer), crm_str(join), crm_str(exp_state)); } return online; } static gboolean determine_online_status_fencing(pe_working_set_t * data_set, xmlNode * node_state, node_t * this_node) { gboolean online = FALSE; gboolean do_terminate = FALSE; bool crmd_online = FALSE; const char *join = crm_element_value(node_state, XML_NODE_JOIN_STATE); const char *is_peer = crm_element_value(node_state, XML_NODE_IS_PEER); const char *in_cluster = crm_element_value(node_state, XML_NODE_IN_CLUSTER); const char *exp_state = crm_element_value(node_state, XML_NODE_EXPECTED); const char *terminate = pe_node_attribute_raw(this_node, "terminate"); /* - XML_NODE_IN_CLUSTER ::= true|false - XML_NODE_IS_PEER ::= online|offline - XML_NODE_JOIN_STATE ::= member|down|pending|banned - XML_NODE_EXPECTED ::= member|down */ if (crm_is_true(terminate)) { do_terminate = TRUE; } else if (terminate != NULL && strlen(terminate) > 0) { /* could be a time() value */ char t = terminate[0]; if (t != '0' && isdigit(t)) { do_terminate = TRUE; } } crm_trace("%s: in_cluster=%s, is_peer=%s, join=%s, expected=%s, term=%d", this_node->details->uname, crm_str(in_cluster), crm_str(is_peer), crm_str(join), crm_str(exp_state), do_terminate); online = crm_is_true(in_cluster); crmd_online = safe_str_eq(is_peer, ONLINESTATUS); if (exp_state == NULL) { exp_state = CRMD_JOINSTATE_DOWN; } if (this_node->details->shutdown) { crm_debug("%s is shutting down", this_node->details->uname); /* Slightly different criteria since we can't shut down a dead peer */ online = crmd_online; } else if (in_cluster == NULL) { pe_fence_node(data_set, this_node, "peer has not been seen by the cluster"); } else if (safe_str_eq(join, CRMD_JOINSTATE_NACK)) { pe_fence_node(data_set, this_node, "peer failed the pacemaker membership criteria"); } else if (do_terminate == FALSE && safe_str_eq(exp_state, CRMD_JOINSTATE_DOWN)) { if (crm_is_true(in_cluster) || crmd_online) { crm_info("- Node %s is not ready to run resources", this_node->details->uname); this_node->details->standby = TRUE; this_node->details->pending = TRUE; } else { crm_trace("%s is down or still coming up", this_node->details->uname); } } else if (do_terminate && safe_str_eq(join, CRMD_JOINSTATE_DOWN) && crm_is_true(in_cluster) == FALSE && !crmd_online) { crm_info("Node %s was just shot", this_node->details->uname); online = FALSE; } else if (crm_is_true(in_cluster) == FALSE) { pe_fence_node(data_set, this_node, "peer is no longer part of the cluster"); } else if (!crmd_online) { pe_fence_node(data_set, this_node, "peer process is no longer available"); /* Everything is running at this point, now check join state */ } else if (do_terminate) { pe_fence_node(data_set, this_node, "termination was requested"); } else if (safe_str_eq(join, CRMD_JOINSTATE_MEMBER)) { crm_info("Node %s is active", this_node->details->uname); } else if (safe_str_eq(join, CRMD_JOINSTATE_PENDING) || safe_str_eq(join, CRMD_JOINSTATE_DOWN)) { crm_info("Node %s is not ready to run resources", this_node->details->uname); this_node->details->standby = TRUE; this_node->details->pending = TRUE; } else { pe_fence_node(data_set, this_node, "peer was in an unknown state"); crm_warn("%s: in-cluster=%s, is-peer=%s, join=%s, expected=%s, term=%d, shutdown=%d", this_node->details->uname, crm_str(in_cluster), crm_str(is_peer), crm_str(join), crm_str(exp_state), do_terminate, this_node->details->shutdown); } return online; } static gboolean determine_remote_online_status(pe_working_set_t * data_set, node_t * this_node) { resource_t *rsc = this_node->details->remote_rsc; resource_t *container = NULL; pe_node_t *host = NULL; /* If there is a node state entry for a (former) Pacemaker Remote node * but no resource creating that node, the node's connection resource will * be NULL. Consider it an offline remote node in that case. */ if (rsc == NULL) { this_node->details->online = FALSE; goto remote_online_done; } container = rsc->container; if (container && (g_list_length(rsc->running_on) == 1)) { host = rsc->running_on->data; } /* If the resource is currently started, mark it online. */ if (rsc->role == RSC_ROLE_STARTED) { crm_trace("%s node %s presumed ONLINE because connection resource is started", (container? "Guest" : "Remote"), this_node->details->id); this_node->details->online = TRUE; } /* consider this node shutting down if transitioning start->stop */ if (rsc->role == RSC_ROLE_STARTED && rsc->next_role == RSC_ROLE_STOPPED) { crm_trace("%s node %s shutting down because connection resource is stopping", (container? "Guest" : "Remote"), this_node->details->id); this_node->details->shutdown = TRUE; } /* Now check all the failure conditions. */ if(container && is_set(container->flags, pe_rsc_failed)) { crm_trace("Guest node %s UNCLEAN because guest resource failed", this_node->details->id); this_node->details->online = FALSE; this_node->details->remote_requires_reset = TRUE; } else if(is_set(rsc->flags, pe_rsc_failed)) { crm_trace("%s node %s OFFLINE because connection resource failed", (container? "Guest" : "Remote"), this_node->details->id); this_node->details->online = FALSE; } else if (rsc->role == RSC_ROLE_STOPPED || (container && container->role == RSC_ROLE_STOPPED)) { crm_trace("%s node %s OFFLINE because its resource is stopped", (container? "Guest" : "Remote"), this_node->details->id); this_node->details->online = FALSE; this_node->details->remote_requires_reset = FALSE; } else if (host && (host->details->online == FALSE) && host->details->unclean) { crm_trace("Guest node %s UNCLEAN because host is unclean", this_node->details->id); this_node->details->online = FALSE; this_node->details->remote_requires_reset = TRUE; } remote_online_done: crm_trace("Remote node %s online=%s", this_node->details->id, this_node->details->online ? "TRUE" : "FALSE"); return this_node->details->online; } gboolean determine_online_status(xmlNode * node_state, node_t * this_node, pe_working_set_t * data_set) { gboolean online = FALSE; const char *exp_state = crm_element_value(node_state, XML_NODE_EXPECTED); if (this_node == NULL) { crm_config_err("No node to check"); return online; } this_node->details->shutdown = FALSE; this_node->details->expected_up = FALSE; if (pe__shutdown_requested(this_node)) { this_node->details->shutdown = TRUE; } else if (safe_str_eq(exp_state, CRMD_JOINSTATE_MEMBER)) { this_node->details->expected_up = TRUE; } if (this_node->details->type == node_ping) { this_node->details->unclean = FALSE; online = FALSE; /* As far as resource management is concerned, * the node is safely offline. * Anyone caught abusing this logic will be shot */ } else if (is_set(data_set->flags, pe_flag_stonith_enabled) == FALSE) { online = determine_online_status_no_fencing(data_set, node_state, this_node); } else { online = determine_online_status_fencing(data_set, node_state, this_node); } if (online) { this_node->details->online = TRUE; } else { /* remove node from contention */ this_node->fixed = TRUE; this_node->weight = -INFINITY; } if (online && this_node->details->shutdown) { /* don't run resources here */ this_node->fixed = TRUE; this_node->weight = -INFINITY; } if (this_node->details->type == node_ping) { crm_info("Node %s is not a pacemaker node", this_node->details->uname); } else if (this_node->details->unclean) { pe_proc_warn("Node %s is unclean", this_node->details->uname); } else if (this_node->details->online) { crm_info("Node %s is %s", this_node->details->uname, this_node->details->shutdown ? "shutting down" : this_node->details->pending ? "pending" : this_node->details->standby ? "standby" : this_node->details->maintenance ? "maintenance" : "online"); } else { crm_trace("Node %s is offline", this_node->details->uname); } return online; } /*! * \internal * \brief Find the end of a resource's name, excluding any clone suffix * * \param[in] id Resource ID to check * * \return Pointer to last character of resource's base name */ const char * pe_base_name_end(const char *id) { if (!crm_strlen_zero(id)) { const char *end = id + strlen(id) - 1; for (const char *s = end; s > id; --s) { switch (*s) { case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': case '8': case '9': break; case ':': return (s == end)? s : (s - 1); default: return end; } } return end; } return NULL; } /*! * \internal * \brief Get a resource name excluding any clone suffix * * \param[in] last_rsc_id Resource ID to check * * \return Pointer to newly allocated string with resource's base name * \note It is the caller's responsibility to free() the result. * This asserts on error, so callers can assume result is not NULL. */ char * clone_strip(const char *last_rsc_id) { const char *end = pe_base_name_end(last_rsc_id); char *basename = NULL; CRM_ASSERT(end); basename = strndup(last_rsc_id, end - last_rsc_id + 1); CRM_ASSERT(basename); return basename; } /*! * \internal * \brief Get the name of the first instance of a cloned resource * * \param[in] last_rsc_id Resource ID to check * * \return Pointer to newly allocated string with resource's base name plus :0 * \note It is the caller's responsibility to free() the result. * This asserts on error, so callers can assume result is not NULL. */ char * clone_zero(const char *last_rsc_id) { const char *end = pe_base_name_end(last_rsc_id); size_t base_name_len = end - last_rsc_id + 1; char *zero = NULL; CRM_ASSERT(end); zero = calloc(base_name_len + 3, sizeof(char)); CRM_ASSERT(zero); memcpy(zero, last_rsc_id, base_name_len); zero[base_name_len] = ':'; zero[base_name_len + 1] = '0'; return zero; } static resource_t * create_fake_resource(const char *rsc_id, xmlNode * rsc_entry, pe_working_set_t * data_set) { resource_t *rsc = NULL; xmlNode *xml_rsc = create_xml_node(NULL, XML_CIB_TAG_RESOURCE); copy_in_properties(xml_rsc, rsc_entry); crm_xml_add(xml_rsc, XML_ATTR_ID, rsc_id); crm_log_xml_debug(xml_rsc, "Orphan resource"); if (!common_unpack(xml_rsc, &rsc, NULL, data_set)) { return NULL; } if (xml_contains_remote_node(xml_rsc)) { node_t *node; crm_debug("Detected orphaned remote node %s", rsc_id); node = pe_find_node(data_set->nodes, rsc_id); if (node == NULL) { node = pe_create_node(rsc_id, rsc_id, "remote", NULL, data_set); } link_rsc2remotenode(data_set, rsc); if (node) { crm_trace("Setting node %s as shutting down due to orphaned connection resource", rsc_id); node->details->shutdown = TRUE; } } if (crm_element_value(rsc_entry, XML_RSC_ATTR_CONTAINER)) { /* This orphaned rsc needs to be mapped to a container. */ crm_trace("Detected orphaned container filler %s", rsc_id); set_bit(rsc->flags, pe_rsc_orphan_container_filler); } set_bit(rsc->flags, pe_rsc_orphan); data_set->resources = g_list_append(data_set->resources, rsc); return rsc; } /*! * \internal * \brief Create orphan instance for anonymous clone resource history */ static pe_resource_t * create_anonymous_orphan(pe_resource_t *parent, const char *rsc_id, pe_node_t *node, pe_working_set_t *data_set) { pe_resource_t *top = pe__create_clone_child(parent, data_set); // find_rsc() because we might be a cloned group pe_resource_t *orphan = top->fns->find_rsc(top, rsc_id, NULL, pe_find_clone); pe_rsc_debug(parent, "Created orphan %s for %s: %s on %s", top->id, parent->id, rsc_id, node->details->uname); return orphan; } /*! * \internal * \brief Check a node for an instance of an anonymous clone * * Return a child instance of the specified anonymous clone, in order of * preference: (1) the instance running on the specified node, if any; * (2) an inactive instance (i.e. within the total of clone-max instances); * (3) a newly created orphan (i.e. clone-max instances are already active). * * \param[in] data_set Cluster information * \param[in] node Node on which to check for instance * \param[in] parent Clone to check * \param[in] rsc_id Name of cloned resource in history (without instance) */ static resource_t * find_anonymous_clone(pe_working_set_t * data_set, node_t * node, resource_t * parent, const char *rsc_id) { GListPtr rIter = NULL; pe_resource_t *rsc = NULL; pe_resource_t *inactive_instance = NULL; gboolean skip_inactive = FALSE; CRM_ASSERT(parent != NULL); CRM_ASSERT(pe_rsc_is_clone(parent)); CRM_ASSERT(is_not_set(parent->flags, pe_rsc_unique)); // Check for active (or partially active, for cloned groups) instance pe_rsc_trace(parent, "Looking for %s on %s in %s", rsc_id, node->details->uname, parent->id); for (rIter = parent->children; rsc == NULL && rIter; rIter = rIter->next) { GListPtr locations = NULL; resource_t *child = rIter->data; /* Check whether this instance is already known to be active or pending * anywhere, at this stage of unpacking. Because this function is called * for a resource before the resource's individual operation history * entries are unpacked, locations will generally not contain the * desired node. * * However, there are three exceptions: * (1) when child is a cloned group and we have already unpacked the * history of another member of the group on the same node; * (2) when we've already unpacked the history of another numbered * instance on the same node (which can happen if globally-unique * was flipped from true to false); and * (3) when we re-run calculations on the same data set as part of a * simulation. */ child->fns->location(child, &locations, 2); if (locations) { /* We should never associate the same numbered anonymous clone * instance with multiple nodes, and clone instances can't migrate, * so there must be only one location, regardless of history. */ CRM_LOG_ASSERT(locations->next == NULL); if (((pe_node_t *)locations->data)->details == node->details) { /* This child instance is active on the requested node, so check * for a corresponding configured resource. We use find_rsc() * instead of child because child may be a cloned group, and we * need the particular member corresponding to rsc_id. * * If the history entry is orphaned, rsc will be NULL. */ rsc = parent->fns->find_rsc(child, rsc_id, NULL, pe_find_clone); if (rsc) { /* If there are multiple instance history entries for an * anonymous clone in a single node's history (which can * happen if globally-unique is switched from true to * false), we want to consider the instances beyond the * first as orphans, even if there are inactive instance * numbers available. */ if (rsc->running_on) { crm_notice("Active (now-)anonymous clone %s has " "multiple (orphan) instance histories on %s", parent->id, node->details->uname); skip_inactive = TRUE; rsc = NULL; } else { pe_rsc_trace(parent, "Resource %s, active", rsc->id); } } } g_list_free(locations); } else { pe_rsc_trace(parent, "Resource %s, skip inactive", child->id); if (!skip_inactive && !inactive_instance && is_not_set(child->flags, pe_rsc_block)) { // Remember one inactive instance in case we don't find active inactive_instance = parent->fns->find_rsc(child, rsc_id, NULL, pe_find_clone); /* ... but don't use it if it was already associated with a * pending action on another node */ if (inactive_instance && inactive_instance->pending_node && (inactive_instance->pending_node->details != node->details)) { inactive_instance = NULL; } } } } if ((rsc == NULL) && !skip_inactive && (inactive_instance != NULL)) { pe_rsc_trace(parent, "Resource %s, empty slot", inactive_instance->id); rsc = inactive_instance; } /* If the resource has "requires" set to "quorum" or "nothing", and we don't * have a clone instance for every node, we don't want to consume a valid * instance number for unclean nodes. Such instances may appear to be active * according to the history, but should be considered inactive, so we can * start an instance elsewhere. Treat such instances as orphans. * * An exception is instances running on guest nodes -- since guest node * "fencing" is actually just a resource stop, requires shouldn't apply. * * @TODO Ideally, we'd use an inactive instance number if it is not needed * for any clean instances. However, we don't know that at this point. */ if ((rsc != NULL) && is_not_set(rsc->flags, pe_rsc_needs_fencing) && (!node->details->online || node->details->unclean) && !pe__is_guest_node(node) && !pe__is_universal_clone(parent, data_set)) { rsc = NULL; } if (rsc == NULL) { rsc = create_anonymous_orphan(parent, rsc_id, node, data_set); pe_rsc_trace(parent, "Resource %s, orphan", rsc->id); } return rsc; } static resource_t * unpack_find_resource(pe_working_set_t * data_set, node_t * node, const char *rsc_id, xmlNode * rsc_entry) { resource_t *rsc = NULL; resource_t *parent = NULL; crm_trace("looking for %s", rsc_id); rsc = pe_find_resource(data_set->resources, rsc_id); if (rsc == NULL) { /* If we didn't find the resource by its name in the operation history, * check it again as a clone instance. Even when clone-max=0, we create * a single :0 orphan to match against here. */ char *clone0_id = clone_zero(rsc_id); resource_t *clone0 = pe_find_resource(data_set->resources, clone0_id); if (clone0 && is_not_set(clone0->flags, pe_rsc_unique)) { rsc = clone0; parent = uber_parent(clone0); crm_trace("%s found as %s (%s)", rsc_id, clone0_id, parent->id); } else { crm_trace("%s is not known as %s either (orphan)", rsc_id, clone0_id); } free(clone0_id); } else if (rsc->variant > pe_native) { crm_trace("Resource history for %s is orphaned because it is no longer primitive", rsc_id); return NULL; } else { parent = uber_parent(rsc); } if (pe_rsc_is_anon_clone(parent)) { if (pe_rsc_is_bundled(parent)) { rsc = pe__find_bundle_replica(parent->parent, node); } else { char *base = clone_strip(rsc_id); rsc = find_anonymous_clone(data_set, node, parent, base); free(base); CRM_ASSERT(rsc != NULL); } } if (rsc && safe_str_neq(rsc_id, rsc->id) && safe_str_neq(rsc_id, rsc->clone_name)) { free(rsc->clone_name); rsc->clone_name = strdup(rsc_id); pe_rsc_debug(rsc, "Internally renamed %s on %s to %s%s", rsc_id, node->details->uname, rsc->id, (is_set(rsc->flags, pe_rsc_orphan)? " (ORPHAN)" : "")); } return rsc; } static resource_t * process_orphan_resource(xmlNode * rsc_entry, node_t * node, pe_working_set_t * data_set) { resource_t *rsc = NULL; const char *rsc_id = crm_element_value(rsc_entry, XML_ATTR_ID); crm_debug("Detected orphan resource %s on %s", rsc_id, node->details->uname); rsc = create_fake_resource(rsc_id, rsc_entry, data_set); if (is_set(data_set->flags, pe_flag_stop_rsc_orphans) == FALSE) { clear_bit(rsc->flags, pe_rsc_managed); } else { - print_resource(LOG_TRACE, "Added orphan", rsc, FALSE); - CRM_CHECK(rsc != NULL, return NULL); + pe_rsc_trace(rsc, "Added orphan %s", rsc->id); resource_location(rsc, NULL, -INFINITY, "__orphan_do_not_run__", data_set); } return rsc; } static void process_rsc_state(resource_t * rsc, node_t * node, enum action_fail_response on_fail, xmlNode * migrate_op, pe_working_set_t * data_set) { node_t *tmpnode = NULL; char *reason = NULL; CRM_ASSERT(rsc); pe_rsc_trace(rsc, "Resource %s is %s on %s: on_fail=%s", rsc->id, role2text(rsc->role), node->details->uname, fail2text(on_fail)); /* process current state */ if (rsc->role != RSC_ROLE_UNKNOWN) { resource_t *iter = rsc; while (iter) { if (g_hash_table_lookup(iter->known_on, node->details->id) == NULL) { node_t *n = node_copy(node); pe_rsc_trace(rsc, "%s (aka. %s) known on %s", rsc->id, rsc->clone_name, n->details->uname); g_hash_table_insert(iter->known_on, (gpointer) n->details->id, n); } if (is_set(iter->flags, pe_rsc_unique)) { break; } iter = iter->parent; } } /* If a managed resource is believed to be running, but node is down ... */ if (rsc->role > RSC_ROLE_STOPPED && node->details->online == FALSE && node->details->maintenance == FALSE && is_set(rsc->flags, pe_rsc_managed)) { gboolean should_fence = FALSE; /* If this is a guest node, fence it (regardless of whether fencing is * enabled, because guest node fencing is done by recovery of the * container resource rather than by the fencer). Mark the resource * we're processing as failed. When the guest comes back up, its * operation history in the CIB will be cleared, freeing the affected * resource to run again once we are sure we know its state. */ if (pe__is_guest_node(node)) { set_bit(rsc->flags, pe_rsc_failed); should_fence = TRUE; } else if (is_set(data_set->flags, pe_flag_stonith_enabled)) { if (pe__is_remote_node(node) && node->details->remote_rsc && is_not_set(node->details->remote_rsc->flags, pe_rsc_failed)) { /* Setting unseen means that fencing of the remote node will * occur only if the connection resource is not going to start * somewhere. This allows connection resources on a failed * cluster node to move to another node without requiring the * remote nodes to be fenced as well. */ node->details->unseen = TRUE; reason = crm_strdup_printf("%s is active there (fencing will be" " revoked if remote connection can " "be re-established elsewhere)", rsc->id); } should_fence = TRUE; } if (should_fence) { if (reason == NULL) { reason = crm_strdup_printf("%s is thought to be active there", rsc->id); } pe_fence_node(data_set, node, reason); } free(reason); } if (node->details->unclean) { /* No extra processing needed * Also allows resources to be started again after a node is shot */ on_fail = action_fail_ignore; } switch (on_fail) { case action_fail_ignore: /* nothing to do */ break; case action_fail_fence: /* treat it as if it is still running * but also mark the node as unclean */ reason = crm_strdup_printf("%s failed there", rsc->id); pe_fence_node(data_set, node, reason); free(reason); break; case action_fail_standby: node->details->standby = TRUE; node->details->standby_onfail = TRUE; break; case action_fail_block: /* is_managed == FALSE will prevent any * actions being sent for the resource */ clear_bit(rsc->flags, pe_rsc_managed); set_bit(rsc->flags, pe_rsc_block); break; case action_fail_migrate: /* make sure it comes up somewhere else * or not at all */ resource_location(rsc, node, -INFINITY, "__action_migration_auto__", data_set); break; case action_fail_stop: rsc->next_role = RSC_ROLE_STOPPED; break; case action_fail_recover: if (rsc->role != RSC_ROLE_STOPPED && rsc->role != RSC_ROLE_UNKNOWN) { set_bit(rsc->flags, pe_rsc_failed); stop_action(rsc, node, FALSE); } break; case action_fail_restart_container: set_bit(rsc->flags, pe_rsc_failed); if (rsc->container && pe_rsc_is_bundled(rsc)) { /* A bundle's remote connection can run on a different node than * the bundle's container. We don't necessarily know where the * container is running yet, so remember it and add a stop * action for it later. */ data_set->stop_needed = g_list_prepend(data_set->stop_needed, rsc->container); } else if (rsc->container) { stop_action(rsc->container, node, FALSE); } else if (rsc->role != RSC_ROLE_STOPPED && rsc->role != RSC_ROLE_UNKNOWN) { stop_action(rsc, node, FALSE); } break; case action_fail_reset_remote: set_bit(rsc->flags, pe_rsc_failed); if (is_set(data_set->flags, pe_flag_stonith_enabled)) { tmpnode = NULL; if (rsc->is_remote_node) { tmpnode = pe_find_node(data_set->nodes, rsc->id); } if (tmpnode && pe__is_remote_node(tmpnode) && tmpnode->details->remote_was_fenced == 0) { /* The remote connection resource failed in a way that * should result in fencing the remote node. */ pe_fence_node(data_set, tmpnode, "remote connection is unrecoverable"); } } /* require the stop action regardless if fencing is occurring or not. */ if (rsc->role > RSC_ROLE_STOPPED) { stop_action(rsc, node, FALSE); } /* if reconnect delay is in use, prevent the connection from exiting the * "STOPPED" role until the failure is cleared by the delay timeout. */ if (rsc->remote_reconnect_ms) { rsc->next_role = RSC_ROLE_STOPPED; } break; } /* ensure a remote-node connection failure forces an unclean remote-node * to be fenced. By setting unseen = FALSE, the remote-node failure will * result in a fencing operation regardless if we're going to attempt to * reconnect to the remote-node in this transition or not. */ if (is_set(rsc->flags, pe_rsc_failed) && rsc->is_remote_node) { tmpnode = pe_find_node(data_set->nodes, rsc->id); if (tmpnode && tmpnode->details->unclean) { tmpnode->details->unseen = FALSE; } } if (rsc->role != RSC_ROLE_STOPPED && rsc->role != RSC_ROLE_UNKNOWN) { if (is_set(rsc->flags, pe_rsc_orphan)) { if (is_set(rsc->flags, pe_rsc_managed)) { crm_config_warn("Detected active orphan %s running on %s", rsc->id, node->details->uname); } else { crm_config_warn("Cluster configured not to stop active orphans." " %s must be stopped manually on %s", rsc->id, node->details->uname); } } native_add_running(rsc, node, data_set); if (on_fail != action_fail_ignore) { set_bit(rsc->flags, pe_rsc_failed); } } else if (rsc->clone_name && strchr(rsc->clone_name, ':') != NULL) { /* Only do this for older status sections that included instance numbers * Otherwise stopped instances will appear as orphans */ pe_rsc_trace(rsc, "Resetting clone_name %s for %s (stopped)", rsc->clone_name, rsc->id); free(rsc->clone_name); rsc->clone_name = NULL; } else { GList *possible_matches = pe__resource_actions(rsc, node, RSC_STOP, FALSE); GListPtr gIter = possible_matches; for (; gIter != NULL; gIter = gIter->next) { action_t *stop = (action_t *) gIter->data; stop->flags |= pe_action_optional; } g_list_free(possible_matches); } } /* create active recurring operations as optional */ static void process_recurring(node_t * node, resource_t * rsc, int start_index, int stop_index, GListPtr sorted_op_list, pe_working_set_t * data_set) { int counter = -1; const char *task = NULL; const char *status = NULL; GListPtr gIter = sorted_op_list; CRM_ASSERT(rsc); pe_rsc_trace(rsc, "%s: Start index %d, stop index = %d", rsc->id, start_index, stop_index); for (; gIter != NULL; gIter = gIter->next) { xmlNode *rsc_op = (xmlNode *) gIter->data; guint interval_ms = 0; char *key = NULL; const char *id = ID(rsc_op); const char *interval_ms_s = NULL; counter++; if (node->details->online == FALSE) { pe_rsc_trace(rsc, "Skipping %s/%s: node is offline", rsc->id, node->details->uname); break; /* Need to check if there's a monitor for role="Stopped" */ } else if (start_index < stop_index && counter <= stop_index) { pe_rsc_trace(rsc, "Skipping %s/%s: resource is not active", id, node->details->uname); continue; } else if (counter < start_index) { pe_rsc_trace(rsc, "Skipping %s/%s: old %d", id, node->details->uname, counter); continue; } interval_ms_s = crm_element_value(rsc_op, XML_LRM_ATTR_INTERVAL_MS); interval_ms = crm_parse_ms(interval_ms_s); if (interval_ms == 0) { pe_rsc_trace(rsc, "Skipping %s/%s: non-recurring", id, node->details->uname); continue; } status = crm_element_value(rsc_op, XML_LRM_ATTR_OPSTATUS); if (safe_str_eq(status, "-1")) { pe_rsc_trace(rsc, "Skipping %s/%s: status", id, node->details->uname); continue; } task = crm_element_value(rsc_op, XML_LRM_ATTR_TASK); /* create the action */ key = generate_op_key(rsc->id, task, interval_ms); pe_rsc_trace(rsc, "Creating %s/%s", key, node->details->uname); custom_action(rsc, key, task, node, TRUE, TRUE, data_set); } } void calculate_active_ops(GListPtr sorted_op_list, int *start_index, int *stop_index) { int counter = -1; int implied_monitor_start = -1; int implied_clone_start = -1; const char *task = NULL; const char *status = NULL; GListPtr gIter = sorted_op_list; *stop_index = -1; *start_index = -1; for (; gIter != NULL; gIter = gIter->next) { xmlNode *rsc_op = (xmlNode *) gIter->data; counter++; task = crm_element_value(rsc_op, XML_LRM_ATTR_TASK); status = crm_element_value(rsc_op, XML_LRM_ATTR_OPSTATUS); if (safe_str_eq(task, CRMD_ACTION_STOP) && safe_str_eq(status, "0")) { *stop_index = counter; } else if (safe_str_eq(task, CRMD_ACTION_START) || safe_str_eq(task, CRMD_ACTION_MIGRATED)) { *start_index = counter; } else if ((implied_monitor_start <= *stop_index) && safe_str_eq(task, CRMD_ACTION_STATUS)) { const char *rc = crm_element_value(rsc_op, XML_LRM_ATTR_RC); if (safe_str_eq(rc, "0") || safe_str_eq(rc, "8")) { implied_monitor_start = counter; } } else if (safe_str_eq(task, CRMD_ACTION_PROMOTE) || safe_str_eq(task, CRMD_ACTION_DEMOTE)) { implied_clone_start = counter; } } if (*start_index == -1) { if (implied_clone_start != -1) { *start_index = implied_clone_start; } else if (implied_monitor_start != -1) { *start_index = implied_monitor_start; } } } static resource_t * unpack_lrm_rsc_state(node_t * node, xmlNode * rsc_entry, pe_working_set_t * data_set) { GListPtr gIter = NULL; int stop_index = -1; int start_index = -1; enum rsc_role_e req_role = RSC_ROLE_UNKNOWN; const char *task = NULL; const char *rsc_id = crm_element_value(rsc_entry, XML_ATTR_ID); resource_t *rsc = NULL; GListPtr op_list = NULL; GListPtr sorted_op_list = NULL; xmlNode *migrate_op = NULL; xmlNode *rsc_op = NULL; xmlNode *last_failure = NULL; enum action_fail_response on_fail = FALSE; enum rsc_role_e saved_role = RSC_ROLE_UNKNOWN; crm_trace("[%s] Processing %s on %s", crm_element_name(rsc_entry), rsc_id, node->details->uname); /* extract operations */ op_list = NULL; sorted_op_list = NULL; for (rsc_op = __xml_first_child_element(rsc_entry); rsc_op != NULL; rsc_op = __xml_next_element(rsc_op)) { if (crm_str_eq((const char *)rsc_op->name, XML_LRM_TAG_RSC_OP, TRUE)) { op_list = g_list_prepend(op_list, rsc_op); } } if (op_list == NULL) { /* if there are no operations, there is nothing to do */ return NULL; } /* find the resource */ rsc = unpack_find_resource(data_set, node, rsc_id, rsc_entry); if (rsc == NULL) { rsc = process_orphan_resource(rsc_entry, node, data_set); } CRM_ASSERT(rsc != NULL); /* process operations */ saved_role = rsc->role; on_fail = action_fail_ignore; rsc->role = RSC_ROLE_UNKNOWN; sorted_op_list = g_list_sort(op_list, sort_op_by_callid); for (gIter = sorted_op_list; gIter != NULL; gIter = gIter->next) { xmlNode *rsc_op = (xmlNode *) gIter->data; task = crm_element_value(rsc_op, XML_LRM_ATTR_TASK); if (safe_str_eq(task, CRMD_ACTION_MIGRATED)) { migrate_op = rsc_op; } unpack_rsc_op(rsc, node, rsc_op, &last_failure, &on_fail, data_set); } /* create active recurring operations as optional */ calculate_active_ops(sorted_op_list, &start_index, &stop_index); process_recurring(node, rsc, start_index, stop_index, sorted_op_list, data_set); /* no need to free the contents */ g_list_free(sorted_op_list); process_rsc_state(rsc, node, on_fail, migrate_op, data_set); if (get_target_role(rsc, &req_role)) { if (rsc->next_role == RSC_ROLE_UNKNOWN || req_role < rsc->next_role) { pe_rsc_debug(rsc, "%s: Overwriting calculated next role %s" " with requested next role %s", rsc->id, role2text(rsc->next_role), role2text(req_role)); rsc->next_role = req_role; } else if (req_role > rsc->next_role) { pe_rsc_info(rsc, "%s: Not overwriting calculated next role %s" " with requested next role %s", rsc->id, role2text(rsc->next_role), role2text(req_role)); } } if (saved_role > rsc->role) { rsc->role = saved_role; } return rsc; } static void handle_orphaned_container_fillers(xmlNode * lrm_rsc_list, pe_working_set_t * data_set) { xmlNode *rsc_entry = NULL; for (rsc_entry = __xml_first_child_element(lrm_rsc_list); rsc_entry != NULL; rsc_entry = __xml_next_element(rsc_entry)) { resource_t *rsc; resource_t *container; const char *rsc_id; const char *container_id; if (safe_str_neq((const char *)rsc_entry->name, XML_LRM_TAG_RESOURCE)) { continue; } container_id = crm_element_value(rsc_entry, XML_RSC_ATTR_CONTAINER); rsc_id = crm_element_value(rsc_entry, XML_ATTR_ID); if (container_id == NULL || rsc_id == NULL) { continue; } container = pe_find_resource(data_set->resources, container_id); if (container == NULL) { continue; } rsc = pe_find_resource(data_set->resources, rsc_id); if (rsc == NULL || is_set(rsc->flags, pe_rsc_orphan_container_filler) == FALSE || rsc->container != NULL) { continue; } pe_rsc_trace(rsc, "Mapped container of orphaned resource %s to %s", rsc->id, container_id); rsc->container = container; container->fillers = g_list_append(container->fillers, rsc); } } gboolean unpack_lrm_resources(node_t * node, xmlNode * lrm_rsc_list, pe_working_set_t * data_set) { xmlNode *rsc_entry = NULL; gboolean found_orphaned_container_filler = FALSE; CRM_CHECK(node != NULL, return FALSE); crm_trace("Unpacking resources on %s", node->details->uname); for (rsc_entry = __xml_first_child_element(lrm_rsc_list); rsc_entry != NULL; rsc_entry = __xml_next_element(rsc_entry)) { if (crm_str_eq((const char *)rsc_entry->name, XML_LRM_TAG_RESOURCE, TRUE)) { resource_t *rsc = unpack_lrm_rsc_state(node, rsc_entry, data_set); if (!rsc) { continue; } if (is_set(rsc->flags, pe_rsc_orphan_container_filler)) { found_orphaned_container_filler = TRUE; } } } /* now that all the resource state has been unpacked for this node * we have to go back and map any orphaned container fillers to their * container resource */ if (found_orphaned_container_filler) { handle_orphaned_container_fillers(lrm_rsc_list, data_set); } return TRUE; } static void set_active(resource_t * rsc) { resource_t *top = uber_parent(rsc); if (top && is_set(top->flags, pe_rsc_promotable)) { rsc->role = RSC_ROLE_SLAVE; } else { rsc->role = RSC_ROLE_STARTED; } } static void set_node_score(gpointer key, gpointer value, gpointer user_data) { node_t *node = value; int *score = user_data; node->weight = *score; } #define STATUS_PATH_MAX 1024 static xmlNode * find_lrm_op(const char *resource, const char *op, const char *node, const char *source, bool success_only, pe_working_set_t *data_set) { int offset = 0; char xpath[STATUS_PATH_MAX]; xmlNode *xml = NULL; offset += snprintf(xpath + offset, STATUS_PATH_MAX - offset, "//node_state[@uname='%s']", node); offset += snprintf(xpath + offset, STATUS_PATH_MAX - offset, "//" XML_LRM_TAG_RESOURCE "[@id='%s']", resource); /* Need to check against transition_magic too? */ if (source && safe_str_eq(op, CRMD_ACTION_MIGRATE)) { offset += snprintf(xpath + offset, STATUS_PATH_MAX - offset, "/" XML_LRM_TAG_RSC_OP "[@operation='%s' and @migrate_target='%s']", op, source); } else if (source && safe_str_eq(op, CRMD_ACTION_MIGRATED)) { offset += snprintf(xpath + offset, STATUS_PATH_MAX - offset, "/" XML_LRM_TAG_RSC_OP "[@operation='%s' and @migrate_source='%s']", op, source); } else { offset += snprintf(xpath + offset, STATUS_PATH_MAX - offset, "/" XML_LRM_TAG_RSC_OP "[@operation='%s']", op); } CRM_LOG_ASSERT(offset > 0); xml = get_xpath_object(xpath, data_set->input, LOG_DEBUG); if (xml && success_only) { int rc = PCMK_OCF_UNKNOWN_ERROR; int status = PCMK_LRM_OP_ERROR; crm_element_value_int(xml, XML_LRM_ATTR_RC, &rc); crm_element_value_int(xml, XML_LRM_ATTR_OPSTATUS, &status); if ((rc != PCMK_OCF_OK) || (status != PCMK_LRM_OP_DONE)) { return NULL; } } return xml; } static int pe__call_id(xmlNode *op_xml) { int id = 0; if (op_xml) { crm_element_value_int(op_xml, XML_LRM_ATTR_CALLID, &id); } return id; } /*! * \brief Check whether a stop happened on the same node after some event * * \param[in] rsc Resource being checked * \param[in] node Node being checked * \param[in] xml_op Event that stop is being compared to * \param[in] data_set Cluster working set * * \return TRUE if stop happened after event, FALSE otherwise * * \note This is really unnecessary, but kept as a safety mechanism. We * currently don't save more than one successful event in history, so this * only matters when processing really old CIB files that we don't * technically support anymore, or as preparation for logging an extended * history in the future. */ static bool stop_happened_after(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op, pe_working_set_t *data_set) { xmlNode *stop_op = find_lrm_op(rsc->id, CRMD_ACTION_STOP, node->details->uname, NULL, TRUE, data_set); return (stop_op && (pe__call_id(stop_op) > pe__call_id(xml_op))); } static void unpack_migrate_to_success(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op, pe_working_set_t *data_set) { /* A successful migration sequence is: * migrate_to on source node * migrate_from on target node * stop on source node * * If a migrate_to is followed by a stop, the entire migration (successful * or failed) is complete, and we don't care what happened on the target. * * If no migrate_from has happened, the migration is considered to be * "partial". If the migrate_from failed, make sure the resource gets * stopped on both source and target (if up). * * If the migrate_to and migrate_from both succeeded (which also implies the * resource is no longer running on the source), but there is no stop, the * migration is considered to be "dangling". Schedule a stop on the source * in this case. */ int from_rc = 0; int from_status = 0; pe_node_t *target_node = NULL; pe_node_t *source_node = NULL; xmlNode *migrate_from = NULL; const char *source = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_SOURCE); const char *target = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_TARGET); // Sanity check CRM_CHECK(source && target && !strcmp(source, node->details->uname), return); if (stop_happened_after(rsc, node, xml_op, data_set)) { return; } // Clones are not allowed to migrate, so role can't be master rsc->role = RSC_ROLE_STARTED; target_node = pe_find_node(data_set->nodes, target); source_node = pe_find_node(data_set->nodes, source); // Check whether there was a migrate_from action on the target migrate_from = find_lrm_op(rsc->id, CRMD_ACTION_MIGRATED, target, source, FALSE, data_set); if (migrate_from) { crm_element_value_int(migrate_from, XML_LRM_ATTR_RC, &from_rc); crm_element_value_int(migrate_from, XML_LRM_ATTR_OPSTATUS, &from_status); pe_rsc_trace(rsc, "%s op on %s exited with status=%d, rc=%d", ID(migrate_from), target, from_status, from_rc); } if (migrate_from && from_rc == PCMK_OCF_OK && from_status == PCMK_LRM_OP_DONE) { /* The migrate_to and migrate_from both succeeded, so mark the migration * as "dangling". This will be used to schedule a stop action on the * source without affecting the target. */ pe_rsc_trace(rsc, "Detected dangling migration op: %s on %s", ID(xml_op), source); rsc->role = RSC_ROLE_STOPPED; rsc->dangling_migrations = g_list_prepend(rsc->dangling_migrations, node); } else if (migrate_from && (from_status != PCMK_LRM_OP_PENDING)) { // Failed if (target_node && target_node->details->online) { pe_rsc_trace(rsc, "Marking active on %s %p %d", target, target_node, target_node->details->online); native_add_running(rsc, target_node, data_set); } } else { // Pending, or complete but erased if (target_node && target_node->details->online) { pe_rsc_trace(rsc, "Marking active on %s %p %d", target, target_node, target_node->details->online); native_add_running(rsc, target_node, data_set); if (source_node && source_node->details->online) { /* This is a partial migration: the migrate_to completed * successfully on the source, but the migrate_from has not * completed. Remember the source and target; if the newly * chosen target remains the same when we schedule actions * later, we may continue with the migration. */ rsc->partial_migration_target = target_node; rsc->partial_migration_source = source_node; } } else { /* Consider it failed here - forces a restart, prevents migration */ set_bit(rsc->flags, pe_rsc_failed); clear_bit(rsc->flags, pe_rsc_allow_migrate); } } } static void unpack_migrate_to_failure(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op, pe_working_set_t *data_set) { int target_stop_id = 0; int target_migrate_from_id = 0; xmlNode *target_stop = NULL; xmlNode *target_migrate_from = NULL; const char *source = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_SOURCE); const char *target = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_TARGET); // Sanity check CRM_CHECK(source && target && !strcmp(source, node->details->uname), return); /* If a migration failed, we have to assume the resource is active. Clones * are not allowed to migrate, so role can't be master. */ rsc->role = RSC_ROLE_STARTED; // Check for stop on the target target_stop = find_lrm_op(rsc->id, CRMD_ACTION_STOP, target, NULL, TRUE, data_set); target_stop_id = pe__call_id(target_stop); // Check for migrate_from on the target target_migrate_from = find_lrm_op(rsc->id, CRMD_ACTION_MIGRATED, target, source, TRUE, data_set); target_migrate_from_id = pe__call_id(target_migrate_from); if ((target_stop == NULL) || (target_stop_id < target_migrate_from_id)) { /* There was no stop on the source, or a stop that happened before a * migrate_from, so assume the resource is still active on the target * (if it is up). */ node_t *target_node = pe_find_node(data_set->nodes, target); pe_rsc_trace(rsc, "stop (%d) + migrate_from (%d)", target_stop_id, target_migrate_from_id); if (target_node && target_node->details->online) { native_add_running(rsc, target_node, data_set); } } else if (target_migrate_from == NULL) { /* We know there was a stop on the target, but there may not have been a * migrate_from (the stop could have happened before migrate_from was * scheduled or attempted). * * That means this could be a "dangling" migration. But first, check * whether there is a newer migrate_from or start on the source node -- * it's possible the failed migration was followed by a successful * full restart or migration in the reverse direction, in which case we * don't want to force it to stop. */ xmlNode *source_migrate_from = NULL; xmlNode *source_start = NULL; int source_migrate_to_id = pe__call_id(xml_op); source_migrate_from = find_lrm_op(rsc->id, CRMD_ACTION_MIGRATED, source, NULL, TRUE, data_set); if (pe__call_id(source_migrate_from) > source_migrate_to_id) { return; } source_start = find_lrm_op(rsc->id, CRMD_ACTION_START, source, NULL, TRUE, data_set); if (pe__call_id(source_start) > source_migrate_to_id) { return; } // Mark node as having dangling migration so we can force a stop later rsc->dangling_migrations = g_list_prepend(rsc->dangling_migrations, node); } } static void unpack_migrate_from_failure(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op, pe_working_set_t *data_set) { xmlNode *source_stop = NULL; xmlNode *source_migrate_to = NULL; const char *source = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_SOURCE); const char *target = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_TARGET); // Sanity check CRM_CHECK(source && target && !strcmp(target, node->details->uname), return); /* If a migration failed, we have to assume the resource is active. Clones * are not allowed to migrate, so role can't be master. */ rsc->role = RSC_ROLE_STARTED; // Check for a stop on the source source_stop = find_lrm_op(rsc->id, CRMD_ACTION_STOP, source, NULL, TRUE, data_set); // Check for a migrate_to on the source source_migrate_to = find_lrm_op(rsc->id, CRMD_ACTION_MIGRATE, source, target, TRUE, data_set); if ((source_stop == NULL) || (pe__call_id(source_stop) < pe__call_id(source_migrate_to))) { /* There was no stop on the source, or a stop that happened before * migrate_to, so assume the resource is still active on the source (if * it is up). */ pe_node_t *source_node = pe_find_node(data_set->nodes, source); if (source_node && source_node->details->online) { native_add_running(rsc, source_node, data_set); } } } static void record_failed_op(xmlNode *op, const pe_node_t *node, const pe_resource_t *rsc, pe_working_set_t *data_set) { xmlNode *xIter = NULL; const char *op_key = crm_element_value(op, XML_LRM_ATTR_TASK_KEY); if (node->details->online == FALSE) { return; } for (xIter = data_set->failed->children; xIter; xIter = xIter->next) { const char *key = crm_element_value(xIter, XML_LRM_ATTR_TASK_KEY); const char *uname = crm_element_value(xIter, XML_ATTR_UNAME); if(safe_str_eq(op_key, key) && safe_str_eq(uname, node->details->uname)) { crm_trace("Skipping duplicate entry %s on %s", op_key, node->details->uname); return; } } crm_trace("Adding entry %s on %s", op_key, node->details->uname); crm_xml_add(op, XML_ATTR_UNAME, node->details->uname); crm_xml_add(op, XML_LRM_ATTR_RSCID, rsc->id); add_node_copy(data_set->failed, op); } static const char *get_op_key(xmlNode *xml_op) { const char *key = crm_element_value(xml_op, XML_LRM_ATTR_TASK_KEY); if(key == NULL) { key = ID(xml_op); } return key; } static void unpack_rsc_op_failure(resource_t * rsc, node_t * node, int rc, xmlNode * xml_op, xmlNode ** last_failure, enum action_fail_response * on_fail, pe_working_set_t * data_set) { guint interval_ms = 0; bool is_probe = FALSE; action_t *action = NULL; const char *key = get_op_key(xml_op); const char *task = crm_element_value(xml_op, XML_LRM_ATTR_TASK); CRM_ASSERT(rsc); *last_failure = xml_op; crm_element_value_ms(xml_op, XML_LRM_ATTR_INTERVAL_MS, &interval_ms); if ((interval_ms == 0) && safe_str_eq(task, CRMD_ACTION_STATUS)) { is_probe = TRUE; pe_rsc_trace(rsc, "is a probe: %s", key); } if (rc != PCMK_OCF_NOT_INSTALLED || is_set(data_set->flags, pe_flag_symmetric_cluster)) { crm_warn("Processing failed %s of %s on %s: %s " CRM_XS " rc=%d", (is_probe? "probe" : task), rsc->id, node->details->uname, services_ocf_exitcode_str(rc), rc); if (is_probe && (rc != PCMK_OCF_OK) && (rc != PCMK_OCF_NOT_RUNNING) && (rc != PCMK_OCF_RUNNING_MASTER)) { /* A failed (not just unexpected) probe result could mean the user * didn't know resources will be probed even where they can't run. */ crm_notice("If it is not possible for %s to run on %s, see " "the resource-discovery option for location constraints", rsc->id, node->details->uname); } record_failed_op(xml_op, node, rsc, data_set); } else { crm_trace("Processing failed op %s for %s on %s: %s (%d)", task, rsc->id, node->details->uname, services_ocf_exitcode_str(rc), rc); } action = custom_action(rsc, strdup(key), task, NULL, TRUE, FALSE, data_set); if ((action->on_fail <= action_fail_fence && *on_fail < action->on_fail) || (action->on_fail == action_fail_reset_remote && *on_fail <= action_fail_recover) || (action->on_fail == action_fail_restart_container && *on_fail <= action_fail_recover) || (*on_fail == action_fail_restart_container && action->on_fail >= action_fail_migrate)) { pe_rsc_trace(rsc, "on-fail %s -> %s for %s (%s)", fail2text(*on_fail), fail2text(action->on_fail), action->uuid, key); *on_fail = action->on_fail; } if (safe_str_eq(task, CRMD_ACTION_STOP)) { resource_location(rsc, node, -INFINITY, "__stop_fail__", data_set); } else if (safe_str_eq(task, CRMD_ACTION_MIGRATE)) { unpack_migrate_to_failure(rsc, node, xml_op, data_set); } else if (safe_str_eq(task, CRMD_ACTION_MIGRATED)) { unpack_migrate_from_failure(rsc, node, xml_op, data_set); } else if (safe_str_eq(task, CRMD_ACTION_PROMOTE)) { rsc->role = RSC_ROLE_MASTER; } else if (safe_str_eq(task, CRMD_ACTION_DEMOTE)) { if (action->on_fail == action_fail_block) { rsc->role = RSC_ROLE_MASTER; rsc->next_role = RSC_ROLE_STOPPED; } else if(rc == PCMK_OCF_NOT_RUNNING) { rsc->role = RSC_ROLE_STOPPED; } else { /* * Staying in master role would put the PE/TE into a loop. Setting * slave role is not dangerous because the resource will be stopped * as part of recovery, and any master promotion will be ordered * after that stop. */ rsc->role = RSC_ROLE_SLAVE; } } if(is_probe && rc == PCMK_OCF_NOT_INSTALLED) { /* leave stopped */ pe_rsc_trace(rsc, "Leaving %s stopped", rsc->id); rsc->role = RSC_ROLE_STOPPED; } else if (rsc->role < RSC_ROLE_STARTED) { pe_rsc_trace(rsc, "Setting %s active", rsc->id); set_active(rsc); } pe_rsc_trace(rsc, "Resource %s: role=%s, unclean=%s, on_fail=%s, fail_role=%s", rsc->id, role2text(rsc->role), node->details->unclean ? "true" : "false", fail2text(action->on_fail), role2text(action->fail_role)); if (action->fail_role != RSC_ROLE_STARTED && rsc->next_role < action->fail_role) { rsc->next_role = action->fail_role; } if (action->fail_role == RSC_ROLE_STOPPED) { int score = -INFINITY; resource_t *fail_rsc = rsc; if (fail_rsc->parent) { resource_t *parent = uber_parent(fail_rsc); if (pe_rsc_is_clone(parent) && is_not_set(parent->flags, pe_rsc_unique)) { /* For clone resources, if a child fails on an operation * with on-fail = stop, all the resources fail. Do this by preventing * the parent from coming up again. */ fail_rsc = parent; } } crm_warn("Making sure %s doesn't come up again", fail_rsc->id); /* make sure it doesn't come up again */ if (fail_rsc->allowed_nodes != NULL) { g_hash_table_destroy(fail_rsc->allowed_nodes); } fail_rsc->allowed_nodes = node_hash_from_list(data_set->nodes); g_hash_table_foreach(fail_rsc->allowed_nodes, set_node_score, &score); } pe_free_action(action); } /*! * \internal * \brief Remap operation status based on action result * * Given an action result, determine an appropriate operation status for the * purposes of responding to the action (the status provided by the executor is * not directly usable since the executor does not know what was expected). * * \param[in,out] rsc Resource that operation history entry is for * \param[in] rc Actual return code of operation * \param[in] target_rc Expected return code of operation * \param[in] node Node where operation was executed * \param[in] xml_op Operation history entry XML from CIB status * \param[in,out] on_fail What should be done about the result * \param[in] data_set Current cluster working set * * \return Operation status based on return code and action info * \note This may update the resource's current and next role. */ static int determine_op_status( resource_t *rsc, int rc, int target_rc, node_t * node, xmlNode * xml_op, enum action_fail_response * on_fail, pe_working_set_t * data_set) { guint interval_ms = 0; int result = PCMK_LRM_OP_DONE; const char *key = get_op_key(xml_op); const char *task = crm_element_value(xml_op, XML_LRM_ATTR_TASK); bool is_probe = FALSE; CRM_ASSERT(rsc); crm_element_value_ms(xml_op, XML_LRM_ATTR_INTERVAL_MS, &interval_ms); if ((interval_ms == 0) && safe_str_eq(task, CRMD_ACTION_STATUS)) { is_probe = TRUE; } if (target_rc < 0) { /* Pre-1.0 Pacemaker versions, and Pacemaker 1.1.6 or earlier with * Heartbeat 2.0.7 or earlier as the cluster layer, did not include the * target_rc in the transition key, which (along with the similar case * of a corrupted transition key in the CIB) will be reported to this * function as -1. Pacemaker 2.0+ does not support rolling upgrades from * those versions or processing of saved CIB files from those versions, * so we do not need to care much about this case. */ result = PCMK_LRM_OP_ERROR; crm_warn("Expected result not found for %s on %s (corrupt or obsolete CIB?)", key, node->details->uname); } else if (target_rc != rc) { result = PCMK_LRM_OP_ERROR; pe_rsc_debug(rsc, "%s on %s returned '%s' (%d) instead of the expected value: '%s' (%d)", key, node->details->uname, services_ocf_exitcode_str(rc), rc, services_ocf_exitcode_str(target_rc), target_rc); } switch (rc) { case PCMK_OCF_OK: // @TODO Should this be (rc != target_rc)? if (is_probe && (target_rc == PCMK_OCF_NOT_RUNNING)) { result = PCMK_LRM_OP_DONE; pe_rsc_info(rsc, "Operation %s found resource %s active on %s", task, rsc->id, node->details->uname); } break; case PCMK_OCF_NOT_RUNNING: if (is_probe || target_rc == rc || is_not_set(rsc->flags, pe_rsc_managed)) { result = PCMK_LRM_OP_DONE; rsc->role = RSC_ROLE_STOPPED; /* clear any previous failure actions */ *on_fail = action_fail_ignore; rsc->next_role = RSC_ROLE_UNKNOWN; } break; case PCMK_OCF_RUNNING_MASTER: if (is_probe && (rc != target_rc)) { result = PCMK_LRM_OP_DONE; pe_rsc_info(rsc, "Operation %s found resource %s active in master mode on %s", task, rsc->id, node->details->uname); } rsc->role = RSC_ROLE_MASTER; break; case PCMK_OCF_DEGRADED_MASTER: case PCMK_OCF_FAILED_MASTER: rsc->role = RSC_ROLE_MASTER; result = PCMK_LRM_OP_ERROR; break; case PCMK_OCF_NOT_CONFIGURED: result = PCMK_LRM_OP_ERROR_FATAL; break; case PCMK_OCF_UNIMPLEMENT_FEATURE: if (interval_ms > 0) { result = PCMK_LRM_OP_NOTSUPPORTED; break; } // fall through case PCMK_OCF_NOT_INSTALLED: case PCMK_OCF_INVALID_PARAM: case PCMK_OCF_INSUFFICIENT_PRIV: if (!pe_can_fence(data_set, node) && safe_str_eq(task, CRMD_ACTION_STOP)) { /* If a stop fails and we can't fence, there's nothing else we can do */ pe_proc_err("No further recovery can be attempted for %s: %s action failed with '%s' (%d)", rsc->id, task, services_ocf_exitcode_str(rc), rc); clear_bit(rsc->flags, pe_rsc_managed); set_bit(rsc->flags, pe_rsc_block); } result = PCMK_LRM_OP_ERROR_HARD; break; default: if (result == PCMK_LRM_OP_DONE) { crm_info("Treating unknown return code %d for %s on %s as failure", rc, key, node->details->uname); result = PCMK_LRM_OP_ERROR; } break; } return result; } // return TRUE if start or monitor last failure but parameters changed static bool should_clear_for_param_change(xmlNode *xml_op, const char *task, pe_resource_t *rsc, pe_node_t *node, pe_working_set_t *data_set) { if (!strcmp(task, "start") || !strcmp(task, "monitor")) { if (pe__bundle_needs_remote_name(rsc)) { /* We haven't allocated resources yet, so we can't reliably * substitute addr parameters for the REMOTE_CONTAINER_HACK. * When that's needed, defer the check until later. */ pe__add_param_check(xml_op, rsc, node, pe_check_last_failure, data_set); } else { op_digest_cache_t *digest_data = NULL; digest_data = rsc_action_digest_cmp(rsc, xml_op, node, data_set); switch (digest_data->rc) { case RSC_DIGEST_UNKNOWN: crm_trace("Resource %s history entry %s on %s" " has no digest to compare", rsc->id, get_op_key(xml_op), node->details->id); break; case RSC_DIGEST_MATCH: break; default: return TRUE; } } } return FALSE; } // Order action after fencing of remote node, given connection rsc static void order_after_remote_fencing(pe_action_t *action, pe_resource_t *remote_conn, pe_working_set_t *data_set) { pe_node_t *remote_node = pe_find_node(data_set->nodes, remote_conn->id); if (remote_node) { pe_action_t *fence = pe_fence_op(remote_node, NULL, TRUE, NULL, data_set); order_actions(fence, action, pe_order_implies_then); } } static bool should_ignore_failure_timeout(pe_resource_t *rsc, xmlNode *xml_op, const char *task, guint interval_ms, bool is_last_failure, pe_working_set_t *data_set) { /* Clearing failures of recurring monitors has special concerns. The * executor reports only changes in the monitor result, so if the * monitor is still active and still getting the same failure result, * that will go undetected after the failure is cleared. * * Also, the operation history will have the time when the recurring * monitor result changed to the given code, not the time when the * result last happened. * * @TODO We probably should clear such failures only when the failure * timeout has passed since the last occurrence of the failed result. * However we don't record that information. We could maybe approximate * that by clearing only if there is a more recent successful monitor or * stop result, but we don't even have that information at this point * since we are still unpacking the resource's operation history. * * This is especially important for remote connection resources with a * reconnect interval, so in that case, we skip clearing failures * if the remote node hasn't been fenced. */ if (rsc->remote_reconnect_ms && is_set(data_set->flags, pe_flag_stonith_enabled) && (interval_ms != 0) && safe_str_eq(task, CRMD_ACTION_STATUS)) { pe_node_t *remote_node = pe_find_node(data_set->nodes, rsc->id); if (remote_node && !remote_node->details->remote_was_fenced) { if (is_last_failure) { crm_info("Waiting to clear monitor failure for remote node %s" " until fencing has occurred", rsc->id); } return TRUE; } } return FALSE; } /*! * \internal * \brief Check operation age and schedule failure clearing when appropriate * * This function has two distinct purposes. The first is to check whether an * operation history entry is expired (i.e. the resource has a failure timeout, * the entry is older than the timeout, and the resource either has no fail * count or its fail count is entirely older than the timeout). The second is to * schedule fail count clearing when appropriate (i.e. the operation is expired * and either the resource has an expired fail count or the operation is a * last_failure for a remote connection resource with a reconnect interval, * or the operation is a last_failure for a start or monitor operation and the * resource's parameters have changed since the operation). * * \param[in] rsc Resource that operation happened to * \param[in] node Node that operation happened on * \param[in] rc Actual result of operation * \param[in] xml_op Operation history entry XML * \param[in] data_set Current working set * * \return TRUE if operation history entry is expired, FALSE otherwise */ static bool check_operation_expiry(pe_resource_t *rsc, pe_node_t *node, int rc, xmlNode *xml_op, pe_working_set_t *data_set) { bool expired = FALSE; bool is_last_failure = crm_ends_with(ID(xml_op), "_last_failure_0"); time_t last_run = 0; guint interval_ms = 0; int unexpired_fail_count = 0; const char *task = crm_element_value(xml_op, XML_LRM_ATTR_TASK); const char *clear_reason = NULL; crm_element_value_ms(xml_op, XML_LRM_ATTR_INTERVAL_MS, &interval_ms); if ((rsc->failure_timeout > 0) && (crm_element_value_epoch(xml_op, XML_RSC_OP_LAST_CHANGE, &last_run) == 0)) { // Resource has a failure-timeout, and history entry has a timestamp time_t now = get_effective_time(data_set); time_t last_failure = 0; // Is this particular operation history older than the failure timeout? if ((now >= (last_run + rsc->failure_timeout)) && !should_ignore_failure_timeout(rsc, xml_op, task, interval_ms, is_last_failure, data_set)) { expired = TRUE; } // Does the resource as a whole have an unexpired fail count? unexpired_fail_count = pe_get_failcount(node, rsc, &last_failure, pe_fc_effective, xml_op, data_set); // Update scheduler recheck time according to *last* failure crm_trace("%s@%lld is %sexpired @%lld with unexpired_failures=%d timeout=%ds" " last-failure@%lld", ID(xml_op), (long long) last_run, (expired? "" : "not "), (long long) now, unexpired_fail_count, rsc->failure_timeout, (long long) last_failure); last_failure += rsc->failure_timeout + 1; if (unexpired_fail_count && (now < last_failure)) { pe__update_recheck_time(last_failure, data_set); } } if (expired) { if (pe_get_failcount(node, rsc, NULL, pe_fc_default, xml_op, data_set)) { // There is a fail count ignoring timeout if (unexpired_fail_count == 0) { // There is no fail count considering timeout clear_reason = "it expired"; } else { /* This operation is old, but there is an unexpired fail count. * In a properly functioning cluster, this should only be * possible if this operation is not a failure (otherwise the * fail count should be expired too), so this is really just a * failsafe. */ expired = FALSE; } } else if (is_last_failure && rsc->remote_reconnect_ms) { /* Clear any expired last failure when reconnect interval is set, * even if there is no fail count. */ clear_reason = "reconnect interval is set"; } } if (!expired && is_last_failure && should_clear_for_param_change(xml_op, task, rsc, node, data_set)) { clear_reason = "resource parameters have changed"; } if (clear_reason != NULL) { // Schedule clearing of the fail count pe_action_t *clear_op = pe__clear_failcount(rsc, node, clear_reason, data_set); if (is_set(data_set->flags, pe_flag_stonith_enabled) && rsc->remote_reconnect_ms) { /* If we're clearing a remote connection due to a reconnect * interval, we want to wait until any scheduled fencing * completes. * * We could limit this to remote_node->details->unclean, but at * this point, that's always true (it won't be reliable until * after unpack_node_loop() is done). */ crm_info("Clearing %s failure will wait until any scheduled " "fencing of %s completes", task, rsc->id); order_after_remote_fencing(clear_op, rsc, data_set); } } if (expired && (interval_ms == 0) && safe_str_eq(task, CRMD_ACTION_STATUS)) { switch(rc) { case PCMK_OCF_OK: case PCMK_OCF_NOT_RUNNING: case PCMK_OCF_RUNNING_MASTER: case PCMK_OCF_DEGRADED: case PCMK_OCF_DEGRADED_MASTER: // Don't expire probes that return these values expired = FALSE; break; } } return expired; } int pe__target_rc_from_xml(xmlNode *xml_op) { int target_rc = 0; const char *key = crm_element_value(xml_op, XML_ATTR_TRANSITION_KEY); if (key == NULL) { return -1; } decode_transition_key(key, NULL, NULL, NULL, &target_rc); return target_rc; } static enum action_fail_response get_action_on_fail(resource_t *rsc, const char *key, const char *task, pe_working_set_t * data_set) { int result = action_fail_recover; action_t *action = custom_action(rsc, strdup(key), task, NULL, TRUE, FALSE, data_set); result = action->on_fail; pe_free_action(action); return result; } static void update_resource_state(resource_t * rsc, node_t * node, xmlNode * xml_op, const char * task, int rc, xmlNode * last_failure, enum action_fail_response * on_fail, pe_working_set_t * data_set) { gboolean clear_past_failure = FALSE; CRM_ASSERT(rsc); CRM_ASSERT(xml_op); if (rc == PCMK_OCF_NOT_RUNNING) { clear_past_failure = TRUE; } else if (rc == PCMK_OCF_NOT_INSTALLED) { rsc->role = RSC_ROLE_STOPPED; } else if (safe_str_eq(task, CRMD_ACTION_STATUS)) { if (last_failure) { const char *op_key = get_op_key(xml_op); const char *last_failure_key = get_op_key(last_failure); if (safe_str_eq(op_key, last_failure_key)) { clear_past_failure = TRUE; } } if (rsc->role < RSC_ROLE_STARTED) { set_active(rsc); } } else if (safe_str_eq(task, CRMD_ACTION_START)) { rsc->role = RSC_ROLE_STARTED; clear_past_failure = TRUE; } else if (safe_str_eq(task, CRMD_ACTION_STOP)) { rsc->role = RSC_ROLE_STOPPED; clear_past_failure = TRUE; } else if (safe_str_eq(task, CRMD_ACTION_PROMOTE)) { rsc->role = RSC_ROLE_MASTER; clear_past_failure = TRUE; } else if (safe_str_eq(task, CRMD_ACTION_DEMOTE)) { /* Demote from Master does not clear an error */ rsc->role = RSC_ROLE_SLAVE; } else if (safe_str_eq(task, CRMD_ACTION_MIGRATED)) { rsc->role = RSC_ROLE_STARTED; clear_past_failure = TRUE; } else if (safe_str_eq(task, CRMD_ACTION_MIGRATE)) { unpack_migrate_to_success(rsc, node, xml_op, data_set); } else if (rsc->role < RSC_ROLE_STARTED) { pe_rsc_trace(rsc, "%s active on %s", rsc->id, node->details->uname); set_active(rsc); } /* clear any previous failure actions */ if (clear_past_failure) { switch (*on_fail) { case action_fail_stop: case action_fail_fence: case action_fail_migrate: case action_fail_standby: pe_rsc_trace(rsc, "%s.%s is not cleared by a completed stop", rsc->id, fail2text(*on_fail)); break; case action_fail_block: case action_fail_ignore: case action_fail_recover: case action_fail_restart_container: *on_fail = action_fail_ignore; rsc->next_role = RSC_ROLE_UNKNOWN; break; case action_fail_reset_remote: if (rsc->remote_reconnect_ms == 0) { /* With no reconnect interval, the connection is allowed to * start again after the remote node is fenced and * completely stopped. (With a reconnect interval, we wait * for the failure to be cleared entirely before attempting * to reconnect.) */ *on_fail = action_fail_ignore; rsc->next_role = RSC_ROLE_UNKNOWN; } break; } } } /*! * \internal * \brief Remap informational monitor results to usual values * * Certain OCF result codes are for providing extended information to the * user about services that aren't yet failed but not entirely healthy either. * These must be treated as the "normal" result by pacemaker. * * \param[in] rc Actual result of a monitor action * \param[in] xml_op Operation history XML * \param[in] node Node that operation happened on * \param[in] rsc Resource that operation happened to * \param[in] data_set Cluster working set * * \return Result code that pacemaker should use * * \note If the result is remapped, and the node is not shutting down or failed, * the operation will be recorded in the data set's list of failed * operations, to highlight it for the user. */ static int remap_monitor_rc(int rc, xmlNode *xml_op, const pe_node_t *node, const pe_resource_t *rsc, pe_working_set_t *data_set) { int remapped_rc = rc; switch (rc) { case PCMK_OCF_DEGRADED: remapped_rc = PCMK_OCF_OK; break; case PCMK_OCF_DEGRADED_MASTER: remapped_rc = PCMK_OCF_RUNNING_MASTER; break; default: break; } if (rc != remapped_rc) { crm_trace("Remapping monitor result %d to %d", rc, remapped_rc); if (!node->details->shutdown || node->details->online) { record_failed_op(xml_op, node, rsc, data_set); } } return remapped_rc; } static void unpack_rsc_op(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op, xmlNode **last_failure, enum action_fail_response *on_fail, pe_working_set_t *data_set) { int task_id = 0; const char *task = NULL; const char *task_key = NULL; int rc = 0; int status = PCMK_LRM_OP_UNKNOWN; int target_rc = pe__target_rc_from_xml(xml_op); guint interval_ms = 0; bool expired = FALSE; resource_t *parent = rsc; enum action_fail_response failure_strategy = action_fail_recover; CRM_CHECK(rsc != NULL, return); CRM_CHECK(node != NULL, return); CRM_CHECK(xml_op != NULL, return); task_key = get_op_key(xml_op); task = crm_element_value(xml_op, XML_LRM_ATTR_TASK); crm_element_value_int(xml_op, XML_LRM_ATTR_RC, &rc); crm_element_value_int(xml_op, XML_LRM_ATTR_CALLID, &task_id); crm_element_value_int(xml_op, XML_LRM_ATTR_OPSTATUS, &status); crm_element_value_ms(xml_op, XML_LRM_ATTR_INTERVAL_MS, &interval_ms); CRM_CHECK(task != NULL, return); CRM_CHECK(status <= PCMK_LRM_OP_INVALID, return); CRM_CHECK(status >= PCMK_LRM_OP_PENDING, return); if (safe_str_eq(task, CRMD_ACTION_NOTIFY) || safe_str_eq(task, CRMD_ACTION_METADATA)) { /* safe to ignore these */ return; } if (is_not_set(rsc->flags, pe_rsc_unique)) { parent = uber_parent(rsc); } pe_rsc_trace(rsc, "Unpacking task %s/%s (call_id=%d, status=%d, rc=%d) on %s (role=%s)", task_key, task, task_id, status, rc, node->details->uname, role2text(rsc->role)); if (node->details->unclean) { pe_rsc_trace(rsc, "Node %s (where %s is running) is unclean." " Further action depends on the value of the stop's on-fail attribute", node->details->uname, rsc->id); } /* It should be possible to call remap_monitor_rc() first then call * check_operation_expiry() only if rc != target_rc, because there should * never be a fail count without at least one unexpected result in the * resource history. That would be more efficient by avoiding having to call * check_operation_expiry() for expected results. * * However, we do have such configurations in the scheduler regression * tests, even if it shouldn't be possible with the current code. It's * probably a good idea anyway, but that would require updating the test * inputs to something currently possible. */ if ((status != PCMK_LRM_OP_NOT_INSTALLED) && check_operation_expiry(rsc, node, rc, xml_op, data_set)) { expired = TRUE; } if (!strcmp(task, CRMD_ACTION_STATUS)) { rc = remap_monitor_rc(rc, xml_op, node, rsc, data_set); } if (expired && (rc != target_rc)) { const char *magic = crm_element_value(xml_op, XML_ATTR_TRANSITION_MAGIC); if (interval_ms == 0) { crm_notice("Ignoring expired %s failure on %s " CRM_XS " actual=%d expected=%d magic=%s", task_key, node->details->uname, rc, target_rc, magic); goto done; } else if(node->details->online && node->details->unclean == FALSE) { /* Reschedule the recurring monitor. CancelXmlOp() won't work at * this stage, so as a hacky workaround, forcibly change the restart * digest so check_action_definition() does what we want later. * * @TODO We should skip this if there is a newer successful monitor. * Also, this causes rescheduling only if the history entry * has an op-digest (which the expire-non-blocked-failure * scheduler regression test doesn't, but that may not be a * realistic scenario in production). */ crm_notice("Rescheduling %s after failure expired on %s " CRM_XS " actual=%d expected=%d magic=%s", task_key, node->details->uname, rc, target_rc, magic); crm_xml_add(xml_op, XML_LRM_ATTR_RESTART_DIGEST, "calculated-failure-timeout"); goto done; } } /* If the executor reported an operation status of anything but done or * error, consider that final. But for done or error, we know better whether * it should be treated as a failure or not, because we know the expected * result. */ if(status == PCMK_LRM_OP_DONE || status == PCMK_LRM_OP_ERROR) { status = determine_op_status(rsc, rc, target_rc, node, xml_op, on_fail, data_set); } pe_rsc_trace(rsc, "Handling status: %d", status); switch (status) { case PCMK_LRM_OP_CANCELLED: /* do nothing?? */ pe_err("Don't know what to do for cancelled ops yet"); break; case PCMK_LRM_OP_PENDING: if (safe_str_eq(task, CRMD_ACTION_START)) { set_bit(rsc->flags, pe_rsc_start_pending); set_active(rsc); } else if (safe_str_eq(task, CRMD_ACTION_PROMOTE)) { rsc->role = RSC_ROLE_MASTER; } else if (safe_str_eq(task, CRMD_ACTION_MIGRATE) && node->details->unclean) { /* If a pending migrate_to action is out on a unclean node, * we have to force the stop action on the target. */ const char *migrate_target = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_TARGET); node_t *target = pe_find_node(data_set->nodes, migrate_target); if (target) { stop_action(rsc, target, FALSE); } } if (rsc->pending_task == NULL) { if (safe_str_eq(task, CRMD_ACTION_STATUS) && (interval_ms == 0)) { /* Pending probes are not printed, even if pending * operations are requested. If someone ever requests that * behavior, uncomment this and the corresponding part of * native.c:native_pending_task(). */ /*rsc->pending_task = strdup("probe");*/ /*rsc->pending_node = node;*/ } else { rsc->pending_task = strdup(task); rsc->pending_node = node; } } break; case PCMK_LRM_OP_DONE: pe_rsc_trace(rsc, "%s/%s completed on %s", rsc->id, task, node->details->uname); update_resource_state(rsc, node, xml_op, task, rc, *last_failure, on_fail, data_set); break; case PCMK_LRM_OP_NOT_INSTALLED: failure_strategy = get_action_on_fail(rsc, task_key, task, data_set); if (failure_strategy == action_fail_ignore) { crm_warn("Cannot ignore failed %s (status=%d, rc=%d) on %s: " "Resource agent doesn't exist", task_key, status, rc, node->details->uname); /* Also for printing it as "FAILED" by marking it as pe_rsc_failed later */ *on_fail = action_fail_migrate; } resource_location(parent, node, -INFINITY, "hard-error", data_set); unpack_rsc_op_failure(rsc, node, rc, xml_op, last_failure, on_fail, data_set); break; case PCMK_LRM_OP_NOT_CONNECTED: if (pe__is_guest_or_remote_node(node) && is_set(node->details->remote_rsc->flags, pe_rsc_managed)) { /* We should never get into a situation where a managed remote * connection resource is considered OK but a resource action * behind the connection gets a "not connected" status. But as a * fail-safe in case a bug or unusual circumstances do lead to * that, ensure the remote connection is considered failed. */ set_bit(node->details->remote_rsc->flags, pe_rsc_failed); } // fall through case PCMK_LRM_OP_ERROR: case PCMK_LRM_OP_ERROR_HARD: case PCMK_LRM_OP_ERROR_FATAL: case PCMK_LRM_OP_TIMEOUT: case PCMK_LRM_OP_NOTSUPPORTED: case PCMK_LRM_OP_INVALID: failure_strategy = get_action_on_fail(rsc, task_key, task, data_set); if ((failure_strategy == action_fail_ignore) || (failure_strategy == action_fail_restart_container && safe_str_eq(task, CRMD_ACTION_STOP))) { crm_warn("Pretending the failure of %s (rc=%d) on %s succeeded", task_key, rc, node->details->uname); update_resource_state(rsc, node, xml_op, task, target_rc, *last_failure, on_fail, data_set); crm_xml_add(xml_op, XML_ATTR_UNAME, node->details->uname); set_bit(rsc->flags, pe_rsc_failure_ignored); record_failed_op(xml_op, node, rsc, data_set); if (failure_strategy == action_fail_restart_container && *on_fail <= action_fail_recover) { *on_fail = failure_strategy; } } else { unpack_rsc_op_failure(rsc, node, rc, xml_op, last_failure, on_fail, data_set); if(status == PCMK_LRM_OP_ERROR_HARD) { do_crm_log(rc != PCMK_OCF_NOT_INSTALLED?LOG_ERR:LOG_NOTICE, "Preventing %s from re-starting on %s: operation %s failed '%s' (%d)", parent->id, node->details->uname, task, services_ocf_exitcode_str(rc), rc); resource_location(parent, node, -INFINITY, "hard-error", data_set); } else if(status == PCMK_LRM_OP_ERROR_FATAL) { crm_err("Preventing %s from re-starting anywhere: operation %s failed '%s' (%d)", parent->id, task, services_ocf_exitcode_str(rc), rc); resource_location(parent, NULL, -INFINITY, "fatal-error", data_set); } } break; } done: pe_rsc_trace(rsc, "Resource %s after %s: role=%s, next=%s", rsc->id, task, role2text(rsc->role), role2text(rsc->next_role)); } static void add_node_attrs(xmlNode *xml_obj, pe_node_t *node, bool overwrite, pe_working_set_t *data_set) { const char *cluster_name = NULL; g_hash_table_insert(node->details->attrs, strdup(CRM_ATTR_UNAME), strdup(node->details->uname)); g_hash_table_insert(node->details->attrs, strdup(CRM_ATTR_ID), strdup(node->details->id)); if (safe_str_eq(node->details->id, data_set->dc_uuid)) { data_set->dc_node = node; node->details->is_dc = TRUE; g_hash_table_insert(node->details->attrs, strdup(CRM_ATTR_IS_DC), strdup(XML_BOOLEAN_TRUE)); } else { g_hash_table_insert(node->details->attrs, strdup(CRM_ATTR_IS_DC), strdup(XML_BOOLEAN_FALSE)); } cluster_name = g_hash_table_lookup(data_set->config_hash, "cluster-name"); if (cluster_name) { g_hash_table_insert(node->details->attrs, strdup(CRM_ATTR_CLUSTER_NAME), strdup(cluster_name)); } pe__unpack_dataset_nvpairs(xml_obj, XML_TAG_ATTR_SETS, NULL, node->details->attrs, NULL, overwrite, data_set); if (pe_node_attribute_raw(node, CRM_ATTR_SITE_NAME) == NULL) { const char *site_name = pe_node_attribute_raw(node, "site-name"); if (site_name) { g_hash_table_insert(node->details->attrs, strdup(CRM_ATTR_SITE_NAME), strdup(site_name)); } else if (cluster_name) { /* Default to cluster-name if unset */ g_hash_table_insert(node->details->attrs, strdup(CRM_ATTR_SITE_NAME), strdup(cluster_name)); } } } static GListPtr extract_operations(const char *node, const char *rsc, xmlNode * rsc_entry, gboolean active_filter) { int counter = -1; int stop_index = -1; int start_index = -1; xmlNode *rsc_op = NULL; GListPtr gIter = NULL; GListPtr op_list = NULL; GListPtr sorted_op_list = NULL; /* extract operations */ op_list = NULL; sorted_op_list = NULL; for (rsc_op = __xml_first_child_element(rsc_entry); rsc_op != NULL; rsc_op = __xml_next_element(rsc_op)) { if (crm_str_eq((const char *)rsc_op->name, XML_LRM_TAG_RSC_OP, TRUE)) { crm_xml_add(rsc_op, "resource", rsc); crm_xml_add(rsc_op, XML_ATTR_UNAME, node); op_list = g_list_prepend(op_list, rsc_op); } } if (op_list == NULL) { /* if there are no operations, there is nothing to do */ return NULL; } sorted_op_list = g_list_sort(op_list, sort_op_by_callid); /* create active recurring operations as optional */ if (active_filter == FALSE) { return sorted_op_list; } op_list = NULL; calculate_active_ops(sorted_op_list, &start_index, &stop_index); for (gIter = sorted_op_list; gIter != NULL; gIter = gIter->next) { xmlNode *rsc_op = (xmlNode *) gIter->data; counter++; if (start_index < stop_index) { crm_trace("Skipping %s: not active", ID(rsc_entry)); break; } else if (counter < start_index) { crm_trace("Skipping %s: old", ID(rsc_op)); continue; } op_list = g_list_append(op_list, rsc_op); } g_list_free(sorted_op_list); return op_list; } GListPtr find_operations(const char *rsc, const char *node, gboolean active_filter, pe_working_set_t * data_set) { GListPtr output = NULL; GListPtr intermediate = NULL; xmlNode *tmp = NULL; xmlNode *status = find_xml_node(data_set->input, XML_CIB_TAG_STATUS, TRUE); node_t *this_node = NULL; xmlNode *node_state = NULL; for (node_state = __xml_first_child_element(status); node_state != NULL; node_state = __xml_next_element(node_state)) { if (crm_str_eq((const char *)node_state->name, XML_CIB_TAG_STATE, TRUE)) { const char *uname = crm_element_value(node_state, XML_ATTR_UNAME); if (node != NULL && safe_str_neq(uname, node)) { continue; } this_node = pe_find_node(data_set->nodes, uname); if(this_node == NULL) { CRM_LOG_ASSERT(this_node != NULL); continue; } else if (pe__is_guest_or_remote_node(this_node)) { determine_remote_online_status(data_set, this_node); } else { determine_online_status(node_state, this_node, data_set); } if (this_node->details->online || is_set(data_set->flags, pe_flag_stonith_enabled)) { /* offline nodes run no resources... * unless stonith is enabled in which case we need to * make sure rsc start events happen after the stonith */ xmlNode *lrm_rsc = NULL; tmp = find_xml_node(node_state, XML_CIB_TAG_LRM, FALSE); tmp = find_xml_node(tmp, XML_LRM_TAG_RESOURCES, FALSE); for (lrm_rsc = __xml_first_child_element(tmp); lrm_rsc != NULL; lrm_rsc = __xml_next_element(lrm_rsc)) { if (crm_str_eq((const char *)lrm_rsc->name, XML_LRM_TAG_RESOURCE, TRUE)) { const char *rsc_id = crm_element_value(lrm_rsc, XML_ATTR_ID); if (rsc != NULL && safe_str_neq(rsc_id, rsc)) { continue; } intermediate = extract_operations(uname, rsc_id, lrm_rsc, active_filter); output = g_list_concat(output, intermediate); } } } } } return output; } diff --git a/lib/pengine/utils.c b/lib/pengine/utils.c index 0058ebe707..3fc072f996 100644 --- a/lib/pengine/utils.c +++ b/lib/pengine/utils.c @@ -1,2586 +1,2592 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU Lesser General Public License * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include extern xmlNode *get_object_root(const char *object_type, xmlNode * the_root); void print_str_str(gpointer key, gpointer value, gpointer user_data); gboolean ghash_free_str_str(gpointer key, gpointer value, gpointer user_data); void unpack_operation(action_t * action, xmlNode * xml_obj, resource_t * container, pe_working_set_t * data_set); static xmlNode *find_rsc_op_entry_helper(resource_t * rsc, const char *key, gboolean include_disabled); #if ENABLE_VERSIONED_ATTRS pe_rsc_action_details_t * pe_rsc_action_details(pe_action_t *action) { pe_rsc_action_details_t *details; CRM_CHECK(action != NULL, return NULL); if (action->action_details == NULL) { action->action_details = calloc(1, sizeof(pe_rsc_action_details_t)); CRM_CHECK(action->action_details != NULL, return NULL); } details = (pe_rsc_action_details_t *) action->action_details; if (details->versioned_parameters == NULL) { details->versioned_parameters = create_xml_node(NULL, XML_TAG_OP_VER_ATTRS); } if (details->versioned_meta == NULL) { details->versioned_meta = create_xml_node(NULL, XML_TAG_OP_VER_META); } return details; } static void pe_free_rsc_action_details(pe_action_t *action) { pe_rsc_action_details_t *details; if ((action == NULL) || (action->action_details == NULL)) { return; } details = (pe_rsc_action_details_t *) action->action_details; if (details->versioned_parameters) { free_xml(details->versioned_parameters); } if (details->versioned_meta) { free_xml(details->versioned_meta); } action->action_details = NULL; } #endif /*! * \internal * \brief Check whether we can fence a particular node * * \param[in] data_set Working set for cluster * \param[in] node Name of node to check * * \return true if node can be fenced, false otherwise */ bool pe_can_fence(pe_working_set_t *data_set, pe_node_t *node) { if (pe__is_guest_node(node)) { /* Guest nodes are fenced by stopping their container resource. We can * do that if the container's host is either online or fenceable. */ pe_resource_t *rsc = node->details->remote_rsc->container; for (GList *n = rsc->running_on; n != NULL; n = n->next) { pe_node_t *container_node = n->data; if (!container_node->details->online && !pe_can_fence(data_set, container_node)) { return false; } } return true; } else if(is_not_set(data_set->flags, pe_flag_stonith_enabled)) { return false; /* Turned off */ } else if (is_not_set(data_set->flags, pe_flag_have_stonith_resource)) { return false; /* No devices */ } else if (is_set(data_set->flags, pe_flag_have_quorum)) { return true; } else if (data_set->no_quorum_policy == no_quorum_ignore) { return true; } else if(node == NULL) { return false; } else if(node->details->online) { crm_notice("We can fence %s without quorum because they're in our membership", node->details->uname); return true; } crm_trace("Cannot fence %s", node->details->uname); return false; } node_t * node_copy(const node_t *this_node) { node_t *new_node = NULL; CRM_CHECK(this_node != NULL, return NULL); new_node = calloc(1, sizeof(node_t)); CRM_ASSERT(new_node != NULL); crm_trace("Copying %p (%s) to %p", this_node, this_node->details->uname, new_node); new_node->rsc_discover_mode = this_node->rsc_discover_mode; new_node->weight = this_node->weight; new_node->fixed = this_node->fixed; new_node->details = this_node->details; return new_node; } /* any node in list1 or list2 and not in the other gets a score of -INFINITY */ void node_list_exclude(GHashTable * hash, GListPtr list, gboolean merge_scores) { GHashTable *result = hash; node_t *other_node = NULL; GListPtr gIter = list; GHashTableIter iter; node_t *node = NULL; g_hash_table_iter_init(&iter, hash); while (g_hash_table_iter_next(&iter, NULL, (void **)&node)) { other_node = pe_find_node_id(list, node->details->id); if (other_node == NULL) { node->weight = -INFINITY; } else if (merge_scores) { node->weight = merge_weights(node->weight, other_node->weight); } } for (; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; other_node = pe_hash_table_lookup(result, node->details->id); if (other_node == NULL) { node_t *new_node = node_copy(node); new_node->weight = -INFINITY; g_hash_table_insert(result, (gpointer) new_node->details->id, new_node); } } } GHashTable * node_hash_from_list(GListPtr list) { GListPtr gIter = list; GHashTable *result = g_hash_table_new_full(crm_str_hash, g_str_equal, NULL, free); for (; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; node_t *n = node_copy(node); g_hash_table_insert(result, (gpointer) n->details->id, n); } return result; } GListPtr node_list_dup(GListPtr list1, gboolean reset, gboolean filter) { GListPtr result = NULL; GListPtr gIter = list1; for (; gIter != NULL; gIter = gIter->next) { node_t *new_node = NULL; node_t *this_node = (node_t *) gIter->data; if (filter && this_node->weight < 0) { continue; } new_node = node_copy(this_node); if (reset) { new_node->weight = 0; } if (new_node != NULL) { result = g_list_prepend(result, new_node); } } return result; } gint sort_node_uname(gconstpointer a, gconstpointer b) { const char *name_a = ((const node_t *) a)->details->uname; const char *name_b = ((const node_t *) b)->details->uname; while (*name_a && *name_b) { if (isdigit(*name_a) && isdigit(*name_b)) { // If node names contain a number, sort numerically char *end_a = NULL; char *end_b = NULL; long num_a = strtol(name_a, &end_a, 10); long num_b = strtol(name_b, &end_b, 10); // allow ordering e.g. 007 > 7 size_t len_a = end_a - name_a; size_t len_b = end_b - name_b; if (num_a < num_b) { return -1; } else if (num_a > num_b) { return 1; } else if (len_a < len_b) { return -1; } else if (len_a > len_b) { return 1; } name_a = end_a; name_b = end_b; } else { // Compare non-digits case-insensitively int lower_a = tolower(*name_a); int lower_b = tolower(*name_b); if (lower_a < lower_b) { return -1; } else if (lower_a > lower_b) { return 1; } ++name_a; ++name_b; } } if (!*name_a && *name_b) { return -1; } else if (*name_a && !*name_b) { return 1; } return 0; } void dump_node_scores_worker(int level, const char *file, const char *function, int line, resource_t * rsc, const char *comment, GHashTable * nodes) { GHashTable *hash = nodes; GHashTableIter iter; node_t *node = NULL; if (rsc) { hash = rsc->allowed_nodes; } if (rsc && is_set(rsc->flags, pe_rsc_orphan)) { /* Don't show the allocation scores for orphans */ return; } if (level == 0) { char score[128]; int len = sizeof(score); /* For now we want this in sorted order to keep the regression tests happy */ GListPtr gIter = NULL; GListPtr list = g_hash_table_get_values(hash); list = g_list_sort(list, sort_node_uname); gIter = list; for (; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; /* This function is called a whole lot, use stack allocated score */ score2char_stack(node->weight, score, len); if (rsc) { printf("%s: %s allocation score on %s: %s\n", comment, rsc->id, node->details->uname, score); } else { printf("%s: %s = %s\n", comment, node->details->uname, score); } } g_list_free(list); } else if (hash) { char score[128]; int len = sizeof(score); g_hash_table_iter_init(&iter, hash); while (g_hash_table_iter_next(&iter, NULL, (void **)&node)) { /* This function is called a whole lot, use stack allocated score */ score2char_stack(node->weight, score, len); if (rsc) { do_crm_log_alias(LOG_TRACE, file, function, line, "%s: %s allocation score on %s: %s", comment, rsc->id, node->details->uname, score); } else { do_crm_log_alias(LOG_TRACE, file, function, line + 1, "%s: %s = %s", comment, node->details->uname, score); } } } if (rsc && rsc->children) { GListPtr gIter = NULL; gIter = rsc->children; for (; gIter != NULL; gIter = gIter->next) { resource_t *child = (resource_t *) gIter->data; dump_node_scores_worker(level, file, function, line, child, comment, nodes); } } } static void append_dump_text(gpointer key, gpointer value, gpointer user_data) { char **dump_text = user_data; char *new_text = crm_strdup_printf("%s %s=%s", *dump_text, (char *)key, (char *)value); free(*dump_text); *dump_text = new_text; } void dump_node_capacity(int level, const char *comment, node_t * node) { char *dump_text = crm_strdup_printf("%s: %s capacity:", comment, node->details->uname); g_hash_table_foreach(node->details->utilization, append_dump_text, &dump_text); if (level == 0) { fprintf(stdout, "%s\n", dump_text); } else { crm_trace("%s", dump_text); } free(dump_text); } void dump_rsc_utilization(int level, const char *comment, resource_t * rsc, node_t * node) { char *dump_text = crm_strdup_printf("%s: %s utilization on %s:", comment, rsc->id, node->details->uname); g_hash_table_foreach(rsc->utilization, append_dump_text, &dump_text); if (level == 0) { fprintf(stdout, "%s\n", dump_text); } else { crm_trace("%s", dump_text); } free(dump_text); } gint sort_rsc_index(gconstpointer a, gconstpointer b) { const resource_t *resource1 = (const resource_t *)a; const resource_t *resource2 = (const resource_t *)b; if (a == NULL && b == NULL) { return 0; } if (a == NULL) { return 1; } if (b == NULL) { return -1; } if (resource1->sort_index > resource2->sort_index) { return -1; } if (resource1->sort_index < resource2->sort_index) { return 1; } return 0; } gint sort_rsc_priority(gconstpointer a, gconstpointer b) { const resource_t *resource1 = (const resource_t *)a; const resource_t *resource2 = (const resource_t *)b; if (a == NULL && b == NULL) { return 0; } if (a == NULL) { return 1; } if (b == NULL) { return -1; } if (resource1->priority > resource2->priority) { return -1; } if (resource1->priority < resource2->priority) { return 1; } return 0; } action_t * custom_action(resource_t * rsc, char *key, const char *task, node_t * on_node, gboolean optional, gboolean save_action, pe_working_set_t * data_set) { action_t *action = NULL; GListPtr possible_matches = NULL; CRM_CHECK(key != NULL, return NULL); CRM_CHECK(task != NULL, free(key); return NULL); if (save_action && rsc != NULL) { possible_matches = find_actions(rsc->actions, key, on_node); } else if(save_action) { #if 0 action = g_hash_table_lookup(data_set->singletons, key); #else /* More expensive but takes 'node' into account */ possible_matches = find_actions(data_set->actions, key, on_node); #endif } if(data_set->singletons == NULL) { data_set->singletons = g_hash_table_new_full(crm_str_hash, g_str_equal, NULL, NULL); } if (possible_matches != NULL) { if (g_list_length(possible_matches) > 1) { pe_warn("Action %s for %s on %s exists %d times", task, rsc ? rsc->id : "", on_node ? on_node->details->uname : "", g_list_length(possible_matches)); } action = g_list_nth_data(possible_matches, 0); pe_rsc_trace(rsc, "Found existing action %d (%s) for %s (%s) on %s", action->id, action->uuid, (rsc? rsc->id : "no resource"), task, (on_node? on_node->details->uname : "no node")); g_list_free(possible_matches); } if (action == NULL) { if (save_action) { pe_rsc_trace(rsc, "Creating %s action %d: %s for %s (%s) on %s", (optional? "optional" : " mandatory"), data_set->action_id, key, (rsc? rsc->id : "no resource"), task, (on_node? on_node->details->uname : "no node")); } action = calloc(1, sizeof(action_t)); if (save_action) { action->id = data_set->action_id++; } else { action->id = 0; } action->rsc = rsc; CRM_ASSERT(task != NULL); action->task = strdup(task); if (on_node) { action->node = node_copy(on_node); } action->uuid = strdup(key); pe_set_action_bit(action, pe_action_runnable); if (optional) { pe_set_action_bit(action, pe_action_optional); } else { pe_clear_action_bit(action, pe_action_optional); } action->extra = crm_str_table_new(); action->meta = crm_str_table_new(); if (save_action) { data_set->actions = g_list_prepend(data_set->actions, action); if(rsc == NULL) { g_hash_table_insert(data_set->singletons, action->uuid, action); } } if (rsc != NULL) { action->op_entry = find_rsc_op_entry_helper(rsc, key, TRUE); unpack_operation(action, action->op_entry, rsc->container, data_set); if (save_action) { rsc->actions = g_list_prepend(rsc->actions, action); } } if (save_action) { pe_rsc_trace(rsc, "Action %d created", action->id); } } if (!optional && is_set(action->flags, pe_action_optional)) { pe_rsc_trace(rsc, "Unset optional on action %d", action->id); pe_clear_action_bit(action, pe_action_optional); } if (rsc != NULL) { enum action_tasks a_task = text2task(action->task); int warn_level = LOG_TRACE; if (save_action) { warn_level = LOG_WARNING; } if (is_set(action->flags, pe_action_have_node_attrs) == FALSE && action->node != NULL && action->op_entry != NULL) { pe_set_action_bit(action, pe_action_have_node_attrs); pe__unpack_dataset_nvpairs(action->op_entry, XML_TAG_ATTR_SETS, action->node->details->attrs, action->extra, NULL, FALSE, data_set); } if (is_set(action->flags, pe_action_pseudo)) { /* leave untouched */ } else if (action->node == NULL) { pe_rsc_trace(rsc, "Unset runnable on %s", action->uuid); pe_clear_action_bit(action, pe_action_runnable); } else if (is_not_set(rsc->flags, pe_rsc_managed) && g_hash_table_lookup(action->meta, XML_LRM_ATTR_INTERVAL_MS) == NULL) { crm_debug("Action %s (unmanaged)", action->uuid); pe_rsc_trace(rsc, "Set optional on %s", action->uuid); pe_set_action_bit(action, pe_action_optional); /* action->runnable = FALSE; */ } else if (action->node->details->online == FALSE && (!pe__is_guest_node(action->node) || action->node->details->remote_requires_reset)) { pe_clear_action_bit(action, pe_action_runnable); do_crm_log(warn_level, "Action %s on %s is unrunnable (offline)", action->uuid, action->node->details->uname); if (is_set(action->rsc->flags, pe_rsc_managed) && save_action && a_task == stop_rsc && action->node->details->unclean == FALSE) { pe_fence_node(data_set, action->node, "resource actions are unrunnable"); } } else if (action->node->details->pending) { pe_clear_action_bit(action, pe_action_runnable); do_crm_log(warn_level, "Action %s on %s is unrunnable (pending)", action->uuid, action->node->details->uname); } else if (action->needs == rsc_req_nothing) { pe_rsc_trace(rsc, "Action %s does not require anything", action->uuid); pe_action_set_reason(action, NULL, TRUE); if (pe__is_guest_node(action->node) && !pe_can_fence(data_set, action->node)) { /* An action that requires nothing usually does not require any * fencing in order to be runnable. However, there is an * exception: an action cannot be completed if it is on a guest * node whose host is unclean and cannot be fenced. */ pe_clear_action_bit(action, pe_action_runnable); crm_debug("%s\t%s (cancelled : host cannot be fenced)", action->node->details->uname, action->uuid); } else { pe_set_action_bit(action, pe_action_runnable); } #if 0 /* * No point checking this * - if we don't have quorum we can't stonith anyway */ } else if (action->needs == rsc_req_stonith) { crm_trace("Action %s requires only stonith", action->uuid); action->runnable = TRUE; #endif } else if (is_set(data_set->flags, pe_flag_have_quorum) == FALSE && data_set->no_quorum_policy == no_quorum_stop) { pe_action_set_flag_reason(__FUNCTION__, __LINE__, action, NULL, "no quorum", pe_action_runnable, TRUE); crm_debug("%s\t%s (cancelled : quorum)", action->node->details->uname, action->uuid); } else if (is_set(data_set->flags, pe_flag_have_quorum) == FALSE && data_set->no_quorum_policy == no_quorum_freeze) { pe_rsc_trace(rsc, "Check resource is already active: %s %s %s %s", rsc->id, action->uuid, role2text(rsc->next_role), role2text(rsc->role)); if (rsc->fns->active(rsc, TRUE) == FALSE || rsc->next_role > rsc->role) { pe_action_set_flag_reason(__FUNCTION__, __LINE__, action, NULL, "quorum freeze", pe_action_runnable, TRUE); pe_rsc_debug(rsc, "%s\t%s (cancelled : quorum freeze)", action->node->details->uname, action->uuid); } } else if(is_not_set(action->flags, pe_action_runnable)) { pe_rsc_trace(rsc, "Action %s is runnable", action->uuid); //pe_action_set_reason(action, NULL, TRUE); pe_set_action_bit(action, pe_action_runnable); } if (save_action) { switch (a_task) { case stop_rsc: set_bit(rsc->flags, pe_rsc_stopping); break; case start_rsc: clear_bit(rsc->flags, pe_rsc_starting); if (is_set(action->flags, pe_action_runnable)) { set_bit(rsc->flags, pe_rsc_starting); } break; default: break; } } } free(key); return action; } static const char * unpack_operation_on_fail(action_t * action) { const char *value = g_hash_table_lookup(action->meta, XML_OP_ATTR_ON_FAIL); if (safe_str_eq(action->task, CRMD_ACTION_STOP) && safe_str_eq(value, "standby")) { crm_config_err("on-fail=standby is not allowed for stop actions: %s", action->rsc->id); return NULL; } else if (safe_str_eq(action->task, CRMD_ACTION_DEMOTE) && !value) { /* demote on_fail defaults to master monitor value if present */ xmlNode *operation = NULL; const char *name = NULL; const char *role = NULL; const char *on_fail = NULL; const char *interval_spec = NULL; const char *enabled = NULL; CRM_CHECK(action->rsc != NULL, return NULL); for (operation = __xml_first_child_element(action->rsc->ops_xml); operation && !value; operation = __xml_next_element(operation)) { if (!crm_str_eq((const char *)operation->name, "op", TRUE)) { continue; } name = crm_element_value(operation, "name"); role = crm_element_value(operation, "role"); on_fail = crm_element_value(operation, XML_OP_ATTR_ON_FAIL); enabled = crm_element_value(operation, "enabled"); interval_spec = crm_element_value(operation, XML_LRM_ATTR_INTERVAL); if (!on_fail) { continue; } else if (enabled && !crm_is_true(enabled)) { continue; } else if (safe_str_neq(name, "monitor") || safe_str_neq(role, "Master")) { continue; } else if (crm_parse_interval_spec(interval_spec) == 0) { continue; } value = on_fail; } } return value; } static xmlNode * find_min_interval_mon(resource_t * rsc, gboolean include_disabled) { guint interval_ms = 0; guint min_interval_ms = G_MAXUINT; const char *name = NULL; const char *value = NULL; const char *interval_spec = NULL; xmlNode *op = NULL; xmlNode *operation = NULL; for (operation = __xml_first_child_element(rsc->ops_xml); operation != NULL; operation = __xml_next_element(operation)) { if (crm_str_eq((const char *)operation->name, "op", TRUE)) { name = crm_element_value(operation, "name"); interval_spec = crm_element_value(operation, XML_LRM_ATTR_INTERVAL); value = crm_element_value(operation, "enabled"); if (!include_disabled && value && crm_is_true(value) == FALSE) { continue; } if (safe_str_neq(name, RSC_STATUS)) { continue; } interval_ms = crm_parse_interval_spec(interval_spec); if (interval_ms && (interval_ms < min_interval_ms)) { min_interval_ms = interval_ms; op = operation; } } } return op; } static int unpack_start_delay(const char *value, GHashTable *meta) { int start_delay = 0; if (value != NULL) { start_delay = crm_get_msec(value); if (start_delay < 0) { start_delay = 0; } if (meta) { g_hash_table_replace(meta, strdup(XML_OP_ATTR_START_DELAY), crm_itoa(start_delay)); } } return start_delay; } // true if value contains valid, non-NULL interval origin for recurring op static bool unpack_interval_origin(const char *value, xmlNode *xml_obj, guint interval_ms, crm_time_t *now, long long *start_delay) { long long result = 0; guint interval_sec = interval_ms / 1000; crm_time_t *origin = NULL; // Ignore unspecified values and non-recurring operations if ((value == NULL) || (interval_ms == 0) || (now == NULL)) { return false; } // Parse interval origin from text origin = crm_time_new(value); if (origin == NULL) { crm_config_err("Operation '%s' contains invalid " XML_OP_ATTR_ORIGIN " '%s'", (ID(xml_obj)? ID(xml_obj) : "(unspecified)"), value); return false; } // Get seconds since origin (negative if origin is in the future) result = crm_time_get_seconds(now) - crm_time_get_seconds(origin); crm_time_free(origin); // Calculate seconds from closest interval to now result = result % interval_sec; // Calculate seconds remaining until next interval result = ((result <= 0)? 0 : interval_sec) - result; crm_info("Calculated a start delay of %llds for operation '%s'", result, (ID(xml_obj)? ID(xml_obj) : "(unspecified)")); if (start_delay != NULL) { *start_delay = result * 1000; // milliseconds } return true; } static int unpack_timeout(const char *value) { int timeout = crm_get_msec(value); if (timeout < 0) { timeout = crm_get_msec(CRM_DEFAULT_OP_TIMEOUT_S); } return timeout; } int pe_get_configured_timeout(resource_t *rsc, const char *action, pe_working_set_t *data_set) { xmlNode *child = NULL; const char *timeout = NULL; int timeout_ms = 0; for (child = first_named_child(rsc->ops_xml, XML_ATTR_OP); child != NULL; child = crm_next_same_xml(child)) { if (safe_str_eq(action, crm_element_value(child, XML_NVPAIR_ATTR_NAME))) { timeout = crm_element_value(child, XML_ATTR_TIMEOUT); break; } } if (timeout == NULL && data_set->op_defaults) { GHashTable *action_meta = crm_str_table_new(); pe__unpack_dataset_nvpairs(data_set->op_defaults, XML_TAG_META_SETS, NULL, action_meta, NULL, FALSE, data_set); timeout = g_hash_table_lookup(action_meta, XML_ATTR_TIMEOUT); } // @TODO check meta-attributes (including versioned meta-attributes) // @TODO maybe use min-interval monitor timeout as default for monitors timeout_ms = crm_get_msec(timeout); if (timeout_ms < 0) { timeout_ms = crm_get_msec(CRM_DEFAULT_OP_TIMEOUT_S); } return timeout_ms; } #if ENABLE_VERSIONED_ATTRS static void unpack_versioned_meta(xmlNode *versioned_meta, xmlNode *xml_obj, guint interval_ms, crm_time_t *now) { xmlNode *attrs = NULL; xmlNode *attr = NULL; for (attrs = __xml_first_child_element(versioned_meta); attrs != NULL; attrs = __xml_next_element(attrs)) { for (attr = __xml_first_child_element(attrs); attr != NULL; attr = __xml_next_element(attr)) { const char *name = crm_element_value(attr, XML_NVPAIR_ATTR_NAME); const char *value = crm_element_value(attr, XML_NVPAIR_ATTR_VALUE); if (safe_str_eq(name, XML_OP_ATTR_START_DELAY)) { int start_delay = unpack_start_delay(value, NULL); crm_xml_add_int(attr, XML_NVPAIR_ATTR_VALUE, start_delay); } else if (safe_str_eq(name, XML_OP_ATTR_ORIGIN)) { long long start_delay = 0; if (unpack_interval_origin(value, xml_obj, interval_ms, now, &start_delay)) { crm_xml_add(attr, XML_NVPAIR_ATTR_NAME, XML_OP_ATTR_START_DELAY); crm_xml_add_ll(attr, XML_NVPAIR_ATTR_VALUE, start_delay); } } else if (safe_str_eq(name, XML_ATTR_TIMEOUT)) { int timeout = unpack_timeout(value); crm_xml_add_int(attr, XML_NVPAIR_ATTR_VALUE, timeout); } } } } #endif /*! * \brief Unpack operation XML into an action structure * * Unpack an operation's meta-attributes (normalizing the interval, timeout, * and start delay values as integer milliseconds), requirements, and * failure policy. * * \param[in,out] action Action to unpack into * \param[in] xml_obj Operation XML (or NULL if all defaults) * \param[in] container Resource that contains affected resource, if any * \param[in] data_set Cluster state */ void unpack_operation(action_t * action, xmlNode * xml_obj, resource_t * container, pe_working_set_t * data_set) { guint interval_ms = 0; int timeout = 0; char *value_ms = NULL; const char *value = NULL; const char *field = NULL; char *default_timeout = NULL; #if ENABLE_VERSIONED_ATTRS pe_rsc_action_details_t *rsc_details = NULL; #endif CRM_CHECK(action && action->rsc, return); // Cluster-wide pe__unpack_dataset_nvpairs(data_set->op_defaults, XML_TAG_META_SETS, NULL, action->meta, NULL, FALSE, data_set); // Probe timeouts default differently, so handle timeout default later default_timeout = g_hash_table_lookup(action->meta, XML_ATTR_TIMEOUT); if (default_timeout) { default_timeout = strdup(default_timeout); g_hash_table_remove(action->meta, XML_ATTR_TIMEOUT); } if (xml_obj) { xmlAttrPtr xIter = NULL; // take precedence over defaults pe__unpack_dataset_nvpairs(xml_obj, XML_TAG_META_SETS, NULL, action->meta, NULL, TRUE, data_set); #if ENABLE_VERSIONED_ATTRS rsc_details = pe_rsc_action_details(action); pe_unpack_versioned_attributes(data_set->input, xml_obj, XML_TAG_ATTR_SETS, NULL, rsc_details->versioned_parameters, data_set->now, NULL); pe_unpack_versioned_attributes(data_set->input, xml_obj, XML_TAG_META_SETS, NULL, rsc_details->versioned_meta, data_set->now, NULL); #endif /* Anything set as an XML property has highest precedence. * This ensures we use the name and interval from the tag. */ for (xIter = xml_obj->properties; xIter; xIter = xIter->next) { const char *prop_name = (const char *)xIter->name; const char *prop_value = crm_element_value(xml_obj, prop_name); g_hash_table_replace(action->meta, strdup(prop_name), strdup(prop_value)); } } g_hash_table_remove(action->meta, "id"); // Normalize interval to milliseconds field = XML_LRM_ATTR_INTERVAL; value = g_hash_table_lookup(action->meta, field); if (value != NULL) { interval_ms = crm_parse_interval_spec(value); } else if ((xml_obj == NULL) && !strcmp(action->task, RSC_STATUS)) { /* An orphaned recurring monitor will not have any XML. However, we * want the interval to be set, so the action can be properly detected * as a recurring monitor. Parse it from the key in this case. */ parse_op_key(action->uuid, NULL, NULL, &interval_ms); } if (interval_ms > 0) { value_ms = crm_strdup_printf("%u", interval_ms); g_hash_table_replace(action->meta, strdup(field), value_ms); } else if (value) { g_hash_table_remove(action->meta, field); } // Handle timeout default, now that we know the interval if (g_hash_table_lookup(action->meta, XML_ATTR_TIMEOUT)) { free(default_timeout); } else { // Probe timeouts default to minimum-interval monitor's if (safe_str_eq(action->task, RSC_STATUS) && (interval_ms == 0)) { xmlNode *min_interval_mon = find_min_interval_mon(action->rsc, FALSE); if (min_interval_mon) { value = crm_element_value(min_interval_mon, XML_ATTR_TIMEOUT); if (value) { crm_trace("\t%s defaults to minimum-interval monitor's timeout '%s'", action->uuid, value); free(default_timeout); default_timeout = strdup(value); } } } if (default_timeout) { g_hash_table_insert(action->meta, strdup(XML_ATTR_TIMEOUT), default_timeout); } } if (safe_str_neq(action->task, RSC_START) && safe_str_neq(action->task, RSC_PROMOTE)) { action->needs = rsc_req_nothing; value = "nothing (not start/promote)"; } else if (is_set(action->rsc->flags, pe_rsc_needs_fencing)) { action->needs = rsc_req_stonith; value = "fencing (resource)"; } else if (is_set(action->rsc->flags, pe_rsc_needs_quorum)) { action->needs = rsc_req_quorum; value = "quorum (resource)"; } else { action->needs = rsc_req_nothing; value = "nothing (resource)"; } pe_rsc_trace(action->rsc, "\tAction %s requires: %s", action->uuid, value); value = unpack_operation_on_fail(action); if (value == NULL) { } else if (safe_str_eq(value, "block")) { action->on_fail = action_fail_block; g_hash_table_insert(action->meta, strdup(XML_OP_ATTR_ON_FAIL), strdup("block")); value = "block"; // The above could destroy the original string } else if (safe_str_eq(value, "fence")) { action->on_fail = action_fail_fence; value = "node fencing"; if (is_set(data_set->flags, pe_flag_stonith_enabled) == FALSE) { crm_config_err("Specifying on_fail=fence and" " stonith-enabled=false makes no sense"); action->on_fail = action_fail_stop; action->fail_role = RSC_ROLE_STOPPED; value = "stop resource"; } } else if (safe_str_eq(value, "standby")) { action->on_fail = action_fail_standby; value = "node standby"; } else if (safe_str_eq(value, "ignore") || safe_str_eq(value, "nothing")) { action->on_fail = action_fail_ignore; value = "ignore"; } else if (safe_str_eq(value, "migrate")) { action->on_fail = action_fail_migrate; value = "force migration"; } else if (safe_str_eq(value, "stop")) { action->on_fail = action_fail_stop; action->fail_role = RSC_ROLE_STOPPED; value = "stop resource"; } else if (safe_str_eq(value, "restart")) { action->on_fail = action_fail_recover; value = "restart (and possibly migrate)"; } else if (safe_str_eq(value, "restart-container")) { if (container) { action->on_fail = action_fail_restart_container; value = "restart container (and possibly migrate)"; } else { value = NULL; } } else { pe_err("Resource %s: Unknown failure type (%s)", action->rsc->id, value); value = NULL; } /* defaults */ if (value == NULL && container) { action->on_fail = action_fail_restart_container; value = "restart container (and possibly migrate) (default)"; /* For remote nodes, ensure that any failure that results in dropping an * active connection to the node results in fencing of the node. * * There are only two action failures that don't result in fencing. * 1. probes - probe failures are expected. * 2. start - a start failure indicates that an active connection does not already * exist. The user can set op on-fail=fence if they really want to fence start * failures. */ } else if (((value == NULL) || !is_set(action->rsc->flags, pe_rsc_managed)) && (pe__resource_is_remote_conn(action->rsc, data_set) && !(safe_str_eq(action->task, CRMD_ACTION_STATUS) && (interval_ms == 0)) && (safe_str_neq(action->task, CRMD_ACTION_START)))) { if (!is_set(action->rsc->flags, pe_rsc_managed)) { action->on_fail = action_fail_stop; action->fail_role = RSC_ROLE_STOPPED; value = "stop unmanaged remote node (enforcing default)"; } else { if (is_set(data_set->flags, pe_flag_stonith_enabled)) { value = "fence remote node (default)"; } else { value = "recover remote node connection (default)"; } if (action->rsc->remote_reconnect_ms) { action->fail_role = RSC_ROLE_STOPPED; } action->on_fail = action_fail_reset_remote; } } else if (value == NULL && safe_str_eq(action->task, CRMD_ACTION_STOP)) { if (is_set(data_set->flags, pe_flag_stonith_enabled)) { action->on_fail = action_fail_fence; value = "resource fence (default)"; } else { action->on_fail = action_fail_block; value = "resource block (default)"; } } else if (value == NULL) { action->on_fail = action_fail_recover; value = "restart (and possibly migrate) (default)"; } pe_rsc_trace(action->rsc, "\t%s failure handling: %s", action->task, value); value = NULL; if (xml_obj != NULL) { value = g_hash_table_lookup(action->meta, "role_after_failure"); if (value) { pe_warn_once(pe_wo_role_after, "Support for role_after_failure is deprecated and will be removed in a future release"); } } if (value != NULL && action->fail_role == RSC_ROLE_UNKNOWN) { action->fail_role = text2role(value); } /* defaults */ if (action->fail_role == RSC_ROLE_UNKNOWN) { if (safe_str_eq(action->task, CRMD_ACTION_PROMOTE)) { action->fail_role = RSC_ROLE_SLAVE; } else { action->fail_role = RSC_ROLE_STARTED; } } pe_rsc_trace(action->rsc, "\t%s failure results in: %s", action->task, role2text(action->fail_role)); value = g_hash_table_lookup(action->meta, XML_OP_ATTR_START_DELAY); if (value) { unpack_start_delay(value, action->meta); } else { long long start_delay = 0; value = g_hash_table_lookup(action->meta, XML_OP_ATTR_ORIGIN); if (unpack_interval_origin(value, xml_obj, interval_ms, data_set->now, &start_delay)) { g_hash_table_replace(action->meta, strdup(XML_OP_ATTR_START_DELAY), crm_strdup_printf("%lld", start_delay)); } } value = g_hash_table_lookup(action->meta, XML_ATTR_TIMEOUT); timeout = unpack_timeout(value); g_hash_table_replace(action->meta, strdup(XML_ATTR_TIMEOUT), crm_itoa(timeout)); #if ENABLE_VERSIONED_ATTRS unpack_versioned_meta(rsc_details->versioned_meta, xml_obj, interval_ms, data_set->now); #endif } static xmlNode * find_rsc_op_entry_helper(resource_t * rsc, const char *key, gboolean include_disabled) { guint interval_ms = 0; gboolean do_retry = TRUE; char *local_key = NULL; const char *name = NULL; const char *value = NULL; const char *interval_spec = NULL; char *match_key = NULL; xmlNode *op = NULL; xmlNode *operation = NULL; retry: for (operation = __xml_first_child_element(rsc->ops_xml); operation != NULL; operation = __xml_next_element(operation)) { if (crm_str_eq((const char *)operation->name, "op", TRUE)) { name = crm_element_value(operation, "name"); interval_spec = crm_element_value(operation, XML_LRM_ATTR_INTERVAL); value = crm_element_value(operation, "enabled"); if (!include_disabled && value && crm_is_true(value) == FALSE) { continue; } interval_ms = crm_parse_interval_spec(interval_spec); match_key = generate_op_key(rsc->id, name, interval_ms); if (safe_str_eq(key, match_key)) { op = operation; } free(match_key); if (rsc->clone_name) { match_key = generate_op_key(rsc->clone_name, name, interval_ms); if (safe_str_eq(key, match_key)) { op = operation; } free(match_key); } if (op != NULL) { free(local_key); return op; } } } free(local_key); if (do_retry == FALSE) { return NULL; } do_retry = FALSE; if (strstr(key, CRMD_ACTION_MIGRATE) || strstr(key, CRMD_ACTION_MIGRATED)) { local_key = generate_op_key(rsc->id, "migrate", 0); key = local_key; goto retry; } else if (strstr(key, "_notify_")) { local_key = generate_op_key(rsc->id, "notify", 0); key = local_key; goto retry; } return NULL; } xmlNode * find_rsc_op_entry(resource_t * rsc, const char *key) { return find_rsc_op_entry_helper(rsc, key, FALSE); } void print_node(const char *pre_text, node_t * node, gboolean details) { if (node == NULL) { crm_trace("%s%s: ", pre_text == NULL ? "" : pre_text, pre_text == NULL ? "" : ": "); return; } CRM_ASSERT(node->details); crm_trace("%s%s%sNode %s: (weight=%d, fixed=%s)", pre_text == NULL ? "" : pre_text, pre_text == NULL ? "" : ": ", node->details->online ? "" : "Unavailable/Unclean ", node->details->uname, node->weight, node->fixed ? "True" : "False"); if (details) { + int log_level = LOG_TRACE; + char *pe_mutable = strdup("\t\t"); GListPtr gIter = node->details->running_rsc; crm_trace("\t\t===Node Attributes"); g_hash_table_foreach(node->details->attrs, print_str_str, pe_mutable); free(pe_mutable); crm_trace("\t\t=== Resources"); for (; gIter != NULL; gIter = gIter->next) { - resource_t *rsc = (resource_t *) gIter->data; + pe_resource_t *rsc = (pe_resource_t *) gIter->data; - print_resource(LOG_TRACE, "\t\t", rsc, FALSE); + rsc->fns->print(rsc, "\t\t", pe_print_log|pe_print_pending, + &log_level); } } } /* * Used by the HashTable for-loop */ void print_str_str(gpointer key, gpointer value, gpointer user_data) { crm_trace("%s%s %s ==> %s", user_data == NULL ? "" : (char *)user_data, user_data == NULL ? "" : ": ", (char *)key, (char *)value); } -void -print_resource(int log_level, const char *pre_text, resource_t * rsc, gboolean details) -{ - long options = pe_print_log | pe_print_pending; - - if (rsc == NULL) { - do_crm_log(log_level - 1, "%s%s: ", - pre_text == NULL ? "" : pre_text, pre_text == NULL ? "" : ": "); - return; - } - if (details) { - options |= pe_print_details; - } - rsc->fns->print(rsc, pre_text, options, &log_level); -} - void pe_free_action(action_t * action) { if (action == NULL) { return; } g_list_free_full(action->actions_before, free); /* action_wrapper_t* */ g_list_free_full(action->actions_after, free); /* action_wrapper_t* */ if (action->extra) { g_hash_table_destroy(action->extra); } if (action->meta) { g_hash_table_destroy(action->meta); } #if ENABLE_VERSIONED_ATTRS if (action->rsc) { pe_free_rsc_action_details(action); } #endif free(action->cancel_task); free(action->reason); free(action->task); free(action->uuid); free(action->node); free(action); } GListPtr find_recurring_actions(GListPtr input, node_t * not_on_node) { const char *value = NULL; GListPtr result = NULL; GListPtr gIter = input; CRM_CHECK(input != NULL, return NULL); for (; gIter != NULL; gIter = gIter->next) { action_t *action = (action_t *) gIter->data; value = g_hash_table_lookup(action->meta, XML_LRM_ATTR_INTERVAL_MS); if (value == NULL) { /* skip */ } else if (safe_str_eq(value, "0")) { /* skip */ } else if (safe_str_eq(CRMD_ACTION_CANCEL, action->task)) { /* skip */ } else if (not_on_node == NULL) { crm_trace("(null) Found: %s", action->uuid); result = g_list_prepend(result, action); } else if (action->node == NULL) { /* skip */ } else if (action->node->details != not_on_node->details) { crm_trace("Found: %s", action->uuid); result = g_list_prepend(result, action); } } return result; } enum action_tasks get_complex_task(resource_t * rsc, const char *name, gboolean allow_non_atomic) { enum action_tasks task = text2task(name); if (rsc == NULL) { return task; } else if (allow_non_atomic == FALSE || rsc->variant == pe_native) { switch (task) { case stopped_rsc: case started_rsc: case action_demoted: case action_promoted: crm_trace("Folding %s back into its atomic counterpart for %s", name, rsc->id); return task - 1; break; default: break; } } return task; } action_t * find_first_action(GListPtr input, const char *uuid, const char *task, node_t * on_node) { GListPtr gIter = NULL; CRM_CHECK(uuid || task, return NULL); for (gIter = input; gIter != NULL; gIter = gIter->next) { action_t *action = (action_t *) gIter->data; if (uuid != NULL && safe_str_neq(uuid, action->uuid)) { continue; } else if (task != NULL && safe_str_neq(task, action->task)) { continue; } else if (on_node == NULL) { return action; } else if (action->node == NULL) { continue; } else if (on_node->details == action->node->details) { return action; } } return NULL; } GListPtr find_actions(GListPtr input, const char *key, const node_t *on_node) { GListPtr gIter = input; GListPtr result = NULL; CRM_CHECK(key != NULL, return NULL); for (; gIter != NULL; gIter = gIter->next) { action_t *action = (action_t *) gIter->data; if (safe_str_neq(key, action->uuid)) { crm_trace("%s does not match action %s", key, action->uuid); continue; } else if (on_node == NULL) { crm_trace("Action %s matches (ignoring node)", key); result = g_list_prepend(result, action); } else if (action->node == NULL) { crm_trace("Action %s matches (unallocated, assigning to %s)", key, on_node->details->uname); action->node = node_copy(on_node); result = g_list_prepend(result, action); } else if (on_node->details == action->node->details) { crm_trace("Action %s on %s matches", key, on_node->details->uname); result = g_list_prepend(result, action); } else { crm_trace("Action %s on node %s does not match requested node %s", key, action->node->details->uname, on_node->details->uname); } } return result; } GList * find_actions_exact(GList *input, const char *key, const pe_node_t *on_node) { GList *result = NULL; CRM_CHECK(key != NULL, return NULL); if (on_node == NULL) { crm_trace("Not searching for action %s because node not specified", key); return NULL; } for (GList *gIter = input; gIter != NULL; gIter = gIter->next) { pe_action_t *action = (pe_action_t *) gIter->data; if (action->node == NULL) { crm_trace("Skipping comparison of %s vs action %s without node", key, action->uuid); } else if (safe_str_neq(key, action->uuid)) { crm_trace("Desired action %s doesn't match %s", key, action->uuid); } else if (safe_str_neq(on_node->details->id, action->node->details->id)) { crm_trace("Action %s desired node ID %s doesn't match %s", key, on_node->details->id, action->node->details->id); } else { crm_trace("Action %s matches", key); result = g_list_prepend(result, action); } } return result; } /*! * \brief Find all actions of given type for a resource * * \param[in] rsc Resource to search * \param[in] node Find only actions scheduled on this node * \param[in] task Action name to search for * \param[in] require_node If TRUE, NULL node or action node will not match * * \return List of actions found (or NULL if none) * \note If node is not NULL and require_node is FALSE, matching actions * without a node will be assigned to node. */ GList * pe__resource_actions(const pe_resource_t *rsc, const pe_node_t *node, const char *task, bool require_node) { GList *result = NULL; char *key = generate_op_key(rsc->id, task, 0); if (require_node) { result = find_actions_exact(rsc->actions, key, node); } else { result = find_actions(rsc->actions, key, node); } free(key); return result; } static void resource_node_score(resource_t * rsc, node_t * node, int score, const char *tag) { node_t *match = NULL; if ((rsc->exclusive_discover || (node->rsc_discover_mode == pe_discover_never)) && safe_str_eq(tag, "symmetric_default")) { /* This string comparision may be fragile, but exclusive resources and * exclusive nodes should not have the symmetric_default constraint * applied to them. */ return; } else if (rsc->children) { GListPtr gIter = rsc->children; for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; resource_node_score(child_rsc, node, score, tag); } } pe_rsc_trace(rsc, "Setting %s for %s on %s: %d", tag, rsc->id, node->details->uname, score); match = pe_hash_table_lookup(rsc->allowed_nodes, node->details->id); if (match == NULL) { match = node_copy(node); g_hash_table_insert(rsc->allowed_nodes, (gpointer) match->details->id, match); } match->weight = merge_weights(match->weight, score); } void resource_location(resource_t * rsc, node_t * node, int score, const char *tag, pe_working_set_t * data_set) { if (node != NULL) { resource_node_score(rsc, node, score, tag); } else if (data_set != NULL) { GListPtr gIter = data_set->nodes; for (; gIter != NULL; gIter = gIter->next) { node_t *node_iter = (node_t *) gIter->data; resource_node_score(rsc, node_iter, score, tag); } } else { GHashTableIter iter; node_t *node_iter = NULL; g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&node_iter)) { resource_node_score(rsc, node_iter, score, tag); } } if (node == NULL && score == -INFINITY) { if (rsc->allocated_to) { crm_info("Deallocating %s from %s", rsc->id, rsc->allocated_to->details->uname); free(rsc->allocated_to); rsc->allocated_to = NULL; } } } #define sort_return(an_int, why) do { \ free(a_uuid); \ free(b_uuid); \ crm_trace("%s (%d) %c %s (%d) : %s", \ a_xml_id, a_call_id, an_int>0?'>':an_int<0?'<':'=', \ b_xml_id, b_call_id, why); \ return an_int; \ } while(0) gint sort_op_by_callid(gconstpointer a, gconstpointer b) { int a_call_id = -1; int b_call_id = -1; char *a_uuid = NULL; char *b_uuid = NULL; const xmlNode *xml_a = a; const xmlNode *xml_b = b; const char *a_xml_id = crm_element_value(xml_a, XML_ATTR_ID); const char *b_xml_id = crm_element_value(xml_b, XML_ATTR_ID); if (safe_str_eq(a_xml_id, b_xml_id)) { /* We have duplicate lrm_rsc_op entries in the status * section which is unlikely to be a good thing * - we can handle it easily enough, but we need to get * to the bottom of why it's happening. */ pe_err("Duplicate lrm_rsc_op entries named %s", a_xml_id); sort_return(0, "duplicate"); } crm_element_value_int(xml_a, XML_LRM_ATTR_CALLID, &a_call_id); crm_element_value_int(xml_b, XML_LRM_ATTR_CALLID, &b_call_id); if (a_call_id == -1 && b_call_id == -1) { /* both are pending ops so it doesn't matter since * stops are never pending */ sort_return(0, "pending"); } else if (a_call_id >= 0 && a_call_id < b_call_id) { sort_return(-1, "call id"); } else if (b_call_id >= 0 && a_call_id > b_call_id) { sort_return(1, "call id"); } else if (b_call_id >= 0 && a_call_id == b_call_id) { /* * The op and last_failed_op are the same * Order on last-rc-change */ time_t last_a = -1; time_t last_b = -1; crm_element_value_epoch(xml_a, XML_RSC_OP_LAST_CHANGE, &last_a); crm_element_value_epoch(xml_b, XML_RSC_OP_LAST_CHANGE, &last_b); crm_trace("rc-change: %lld vs %lld", (long long) last_a, (long long) last_b); if (last_a >= 0 && last_a < last_b) { sort_return(-1, "rc-change"); } else if (last_b >= 0 && last_a > last_b) { sort_return(1, "rc-change"); } sort_return(0, "rc-change"); } else { /* One of the inputs is a pending operation * Attempt to use XML_ATTR_TRANSITION_MAGIC to determine its age relative to the other */ int a_id = -1; int b_id = -1; const char *a_magic = crm_element_value(xml_a, XML_ATTR_TRANSITION_MAGIC); const char *b_magic = crm_element_value(xml_b, XML_ATTR_TRANSITION_MAGIC); CRM_CHECK(a_magic != NULL && b_magic != NULL, sort_return(0, "No magic")); if (!decode_transition_magic(a_magic, &a_uuid, &a_id, NULL, NULL, NULL, NULL)) { sort_return(0, "bad magic a"); } if (!decode_transition_magic(b_magic, &b_uuid, &b_id, NULL, NULL, NULL, NULL)) { sort_return(0, "bad magic b"); } /* try to determine the relative age of the operation... * some pending operations (e.g. a start) may have been superseded * by a subsequent stop * * [a|b]_id == -1 means it's a shutdown operation and _always_ comes last */ if (safe_str_neq(a_uuid, b_uuid) || a_id == b_id) { /* * some of the logic in here may be redundant... * * if the UUID from the TE doesn't match then one better * be a pending operation. * pending operations don't survive between elections and joins * because we query the LRM directly */ if (b_call_id == -1) { sort_return(-1, "transition + call"); } else if (a_call_id == -1) { sort_return(1, "transition + call"); } } else if ((a_id >= 0 && a_id < b_id) || b_id == -1) { sort_return(-1, "transition"); } else if ((b_id >= 0 && a_id > b_id) || a_id == -1) { sort_return(1, "transition"); } } /* we should never end up here */ CRM_CHECK(FALSE, sort_return(0, "default")); } time_t get_effective_time(pe_working_set_t * data_set) { if(data_set) { if (data_set->now == NULL) { crm_trace("Recording a new 'now'"); data_set->now = crm_time_new(NULL); } return crm_time_get_seconds_since_epoch(data_set->now); } crm_trace("Defaulting to 'now'"); return time(NULL); } gboolean get_target_role(resource_t * rsc, enum rsc_role_e * role) { enum rsc_role_e local_role = RSC_ROLE_UNKNOWN; const char *value = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_TARGET_ROLE); CRM_CHECK(role != NULL, return FALSE); if (value == NULL || safe_str_eq("started", value) || safe_str_eq("default", value)) { return FALSE; } local_role = text2role(value); if (local_role == RSC_ROLE_UNKNOWN) { crm_config_err("%s: Unknown value for %s: %s", rsc->id, XML_RSC_ATTR_TARGET_ROLE, value); return FALSE; } else if (local_role > RSC_ROLE_STARTED) { if (is_set(uber_parent(rsc)->flags, pe_rsc_promotable)) { if (local_role > RSC_ROLE_SLAVE) { /* This is what we'd do anyway, just leave the default to avoid messing up the placement algorithm */ return FALSE; } } else { crm_config_err("%s is not part of a promotable clone resource, a %s of '%s' makes no sense", rsc->id, XML_RSC_ATTR_TARGET_ROLE, value); return FALSE; } } *role = local_role; return TRUE; } gboolean order_actions(action_t * lh_action, action_t * rh_action, enum pe_ordering order) { GListPtr gIter = NULL; action_wrapper_t *wrapper = NULL; GListPtr list = NULL; if (order == pe_order_none) { return FALSE; } if (lh_action == NULL || rh_action == NULL) { return FALSE; } crm_trace("Ordering Action %s before %s", lh_action->uuid, rh_action->uuid); /* Ensure we never create a dependency on ourselves... it's happened */ CRM_ASSERT(lh_action != rh_action); /* Filter dups, otherwise update_action_states() has too much work to do */ gIter = lh_action->actions_after; for (; gIter != NULL; gIter = gIter->next) { action_wrapper_t *after = (action_wrapper_t *) gIter->data; if (after->action == rh_action && (after->type & order)) { return FALSE; } } wrapper = calloc(1, sizeof(action_wrapper_t)); wrapper->action = rh_action; wrapper->type = order; list = lh_action->actions_after; list = g_list_prepend(list, wrapper); lh_action->actions_after = list; wrapper = NULL; /* order |= pe_order_implies_then; */ /* order ^= pe_order_implies_then; */ wrapper = calloc(1, sizeof(action_wrapper_t)); wrapper->action = lh_action; wrapper->type = order; list = rh_action->actions_before; list = g_list_prepend(list, wrapper); rh_action->actions_before = list; return TRUE; } action_t * get_pseudo_op(const char *name, pe_working_set_t * data_set) { action_t *op = NULL; if(data_set->singletons) { op = g_hash_table_lookup(data_set->singletons, name); } if (op == NULL) { op = custom_action(NULL, strdup(name), name, NULL, TRUE, TRUE, data_set); set_bit(op->flags, pe_action_pseudo); set_bit(op->flags, pe_action_runnable); } return op; } void destroy_ticket(gpointer data) { ticket_t *ticket = data; if (ticket->state) { g_hash_table_destroy(ticket->state); } free(ticket->id); free(ticket); } ticket_t * ticket_new(const char *ticket_id, pe_working_set_t * data_set) { ticket_t *ticket = NULL; if (ticket_id == NULL || strlen(ticket_id) == 0) { return NULL; } if (data_set->tickets == NULL) { data_set->tickets = g_hash_table_new_full(crm_str_hash, g_str_equal, free, destroy_ticket); } ticket = g_hash_table_lookup(data_set->tickets, ticket_id); if (ticket == NULL) { ticket = calloc(1, sizeof(ticket_t)); if (ticket == NULL) { crm_err("Cannot allocate ticket '%s'", ticket_id); return NULL; } crm_trace("Creaing ticket entry for %s", ticket_id); ticket->id = strdup(ticket_id); ticket->granted = FALSE; ticket->last_granted = -1; ticket->standby = FALSE; ticket->state = crm_str_table_new(); g_hash_table_insert(data_set->tickets, strdup(ticket->id), ticket); } return ticket; } static void filter_parameters(xmlNode * param_set, const char *param_string, bool need_present) { if (param_set && param_string) { xmlAttrPtr xIter = param_set->properties; while (xIter) { const char *prop_name = (const char *)xIter->name; char *name = crm_strdup_printf(" %s ", prop_name); char *match = strstr(param_string, name); free(name); // Do now, because current entry might get removed below xIter = xIter->next; if (need_present && match == NULL) { crm_trace("%s not found in %s", prop_name, param_string); xml_remove_prop(param_set, prop_name); } else if (need_present == FALSE && match) { crm_trace("%s found in %s", prop_name, param_string); xml_remove_prop(param_set, prop_name); } } } } #if ENABLE_VERSIONED_ATTRS static void append_versioned_params(xmlNode *versioned_params, const char *ra_version, xmlNode *params) { GHashTable *hash = pe_unpack_versioned_parameters(versioned_params, ra_version); char *key = NULL; char *value = NULL; GHashTableIter iter; g_hash_table_iter_init(&iter, hash); while (g_hash_table_iter_next(&iter, (gpointer *) &key, (gpointer *) &value)) { crm_xml_add(params, key, value); } g_hash_table_destroy(hash); } #endif /*! * \internal * \brief Calculate action digests and store in node's digest cache * * \param[in] rsc Resource that action was for * \param[in] task Name of action performed * \param[in] key Action's task key * \param[in] node Node action was performed on * \param[in] xml_op XML of operation in CIB status (if available) * \param[in] calc_secure Whether to calculate secure digest * \param[in] data_set Cluster working set * * \return Pointer to node's digest cache entry */ static op_digest_cache_t * rsc_action_digest(pe_resource_t *rsc, const char *task, const char *key, pe_node_t *node, xmlNode *xml_op, bool calc_secure, pe_working_set_t *data_set) { op_digest_cache_t *data = NULL; data = g_hash_table_lookup(node->details->digest_cache, key); if (data == NULL) { GHashTable *local_rsc_params = crm_str_table_new(); action_t *action = custom_action(rsc, strdup(key), task, node, TRUE, FALSE, data_set); #if ENABLE_VERSIONED_ATTRS xmlNode *local_versioned_params = create_xml_node(NULL, XML_TAG_RSC_VER_ATTRS); const char *ra_version = NULL; #endif const char *op_version; const char *restart_list = NULL; const char *secure_list = " passwd password "; data = calloc(1, sizeof(op_digest_cache_t)); CRM_ASSERT(data != NULL); get_rsc_attributes(local_rsc_params, rsc, node, data_set); #if ENABLE_VERSIONED_ATTRS pe_get_versioned_attributes(local_versioned_params, rsc, node, data_set); #endif data->params_all = create_xml_node(NULL, XML_TAG_PARAMS); // REMOTE_CONTAINER_HACK: Allow remote nodes that start containers with pacemaker remote inside if (pe__add_bundle_remote_name(rsc, data->params_all, XML_RSC_ATTR_REMOTE_RA_ADDR)) { crm_trace("Set address for bundle connection %s (on %s)", rsc->id, node->details->uname); } g_hash_table_foreach(local_rsc_params, hash2field, data->params_all); g_hash_table_foreach(action->extra, hash2field, data->params_all); g_hash_table_foreach(rsc->parameters, hash2field, data->params_all); g_hash_table_foreach(action->meta, hash2metafield, data->params_all); if(xml_op) { secure_list = crm_element_value(xml_op, XML_LRM_ATTR_OP_SECURE); restart_list = crm_element_value(xml_op, XML_LRM_ATTR_OP_RESTART); op_version = crm_element_value(xml_op, XML_ATTR_CRM_VERSION); #if ENABLE_VERSIONED_ATTRS ra_version = crm_element_value(xml_op, XML_ATTR_RA_VERSION); #endif } else { op_version = CRM_FEATURE_SET; } #if ENABLE_VERSIONED_ATTRS append_versioned_params(local_versioned_params, ra_version, data->params_all); append_versioned_params(rsc->versioned_parameters, ra_version, data->params_all); { pe_rsc_action_details_t *details = pe_rsc_action_details(action); append_versioned_params(details->versioned_parameters, ra_version, data->params_all); } #endif filter_action_parameters(data->params_all, op_version); g_hash_table_destroy(local_rsc_params); pe_free_action(action); data->digest_all_calc = calculate_operation_digest(data->params_all, op_version); if (calc_secure) { data->params_secure = copy_xml(data->params_all); if(secure_list) { filter_parameters(data->params_secure, secure_list, FALSE); } data->digest_secure_calc = calculate_operation_digest(data->params_secure, op_version); } if(xml_op && crm_element_value(xml_op, XML_LRM_ATTR_RESTART_DIGEST) != NULL) { data->params_restart = copy_xml(data->params_all); if (restart_list) { filter_parameters(data->params_restart, restart_list, TRUE); } data->digest_restart_calc = calculate_operation_digest(data->params_restart, op_version); } g_hash_table_insert(node->details->digest_cache, strdup(key), data); } return data; } op_digest_cache_t * rsc_action_digest_cmp(resource_t * rsc, xmlNode * xml_op, node_t * node, pe_working_set_t * data_set) { op_digest_cache_t *data = NULL; char *key = NULL; guint interval_ms = 0; const char *op_version; const char *task = crm_element_value(xml_op, XML_LRM_ATTR_TASK); const char *interval_ms_s = crm_element_value(xml_op, XML_LRM_ATTR_INTERVAL_MS); const char *digest_all; const char *digest_restart; CRM_ASSERT(node != NULL); op_version = crm_element_value(xml_op, XML_ATTR_CRM_VERSION); digest_all = crm_element_value(xml_op, XML_LRM_ATTR_OP_DIGEST); digest_restart = crm_element_value(xml_op, XML_LRM_ATTR_RESTART_DIGEST); interval_ms = crm_parse_ms(interval_ms_s); key = generate_op_key(rsc->id, task, interval_ms); data = rsc_action_digest(rsc, task, key, node, xml_op, is_set(data_set->flags, pe_flag_sanitized), data_set); data->rc = RSC_DIGEST_MATCH; if (digest_restart && data->digest_restart_calc && strcmp(data->digest_restart_calc, digest_restart) != 0) { pe_rsc_info(rsc, "Parameters to %s on %s changed: was %s vs. now %s (restart:%s) %s", key, node->details->uname, crm_str(digest_restart), data->digest_restart_calc, op_version, crm_element_value(xml_op, XML_ATTR_TRANSITION_MAGIC)); data->rc = RSC_DIGEST_RESTART; } else if (digest_all == NULL) { /* it is unknown what the previous op digest was */ data->rc = RSC_DIGEST_UNKNOWN; } else if (strcmp(digest_all, data->digest_all_calc) != 0) { pe_rsc_info(rsc, "Parameters to %s on %s changed: was %s vs. now %s (%s:%s) %s", key, node->details->uname, crm_str(digest_all), data->digest_all_calc, (interval_ms > 0)? "reschedule" : "reload", op_version, crm_element_value(xml_op, XML_ATTR_TRANSITION_MAGIC)); data->rc = RSC_DIGEST_ALL; } free(key); return data; } /*! * \internal * \brief Create an unfencing summary for use in special node attribute * * Create a string combining a fence device's resource ID, agent type, and * parameter digest (whether for all parameters or just non-private parameters). * This can be stored in a special node attribute, allowing us to detect changes * in either the agent type or parameters, to know whether unfencing must be * redone or can be safely skipped when the device's history is cleaned. * * \param[in] rsc_id Fence device resource ID * \param[in] agent_type Fence device agent * \param[in] param_digest Fence device parameter digest * * \return Newly allocated string with unfencing digest * \note The caller is responsible for freeing the result. */ static inline char * create_unfencing_summary(const char *rsc_id, const char *agent_type, const char *param_digest) { return crm_strdup_printf("%s:%s:%s", rsc_id, agent_type, param_digest); } /*! * \internal * \brief Check whether a node can skip unfencing * * Check whether a fence device's current definition matches a node's * stored summary of when it was last unfenced by the device. * * \param[in] rsc_id Fence device's resource ID * \param[in] agent Fence device's agent type * \param[in] digest_calc Fence device's current parameter digest * \param[in] node_summary Value of node's special unfencing node attribute * (a comma-separated list of unfencing summaries for * all devices that have unfenced this node) * * \return TRUE if digest matches, FALSE otherwise */ static bool unfencing_digest_matches(const char *rsc_id, const char *agent, const char *digest_calc, const char *node_summary) { bool matches = FALSE; if (rsc_id && agent && digest_calc && node_summary) { char *search_secure = create_unfencing_summary(rsc_id, agent, digest_calc); /* The digest was calculated including the device ID and agent, * so there is no risk of collision using strstr(). */ matches = (strstr(node_summary, search_secure) != NULL); crm_trace("Calculated unfencing digest '%s' %sfound in '%s'", search_secure, matches? "" : "not ", node_summary); free(search_secure); } return matches; } /* Magic string to use as action name for digest cache entries used for * unfencing checks. This is not a real action name (i.e. "on"), so * check_action_definition() won't confuse these entries with real actions. */ #define STONITH_DIGEST_TASK "stonith-on" /*! * \internal * \brief Calculate fence device digests and digest comparison result * * \param[in] rsc Fence device resource * \param[in] agent Fence device's agent type * \param[in] node Node with digest cache to use * \param[in] data_set Cluster working set * * \return Node's digest cache entry */ static op_digest_cache_t * fencing_action_digest_cmp(pe_resource_t *rsc, const char *agent, pe_node_t *node, pe_working_set_t *data_set) { const char *node_summary = NULL; // Calculate device's current parameter digests char *key = generate_op_key(rsc->id, STONITH_DIGEST_TASK, 0); op_digest_cache_t *data = rsc_action_digest(rsc, STONITH_DIGEST_TASK, key, node, NULL, TRUE, data_set); free(key); // Check whether node has special unfencing summary node attribute node_summary = pe_node_attribute_raw(node, CRM_ATTR_DIGESTS_ALL); if (node_summary == NULL) { data->rc = RSC_DIGEST_UNKNOWN; return data; } // Check whether full parameter digest matches if (unfencing_digest_matches(rsc->id, agent, data->digest_all_calc, node_summary)) { data->rc = RSC_DIGEST_MATCH; return data; } // Check whether secure parameter digest matches node_summary = pe_node_attribute_raw(node, CRM_ATTR_DIGESTS_SECURE); if (unfencing_digest_matches(rsc->id, agent, data->digest_secure_calc, node_summary)) { data->rc = RSC_DIGEST_MATCH; if (is_set(data_set->flags, pe_flag_stdout)) { printf("Only 'private' parameters to %s for unfencing %s changed\n", rsc->id, node->details->uname); } return data; } // Parameters don't match data->rc = RSC_DIGEST_ALL; if (is_set(data_set->flags, (pe_flag_sanitized|pe_flag_stdout)) && data->digest_secure_calc) { char *digest = create_unfencing_summary(rsc->id, agent, data->digest_secure_calc); printf("Parameters to %s for unfencing %s changed, try '%s'\n", rsc->id, node->details->uname, digest); free(digest); } return data; } const char *rsc_printable_id(resource_t *rsc) { if (is_not_set(rsc->flags, pe_rsc_unique)) { return ID(rsc->xml); } return rsc->id; } void clear_bit_recursive(resource_t * rsc, unsigned long long flag) { GListPtr gIter = rsc->children; clear_bit(rsc->flags, flag); for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; clear_bit_recursive(child_rsc, flag); } } void set_bit_recursive(resource_t * rsc, unsigned long long flag) { GListPtr gIter = rsc->children; set_bit(rsc->flags, flag); for (; gIter != NULL; gIter = gIter->next) { resource_t *child_rsc = (resource_t *) gIter->data; set_bit_recursive(child_rsc, flag); } } static GListPtr find_unfencing_devices(GListPtr candidates, GListPtr matches) { for (GListPtr gIter = candidates; gIter != NULL; gIter = gIter->next) { resource_t *candidate = gIter->data; const char *provides = g_hash_table_lookup(candidate->meta, XML_RSC_ATTR_PROVIDES); const char *requires = g_hash_table_lookup(candidate->meta, XML_RSC_ATTR_REQUIRES); if(candidate->children) { matches = find_unfencing_devices(candidate->children, matches); } else if (is_not_set(candidate->flags, pe_rsc_fence_device)) { continue; } else if (crm_str_eq(provides, "unfencing", FALSE) || crm_str_eq(requires, "unfencing", FALSE)) { matches = g_list_prepend(matches, candidate); } } return matches; } action_t * pe_fence_op(node_t * node, const char *op, bool optional, const char *reason, pe_working_set_t * data_set) { char *op_key = NULL; action_t *stonith_op = NULL; if(op == NULL) { op = data_set->stonith_action; } op_key = crm_strdup_printf("%s-%s-%s", CRM_OP_FENCE, node->details->uname, op); if(data_set->singletons) { stonith_op = g_hash_table_lookup(data_set->singletons, op_key); } if(stonith_op == NULL) { stonith_op = custom_action(NULL, op_key, CRM_OP_FENCE, node, TRUE, TRUE, data_set); add_hash_param(stonith_op->meta, XML_LRM_ATTR_TARGET, node->details->uname); add_hash_param(stonith_op->meta, XML_LRM_ATTR_TARGET_UUID, node->details->id); add_hash_param(stonith_op->meta, "stonith_action", op); if (pe__is_guest_or_remote_node(node) && is_set(data_set->flags, pe_flag_enable_unfencing)) { /* Extra work to detect device changes on remotes * * We may do this for all nodes in the future, but for now * the check_action_definition() based stuff works fine. */ long max = 1024; long digests_all_offset = 0; long digests_secure_offset = 0; char *digests_all = calloc(max, sizeof(char)); char *digests_secure = calloc(max, sizeof(char)); GListPtr matches = find_unfencing_devices(data_set->resources, NULL); for (GListPtr gIter = matches; gIter != NULL; gIter = gIter->next) { resource_t *match = gIter->data; const char *agent = g_hash_table_lookup(match->meta, XML_ATTR_TYPE); op_digest_cache_t *data = NULL; data = fencing_action_digest_cmp(match, agent, node, data_set); if(data->rc == RSC_DIGEST_ALL) { optional = FALSE; crm_notice("Unfencing %s (remote): because the definition of %s changed", node->details->uname, match->id); if (is_set(data_set->flags, pe_flag_stdout)) { fprintf(stdout, " notice: Unfencing %s (remote): because the definition of %s changed\n", node->details->uname, match->id); } } digests_all_offset += snprintf( digests_all+digests_all_offset, max-digests_all_offset, "%s:%s:%s,", match->id, agent, data->digest_all_calc); digests_secure_offset += snprintf( digests_secure+digests_secure_offset, max-digests_secure_offset, "%s:%s:%s,", match->id, agent, data->digest_secure_calc); } g_hash_table_insert(stonith_op->meta, strdup(XML_OP_ATTR_DIGESTS_ALL), digests_all); g_hash_table_insert(stonith_op->meta, strdup(XML_OP_ATTR_DIGESTS_SECURE), digests_secure); } } else { free(op_key); } if(optional == FALSE && pe_can_fence(data_set, node)) { pe_action_required(stonith_op, NULL, reason); } else if(reason && stonith_op->reason == NULL) { stonith_op->reason = strdup(reason); } return stonith_op; } void trigger_unfencing( resource_t * rsc, node_t *node, const char *reason, action_t *dependency, pe_working_set_t * data_set) { if(is_not_set(data_set->flags, pe_flag_enable_unfencing)) { /* No resources require it */ return; } else if (rsc != NULL && is_not_set(rsc->flags, pe_rsc_fence_device)) { /* Wasn't a stonith device */ return; } else if(node && node->details->online && node->details->unclean == FALSE && node->details->shutdown == FALSE) { action_t *unfence = pe_fence_op(node, "on", FALSE, reason, data_set); if(dependency) { order_actions(unfence, dependency, pe_order_optional); } } else if(rsc) { GHashTableIter iter; g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&node)) { if(node->details->online && node->details->unclean == FALSE && node->details->shutdown == FALSE) { trigger_unfencing(rsc, node, reason, dependency, data_set); } } } } gboolean add_tag_ref(GHashTable * tags, const char * tag_name, const char * obj_ref) { tag_t *tag = NULL; GListPtr gIter = NULL; gboolean is_existing = FALSE; CRM_CHECK(tags && tag_name && obj_ref, return FALSE); tag = g_hash_table_lookup(tags, tag_name); if (tag == NULL) { tag = calloc(1, sizeof(tag_t)); if (tag == NULL) { return FALSE; } tag->id = strdup(tag_name); tag->refs = NULL; g_hash_table_insert(tags, strdup(tag_name), tag); } for (gIter = tag->refs; gIter != NULL; gIter = gIter->next) { const char *existing_ref = (const char *) gIter->data; if (crm_str_eq(existing_ref, obj_ref, TRUE)){ is_existing = TRUE; break; } } if (is_existing == FALSE) { tag->refs = g_list_append(tag->refs, strdup(obj_ref)); crm_trace("Added: tag=%s ref=%s", tag->id, obj_ref); } return TRUE; } void pe_action_set_flag_reason(const char *function, long line, pe_action_t *action, pe_action_t *reason, const char *text, enum pe_action_flags flags, bool overwrite) { bool unset = FALSE; bool update = FALSE; const char *change = NULL; if(is_set(flags, pe_action_runnable)) { unset = TRUE; change = "unrunnable"; } else if(is_set(flags, pe_action_optional)) { unset = TRUE; change = "required"; } else if(is_set(flags, pe_action_migrate_runnable)) { unset = TRUE; overwrite = TRUE; change = "unrunnable"; } else if(is_set(flags, pe_action_dangle)) { change = "dangling"; } else if(is_set(flags, pe_action_requires_any)) { change = "required"; } else { crm_err("Unknown flag change to %x by %s: 0x%s", flags, action->uuid, (reason? reason->uuid : "0")); } if(unset) { if(is_set(action->flags, flags)) { action->flags = crm_clear_bit(function, line, action->uuid, action->flags, flags); update = TRUE; } } else { if(is_not_set(action->flags, flags)) { action->flags = crm_set_bit(function, line, action->uuid, action->flags, flags); update = TRUE; } } if((change && update) || text) { char *reason_text = NULL; if(reason == NULL) { pe_action_set_reason(action, text, overwrite); } else if(reason->rsc == NULL) { reason_text = crm_strdup_printf("%s %s%c %s", change, reason->task, text?':':0, text?text:""); } else { reason_text = crm_strdup_printf("%s %s %s%c %s", change, reason->rsc->id, reason->task, text?':':0, text?text:"NA"); } if(reason_text && action->rsc != reason->rsc) { pe_action_set_reason(action, reason_text, overwrite); } free(reason_text); } } void pe_action_set_reason(pe_action_t *action, const char *reason, bool overwrite) { if(action->reason && overwrite) { pe_rsc_trace(action->rsc, "Changing %s reason from '%s' to '%s'", action->uuid, action->reason, reason); free(action->reason); action->reason = NULL; } if(action->reason == NULL) { if(reason) { pe_rsc_trace(action->rsc, "Set %s reason to '%s'", action->uuid, reason); action->reason = strdup(reason); } else { action->reason = NULL; } } } /*! * \internal * \brief Check whether shutdown has been requested for a node * * \param[in] node Node to check * * \return TRUE if node has shutdown attribute set and nonzero, FALSE otherwise * \note This differs from simply using node->details->shutdown in that it can * be used before that has been determined (and in fact to determine it), * and it can also be used to distinguish requested shutdown from implicit * shutdown of remote nodes by virtue of their connection stopping. */ bool pe__shutdown_requested(pe_node_t *node) { const char *shutdown = pe_node_attribute_raw(node, XML_CIB_ATTR_SHUTDOWN); return shutdown && strcmp(shutdown, "0"); } /*! * \internal * \brief Update a data set's "recheck by" time * * \param[in] recheck Epoch time when recheck should happen * \param[in,out] data_set Current working set */ void pe__update_recheck_time(time_t recheck, pe_working_set_t *data_set) { if ((recheck > get_effective_time(data_set)) && ((data_set->recheck_by == 0) || (data_set->recheck_by > recheck))) { data_set->recheck_by = recheck; } } /*! * \internal * \brief Wrapper for pe_unpack_nvpairs() using a cluster working set */ void pe__unpack_dataset_nvpairs(xmlNode *xml_obj, const char *set_name, GHashTable *node_hash, GHashTable *hash, const char *always_first, gboolean overwrite, pe_working_set_t *data_set) { crm_time_t *next_change = crm_time_new_undefined(); pe_unpack_nvpairs(data_set->input, xml_obj, set_name, node_hash, hash, always_first, overwrite, data_set->now, next_change); if (crm_time_is_defined(next_change)) { time_t recheck = (time_t) crm_time_get_seconds_since_epoch(next_change); pe__update_recheck_time(recheck, data_set); } crm_time_free(next_change); } + +bool +pe__resource_is_disabled(pe_resource_t *rsc) +{ + const char *target_role = NULL; + + CRM_CHECK(rsc != NULL, return false); + target_role = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_TARGET_ROLE); + if (target_role) { + enum rsc_role_e target_role_e = text2role(target_role); + + if ((target_role_e == RSC_ROLE_STOPPED) + || ((target_role_e == RSC_ROLE_SLAVE) + && is_set(uber_parent(rsc)->flags, pe_rsc_promotable))) { + return true; + } + } + return false; +} diff --git a/tools/crm_mon.c b/tools/crm_mon.c index 548637a668..34cca86c17 100644 --- a/tools/crm_mon.c +++ b/tools/crm_mon.c @@ -1,1901 +1,1902 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* crm_ends_with_ext */ #include #include #include #include #include #include #include #include #include #include #include #include #include "crm_mon.h" #define SUMMARY "Provides a summary of cluster's current state.\n\n" \ "Outputs varying levels of detail in a number of different formats." /* * Definitions indicating which items to print */ static unsigned int show = mon_show_default; /* * Definitions indicating how to output */ static mon_output_format_t output_format = mon_output_unset; /* other globals */ static GMainLoop *mainloop = NULL; static guint timer_id = 0; static mainloop_timer_t *refresh_timer = NULL; static pe_working_set_t *mon_data_set = NULL; static cib_t *cib = NULL; static stonith_t *st = NULL; static xmlNode *current_cib = NULL; static pcmk__common_args_t *args = NULL; static pcmk__output_t *out = NULL; static GOptionContext *context = NULL; /* FIXME allow, detect, and correctly interpret glob pattern or regex? */ const char *print_neg_location_prefix = ""; static time_t last_refresh = 0; crm_trigger_t *refresh_trigger = NULL; static pcmk__supported_format_t formats[] = { #if CURSES_ENABLED CRM_MON_SUPPORTED_FORMAT_CURSES, #endif PCMK__SUPPORTED_FORMAT_HTML, PCMK__SUPPORTED_FORMAT_NONE, PCMK__SUPPORTED_FORMAT_TEXT, CRM_MON_SUPPORTED_FORMAT_XML, { NULL, NULL, NULL } }; /* Define exit codes for monitoring-compatible output * For nagios plugins, the possibilities are * OK=0, WARN=1, CRIT=2, and UNKNOWN=3 */ #define MON_STATUS_WARN CRM_EX_ERROR #define MON_STATUS_CRIT CRM_EX_INVALID_PARAM #define MON_STATUS_UNKNOWN CRM_EX_UNIMPLEMENT_FEATURE #define RECONNECT_MSECS 5000 struct { int reconnect_msec; int fence_history_level; gboolean daemonize; gboolean show_bans; char *pid_file; char *external_agent; char *external_recipient; unsigned int mon_ops; } options = { .reconnect_msec = RECONNECT_MSECS, .fence_history_level = 1, .mon_ops = mon_op_default }; static void clean_up_connections(void); static crm_exit_t clean_up(crm_exit_t exit_code); static void crm_diff_update(const char *event, xmlNode * msg); static gboolean mon_refresh_display(gpointer user_data); static int cib_connect(gboolean full); static void mon_st_callback_event(stonith_t * st, stonith_event_t * e); static void mon_st_callback_display(stonith_t * st, stonith_event_t * e); static void kick_refresh(gboolean data_updated); static gboolean as_cgi_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { if (args->output_ty != NULL) { free(args->output_ty); } args->output_ty = strdup("html"); output_format = mon_output_cgi; options.mon_ops |= mon_op_one_shot; return TRUE; } static gboolean as_html_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { if (args->output_ty != NULL) { free(args->output_ty); } if (args->output_dest != NULL) { free(args->output_dest); } if (optarg != NULL) { args->output_dest = strdup(optarg); } args->output_ty = strdup("html"); output_format = mon_output_html; umask(S_IWGRP | S_IWOTH); return TRUE; } static gboolean as_simple_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { if (args->output_ty != NULL) { free(args->output_ty); } args->output_ty = strdup("text"); output_format = mon_output_monitor; options.mon_ops |= mon_op_one_shot; return TRUE; } static gboolean as_xml_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { if (args->output_ty != NULL) { free(args->output_ty); } args->output_ty = strdup("xml"); output_format = mon_output_legacy_xml; options.mon_ops |= mon_op_one_shot; return TRUE; } static gboolean fence_history_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { int rc = crm_atoi(optarg, "2"); if (rc == -1 || rc > 3) { g_set_error(error, G_OPTION_ERROR, CRM_EX_INVALID_PARAM, "Fence history must be 0-3"); return FALSE; } else { options.fence_history_level = rc; } return TRUE; } static gboolean group_by_node_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_group_by_node; return TRUE; } static gboolean hide_headers_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { show &= ~mon_show_headers; return TRUE; } static gboolean inactive_resources_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_inactive_resources; return TRUE; } static gboolean no_curses_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { output_format = mon_output_plain; return TRUE; } static gboolean one_shot_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_one_shot; return TRUE; } static gboolean print_brief_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_print_brief; return TRUE; } static gboolean print_clone_detail_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_print_clone_detail; return TRUE; } static gboolean print_pending_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_print_pending; return TRUE; } static gboolean print_timing_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_print_timing; show |= mon_show_operations; return TRUE; } static gboolean reconnect_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { int rc = crm_get_msec(optarg); if (rc == -1) { g_set_error(error, G_OPTION_ERROR, CRM_EX_INVALID_PARAM, "Invalid value for -i: %s", optarg); return FALSE; } else { options.reconnect_msec = crm_parse_interval_spec(optarg); } return TRUE; } static gboolean show_attributes_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { show |= mon_show_attributes; return TRUE; } static gboolean show_bans_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { show |= mon_show_bans; if (optarg != NULL) { print_neg_location_prefix = optarg; } return TRUE; } static gboolean show_failcounts_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { show |= mon_show_failcounts; return TRUE; } static gboolean show_operations_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { show |= mon_show_operations; return TRUE; } static gboolean show_tickets_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { show |= mon_show_tickets; return TRUE; } static gboolean use_cib_file_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { setenv("CIB_file", optarg, 1); options.mon_ops |= mon_op_one_shot; return TRUE; } static gboolean watch_fencing_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_watch_fencing; return TRUE; } #define INDENT " " /* *INDENT-OFF* */ static GOptionEntry addl_entries[] = { { "interval", 'i', 0, G_OPTION_ARG_CALLBACK, reconnect_cb, "Update frequency (default is 5 seconds)", "TIMESPEC" }, { "one-shot", '1', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, one_shot_cb, "Display the cluster status once on the console and exit", NULL }, { "daemonize", 'd', 0, G_OPTION_ARG_NONE, &options.daemonize, "Run in the background as a daemon.\n" INDENT "Requires at least one of --output-to and --external-agent.", NULL }, { "pid-file", 'p', 0, G_OPTION_ARG_FILENAME, &options.pid_file, "(Advanced) Daemon pid file location", "FILE" }, { "external-agent", 'E', 0, G_OPTION_ARG_FILENAME, &options.external_agent, "A program to run when resource operations take place", "FILE" }, { "external-recipient", 'e', 0, G_OPTION_ARG_STRING, &options.external_recipient, "A recipient for your program (assuming you want the program to send something to someone).", "RCPT" }, { "xml-file", 'x', G_OPTION_FLAG_HIDDEN, G_OPTION_ARG_CALLBACK, use_cib_file_cb, NULL, NULL }, { NULL } }; static GOptionEntry display_entries[] = { { "group-by-node", 'n', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, group_by_node_cb, "Group resources by node", NULL }, { "inactive", 'r', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, inactive_resources_cb, "Display inactive resources", NULL }, { "failcounts", 'f', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, show_failcounts_cb, "Display resource fail counts", NULL }, { "operations", 'o', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, show_operations_cb, "Display resource operation history", NULL }, { "timing-details", 't', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, print_timing_cb, "Display resource operation history with timing details", NULL }, { "tickets", 'c', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, show_tickets_cb, "Display cluster tickets", NULL }, { "watch-fencing", 'W', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, watch_fencing_cb, "Listen for fencing events. For use with --external-agent", NULL }, { "fence-history", 'm', G_OPTION_FLAG_OPTIONAL_ARG, G_OPTION_ARG_CALLBACK, fence_history_cb, "Show fence history:\n" INDENT "0=off, 1=failures and pending (default without option),\n" INDENT "2=add successes (default without value for option),\n" INDENT "3=show full history without reduction to most recent of each flavor", "LEVEL" }, { "neg-locations", 'L', G_OPTION_FLAG_OPTIONAL_ARG, G_OPTION_ARG_CALLBACK, show_bans_cb, "Display negative location constraints [optionally filtered by id prefix]", NULL }, { "show-node-attributes", 'A', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, show_attributes_cb, "Display node attributes", NULL }, { "hide-headers", 'D', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, hide_headers_cb, "Hide all headers", NULL }, { "show-detail", 'R', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, print_clone_detail_cb, "Show more details (node IDs, individual clone instances)", NULL }, { "brief", 'b', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, print_brief_cb, "Brief output", NULL }, { "pending", 'j', G_OPTION_FLAG_HIDDEN|G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, print_pending_cb, "Display pending state if 'record-pending' is enabled", NULL }, { "simple-status", 's', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, as_simple_cb, "Display the cluster status once as a simple one line output (suitable for nagios)", NULL }, { NULL } }; static GOptionEntry deprecated_entries[] = { { "as-html", 'h', G_OPTION_FLAG_FILENAME, G_OPTION_ARG_CALLBACK, as_html_cb, "Write cluster status to the named HTML file.\n" INDENT "Use --output-as=html --output-to=FILE instead.", "FILE" }, { "as-xml", 'X', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, as_xml_cb, "Write cluster status as XML to stdout. This will enable one-shot mode.\n" INDENT "Use --output-as=xml instead.", NULL }, { "disable-ncurses", 'N', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, no_curses_cb, "Disable the use of ncurses.\n" INDENT "Use --output-as=text instead.", NULL }, { "web-cgi", 'w', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, as_cgi_cb, "Web mode with output suitable for CGI (preselected when run as *.cgi).\n" INDENT "Use --output-as=html --html-cgi instead.", NULL }, { NULL } }; /* *INDENT-ON* */ static gboolean mon_timer_popped(gpointer data) { int rc = pcmk_ok; #if CURSES_ENABLED if (output_format == mon_output_console) { clear(); refresh(); } #endif if (timer_id > 0) { g_source_remove(timer_id); timer_id = 0; } print_as(output_format, "Reconnecting...\n"); rc = cib_connect(TRUE); if (rc != pcmk_ok) { timer_id = g_timeout_add(options.reconnect_msec, mon_timer_popped, NULL); } return FALSE; } static void mon_cib_connection_destroy(gpointer user_data) { print_as(output_format, "Connection to the cluster-daemons terminated\n"); if (refresh_timer != NULL) { /* we'll trigger a refresh after reconnect */ mainloop_timer_stop(refresh_timer); } if (timer_id) { /* we'll trigger a new reconnect-timeout at the end */ g_source_remove(timer_id); timer_id = 0; } if (st) { /* the client API won't properly reconnect notifications * if they are still in the table - so remove them */ st->cmds->remove_notification(st, T_STONITH_NOTIFY_DISCONNECT); st->cmds->remove_notification(st, T_STONITH_NOTIFY_FENCE); st->cmds->remove_notification(st, T_STONITH_NOTIFY_HISTORY); if (st->state != stonith_disconnected) { st->cmds->disconnect(st); } } if (cib) { cib->cmds->signoff(cib); timer_id = g_timeout_add(options.reconnect_msec, mon_timer_popped, NULL); } return; } /* * Mainloop signal handler. */ static void mon_shutdown(int nsig) { clean_up(CRM_EX_OK); } #if CURSES_ENABLED static sighandler_t ncurses_winch_handler; static void mon_winresize(int nsig) { static int not_done; int lines = 0, cols = 0; if (!not_done++) { if (ncurses_winch_handler) /* the original ncurses WINCH signal handler does the * magic of retrieving the new window size; * otherwise, we'd have to use ioctl or tgetent */ (*ncurses_winch_handler) (SIGWINCH); getmaxyx(stdscr, lines, cols); resizeterm(lines, cols); mainloop_set_trigger(refresh_trigger); } not_done--; } #endif static int cib_connect(gboolean full) { int rc = pcmk_ok; static gboolean need_pass = TRUE; CRM_CHECK(cib != NULL, return -EINVAL); if (getenv("CIB_passwd") != NULL) { need_pass = FALSE; } if (is_set(options.mon_ops, mon_op_fence_connect) && st == NULL) { st = stonith_api_new(); } if (is_set(options.mon_ops, mon_op_fence_connect) && st != NULL && st->state == stonith_disconnected) { rc = st->cmds->connect(st, crm_system_name, NULL); if (rc == pcmk_ok) { crm_trace("Setting up stonith callbacks"); if (is_set(options.mon_ops, mon_op_watch_fencing)) { st->cmds->register_notification(st, T_STONITH_NOTIFY_DISCONNECT, mon_st_callback_event); st->cmds->register_notification(st, T_STONITH_NOTIFY_FENCE, mon_st_callback_event); } else { st->cmds->register_notification(st, T_STONITH_NOTIFY_DISCONNECT, mon_st_callback_display); st->cmds->register_notification(st, T_STONITH_NOTIFY_HISTORY, mon_st_callback_display); } } } if (cib->state != cib_connected_query && cib->state != cib_connected_command) { crm_trace("Connecting to the CIB"); if ((output_format == mon_output_console) && need_pass && (cib->variant == cib_remote)) { need_pass = FALSE; print_as(output_format, "Password:"); } rc = cib->cmds->signon(cib, crm_system_name, cib_query); if (rc != pcmk_ok) { return rc; } rc = cib->cmds->query(cib, NULL, ¤t_cib, cib_scope_local | cib_sync_call); if (rc == pcmk_ok) { mon_refresh_display(&output_format); } if (rc == pcmk_ok && full) { if (rc == pcmk_ok) { rc = cib->cmds->set_connection_dnotify(cib, mon_cib_connection_destroy); if (rc == -EPROTONOSUPPORT) { print_as (output_format, "Notification setup not supported, won't be able to reconnect after failure"); if (output_format == mon_output_console) { sleep(2); } rc = pcmk_ok; } } if (rc == pcmk_ok) { cib->cmds->del_notify_callback(cib, T_CIB_DIFF_NOTIFY, crm_diff_update); rc = cib->cmds->add_notify_callback(cib, T_CIB_DIFF_NOTIFY, crm_diff_update); } if (rc != pcmk_ok) { print_as(output_format, "Notification setup failed, could not monitor CIB actions"); if (output_format == mon_output_console) { sleep(2); } clean_up_connections(); } } } return rc; } #if CURSES_ENABLED static const char * get_option_desc(char c) { const char *desc = "No help available"; for (GOptionEntry *entry = display_entries; entry != NULL; entry++) { if (entry->short_name == c) { desc = entry->description; break; } } return desc; } #define print_option_help(output_format, option, condition) \ out->info(out, "%c %c: \t%s", ((condition)? '*': ' '), option, get_option_desc(option)); static gboolean detect_user_input(GIOChannel *channel, GIOCondition condition, gpointer user_data) { int c; gboolean config_mode = FALSE; while (1) { /* Get user input */ c = getchar(); switch (c) { case 'm': if (!options.fence_history_level) { options.mon_ops |= mon_op_fence_history; options.mon_ops |= mon_op_fence_connect; if (st == NULL) { mon_cib_connection_destroy(NULL); } } show ^= mon_show_fence_history; break; case 'c': show ^= mon_show_tickets; break; case 'f': show ^= mon_show_failcounts; break; case 'n': options.mon_ops ^= mon_op_group_by_node; break; case 'o': show ^= mon_show_operations; if ((show & mon_show_operations) == 0) { options.mon_ops &= ~mon_op_print_timing; } break; case 'r': options.mon_ops ^= mon_op_inactive_resources; break; case 'R': options.mon_ops ^= mon_op_print_clone_detail; break; case 't': options.mon_ops ^= mon_op_print_timing; if (is_set(options.mon_ops, mon_op_print_timing)) { show |= mon_show_operations; } break; case 'A': show ^= mon_show_attributes; break; case 'L': show ^= mon_show_bans; break; case 'D': /* If any header is shown, clear them all, otherwise set them all */ if (show & mon_show_headers) { show &= ~mon_show_headers; } else { show |= mon_show_headers; } break; case 'b': options.mon_ops ^= mon_op_print_brief; break; case 'j': options.mon_ops ^= mon_op_print_pending; break; case '?': config_mode = TRUE; break; default: goto refresh; } if (!config_mode) goto refresh; blank_screen(); out->info(out, "%s", "Display option change mode\n"); print_option_help(out, 'c', show & mon_show_tickets); print_option_help(out, 'f', show & mon_show_failcounts); print_option_help(out, 'n', is_set(options.mon_ops, mon_op_group_by_node)); print_option_help(out, 'o', show & mon_show_operations); print_option_help(out, 'r', is_set(options.mon_ops, mon_op_inactive_resources)); print_option_help(out, 't', is_set(options.mon_ops, mon_op_print_timing)); print_option_help(out, 'A', show & mon_show_attributes); print_option_help(out, 'L', show & mon_show_bans); print_option_help(out, 'D', (show & mon_show_headers) == 0); print_option_help(out, 'R', is_set(options.mon_ops, mon_op_print_clone_detail)); print_option_help(out, 'b', is_set(options.mon_ops, mon_op_print_brief)); print_option_help(out, 'j', is_set(options.mon_ops, mon_op_print_pending)); print_option_help(out, 'm', (show & mon_show_fence_history)); out->info(out, "%s", "\nToggle fields via field letter, type any other key to return"); } refresh: mon_refresh_display(NULL); return TRUE; } #endif // Basically crm_signal_handler(SIGCHLD, SIG_IGN) plus the SA_NOCLDWAIT flag static void avoid_zombies() { struct sigaction sa; memset(&sa, 0, sizeof(struct sigaction)); if (sigemptyset(&sa.sa_mask) < 0) { crm_warn("Cannot avoid zombies: %s", pcmk_strerror(errno)); return; } sa.sa_handler = SIG_IGN; sa.sa_flags = SA_RESTART|SA_NOCLDWAIT; if (sigaction(SIGCHLD, &sa, NULL) < 0) { crm_warn("Cannot avoid zombies: %s", pcmk_strerror(errno)); } } static GOptionContext * build_arg_context(pcmk__common_args_t *args, GOptionGroup **group) { GOptionContext *context = NULL; GOptionEntry extra_prog_entries[] = { { "quiet", 'Q', 0, G_OPTION_ARG_NONE, &(args->quiet), "Be less descriptive in output.", NULL }, { NULL } }; const char *description = "*Examples*\n\n" "Display the cluster status on the console with updates as they occur:\n\n" "\tcrm_mon\n\n" "Display the cluster status on the console just once then exit:\n\n" "\tcrm_mon -1\n\n" "Display your cluster status, group resources by node, and include inactive resources in the list:\n\n" "\tcrm_mon --group-by-node --inactive\n\n" "Start crm_mon as a background daemon and have it write the cluster status to an HTML file:\n\n" "\tcrm_mon --daemonize --output-as html --output-to /path/to/docroot/filename.html\n\n" "Start crm_mon and export the current cluster status as XML to stdout, then exit:\n\n" "\tcrm_mon --output-as xml\n\n" "*Time Specification*\n\n" "The TIMESPEC in any command line option can be specified in many different\n" "formats. It can be just an integer number of seconds, a number plus units\n" "(ms/msec/us/usec/s/sec/m/min/h/hr), or an ISO 8601 period specification.\n"; context = pcmk__build_arg_context(args, "console (default), html, text, xml", group); pcmk__add_main_args(context, extra_prog_entries); g_option_context_set_description(context, description); pcmk__add_arg_group(context, "display", "Display Options:", "Show display options", display_entries); pcmk__add_arg_group(context, "additional", "Additional Options:", "Show additional options", addl_entries); pcmk__add_arg_group(context, "deprecated", "Deprecated Options:", "Show deprecated options", deprecated_entries); return context; } /* If certain format options were specified, we want to set some extra * options. We can just process these like they were given on the * command line. */ static void add_output_args() { GError *error = NULL; if (output_format == mon_output_plain) { if (!pcmk__force_args(context, &error, "%s --text-fancy", g_get_prgname())) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); clean_up(CRM_EX_USAGE); } } else if (output_format == mon_output_html) { if (!pcmk__force_args(context, &error, "%s --html-title \"Cluster Status\"", g_get_prgname())) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); clean_up(CRM_EX_USAGE); } } else if (output_format == mon_output_cgi) { if (!pcmk__force_args(context, &error, "%s --html-cgi --html-title \"Cluster Status\"", g_get_prgname())) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); clean_up(CRM_EX_USAGE); } } else if (output_format == mon_output_xml) { if (!pcmk__force_args(context, &error, "%s --xml-simple-list", g_get_prgname())) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); clean_up(CRM_EX_USAGE); } } else if (output_format == mon_output_legacy_xml) { output_format = mon_output_xml; if (!pcmk__force_args(context, &error, "%s --xml-legacy", g_get_prgname())) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); clean_up(CRM_EX_USAGE); } } } /* Which output format to use could come from two places: The --as-xml * style arguments we gave in deprecated_entries above, or the formatted output * arguments added by pcmk__register_formats. If the latter were used, * output_format will be mon_output_unset. * * Call the callbacks as if those older style arguments were provided so * the various things they do get done. */ static void reconcile_output_format(pcmk__common_args_t *args) { gboolean retval = TRUE; GError *error = NULL; if (output_format != mon_output_unset) { return; } if (safe_str_eq(args->output_ty, "html")) { char *dest = NULL; if (args->output_dest != NULL) { dest = strdup(args->output_dest); } retval = as_html_cb("h", dest, NULL, &error); free(dest); } else if (safe_str_eq(args->output_ty, "text")) { retval = no_curses_cb("N", NULL, NULL, &error); } else if (safe_str_eq(args->output_ty, "xml")) { if (args->output_ty != NULL) { free(args->output_ty); } args->output_ty = strdup("xml"); output_format = mon_output_xml; options.mon_ops |= mon_op_one_shot; } else if (is_set(options.mon_ops, mon_op_one_shot)) { if (args->output_ty != NULL) { free(args->output_ty); } args->output_ty = strdup("text"); output_format = mon_output_plain; } else { /* Neither old nor new arguments were given, so set the default. */ if (args->output_ty != NULL) { free(args->output_ty); } args->output_ty = strdup("console"); output_format = mon_output_console; } if (!retval) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); clean_up(CRM_EX_USAGE); } } int main(int argc, char **argv) { int rc = pcmk_ok; char **processed_args = NULL; GOptionGroup *output_group = NULL; GError *error = NULL; args = pcmk__new_common_args(SUMMARY); context = build_arg_context(args, &output_group); pcmk__register_formats(output_group, formats); options.pid_file = strdup("/tmp/ClusterMon.pid"); crm_log_cli_init("crm_mon"); // Avoid needing to wait for subprocesses forked for -E/--external-agent avoid_zombies(); if (crm_ends_with_ext(argv[0], ".cgi") == TRUE) { output_format = mon_output_cgi; options.mon_ops |= mon_op_one_shot; } processed_args = pcmk__cmdline_preproc(argc, argv, "ehimpxEL"); if (!g_option_context_parse_strv(context, &processed_args, &error)) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); return clean_up(CRM_EX_USAGE); } for (int i = 0; i < args->verbosity; i++) { crm_bump_log_level(argc, argv); } if (!args->version) { if (args->quiet) { show &= ~mon_show_times; } if (is_set(options.mon_ops, mon_op_watch_fencing)) { options.mon_ops |= mon_op_fence_connect; /* don't moan as fence_history_level == 1 is default */ options.fence_history_level = 0; } /* create the cib-object early to be able to do further * decisions based on the cib-source */ cib = cib_new(); if (cib == NULL) { rc = -EINVAL; } else { switch (cib->variant) { case cib_native: /* cib & fencing - everything available */ break; case cib_file: /* Don't try to connect to fencing as we * either don't have a running cluster or * the fencing-information would possibly * not match the cib data from a file. * As we don't expect cib-updates coming * in enforce one-shot. */ options.fence_history_level = 0; options.mon_ops |= mon_op_one_shot; break; case cib_remote: /* updates coming in but no fencing */ options.fence_history_level = 0; break; case cib_undefined: case cib_database: default: /* something is odd */ rc = -EINVAL; break; } } switch (options.fence_history_level) { case 3: options.mon_ops |= mon_op_fence_full_history; /* fall through to next lower level */ case 2: show |= mon_show_fence_history; /* fall through to next lower level */ case 1: options.mon_ops |= mon_op_fence_history; options.mon_ops |= mon_op_fence_connect; break; default: break; } if (is_set(options.mon_ops, mon_op_one_shot)) { if (output_format == mon_output_console) { output_format = mon_output_plain; } } else if (options.daemonize) { if ((output_format == mon_output_console) || (output_format == mon_output_plain)) { output_format = mon_output_none; } crm_enable_stderr(FALSE); if ((args->output_dest == NULL || safe_str_eq(args->output_dest, "-")) && !options.external_agent) { printf("--daemonize requires at least one of --output-to and --external-agent\n"); return clean_up(CRM_EX_USAGE); } if (cib) { /* to be on the safe side don't have cib-object around * when we are forking */ cib_delete(cib); cib = NULL; crm_make_daemon(crm_system_name, TRUE, options.pid_file); cib = cib_new(); if (cib == NULL) { rc = -EINVAL; } /* otherwise assume we've got the same cib-object we've just destroyed * in our parent */ } } else if (output_format == mon_output_console) { #if CURSES_ENABLED crm_enable_stderr(FALSE); #else options.mon_ops |= mon_op_one_shot; output_format = mon_output_plain; printf("Defaulting to one-shot mode\n"); printf("You need to have curses available at compile time to enable console mode\n"); #endif } } if (rc != pcmk_ok) { // Shouldn't really be possible fprintf(stderr, "Invalid CIB source\n"); return clean_up(CRM_EX_ERROR); } reconcile_output_format(args); add_output_args(); /* Create the output format - output_format must not be changed after this point. */ if (args->version && output_format == mon_output_console) { /* Use the text output format here if we are in curses mode but were given * --version. Displaying version information uses printf, and then we * immediately exit. We don't want to initialize curses for that. */ rc = pcmk__output_new(&out, "text", args->output_dest, argv); } else { rc = pcmk__output_new(&out, args->output_ty, args->output_dest, argv); } if (rc != 0) { fprintf(stderr, "Error creating output format %s: %s\n", args->output_ty, pcmk_strerror(rc)); return clean_up(CRM_EX_ERROR); } crm_mon_register_messages(out); pe__register_messages(out); stonith__register_messages(out); if (args->version) { out->version(out, false); return clean_up(CRM_EX_OK); } /* Extra sanity checks when in CGI mode */ if (output_format == mon_output_cgi) { if (cib && cib->variant == cib_file) { fprintf(stderr, "CGI mode used with CIB file\n"); return clean_up(CRM_EX_USAGE); } else if (options.external_agent != NULL) { fprintf(stderr, "CGI mode cannot be used with --external-agent\n"); return clean_up(CRM_EX_USAGE); } else if (options.daemonize == TRUE) { fprintf(stderr, "CGI mode cannot be used with -d\n"); return clean_up(CRM_EX_USAGE); } } /* XML output always prints everything */ if (output_format == mon_output_xml || output_format == mon_output_legacy_xml) { show = mon_show_all; options.mon_ops |= mon_op_print_timing; } crm_info("Starting %s", crm_system_name); if (cib) { do { if (is_not_set(options.mon_ops, mon_op_one_shot)) { print_as(output_format ,"Waiting until cluster is available on this node ...\n"); } rc = cib_connect(is_not_set(options.mon_ops, mon_op_one_shot)); if (is_set(options.mon_ops, mon_op_one_shot)) { break; } else if (rc != pcmk_ok) { sleep(options.reconnect_msec / 1000); #if CURSES_ENABLED if (output_format == mon_output_console) { clear(); refresh(); } #endif } else { if (output_format == mon_output_html && out->dest != stdout) { printf("Writing html to %s ...\n", args->output_dest); } } } while (rc == -ENOTCONN); } if (rc != pcmk_ok) { if (output_format == mon_output_monitor) { printf("CLUSTER CRIT: Connection to cluster failed: %s\n", pcmk_strerror(rc)); return clean_up(MON_STATUS_CRIT); } else { if (rc == -ENOTCONN) { out->err(out, "%s", "\nError: cluster is not available on this node"); } else { out->err(out, "\nConnection to cluster failed: %s", pcmk_strerror(rc)); } } if (output_format == mon_output_console) { sleep(2); } return clean_up(crm_errno2exit(rc)); } if (is_set(options.mon_ops, mon_op_one_shot)) { return clean_up(CRM_EX_OK); } mainloop = g_main_loop_new(NULL, FALSE); mainloop_add_signal(SIGTERM, mon_shutdown); mainloop_add_signal(SIGINT, mon_shutdown); #if CURSES_ENABLED if (output_format == mon_output_console) { ncurses_winch_handler = crm_signal_handler(SIGWINCH, mon_winresize); if (ncurses_winch_handler == SIG_DFL || ncurses_winch_handler == SIG_IGN || ncurses_winch_handler == SIG_ERR) ncurses_winch_handler = NULL; g_io_add_watch(g_io_channel_unix_new(STDIN_FILENO), G_IO_IN, detect_user_input, NULL); } #endif refresh_trigger = mainloop_add_trigger(G_PRIORITY_LOW, mon_refresh_display, NULL); g_main_loop_run(mainloop); g_main_loop_unref(mainloop); crm_info("Exiting %s", crm_system_name); return clean_up(CRM_EX_OK); } /*! * \internal * \brief Print one-line status suitable for use with monitoring software * * \param[in] data_set Working set of CIB state * \param[in] history List of stonith actions * * \note This function's output (and the return code when the program exits) * should conform to https://www.monitoring-plugins.org/doc/guidelines.html */ static void print_simple_status(pcmk__output_t *out, pe_working_set_t * data_set, stonith_history_t *history, unsigned int mon_ops) { GListPtr gIter = NULL; int nodes_online = 0; int nodes_standby = 0; int nodes_maintenance = 0; char *offline_nodes = NULL; gboolean no_dc = FALSE; gboolean offline = FALSE; if (data_set->dc_node == NULL) { mon_ops |= mon_op_has_warnings; no_dc = TRUE; } for (gIter = data_set->nodes; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; if (node->details->standby && node->details->online) { nodes_standby++; } else if (node->details->maintenance && node->details->online) { nodes_maintenance++; } else if (node->details->online) { nodes_online++; } else { char *s = crm_strdup_printf("offline node: %s", node->details->uname); offline_nodes = add_list_element(offline_nodes, s); free(s); mon_ops |= mon_op_has_warnings; offline = TRUE; } } if (is_set(mon_ops, mon_op_has_warnings)) { out->info(out, "CLUSTER WARN:%s%s%s", no_dc ? " No DC" : "", no_dc && offline ? "," : "", offline ? offline_nodes : ""); free(offline_nodes); } else { - int nresources = count_resources(data_set, NULL); char *nodes_standby_s = NULL; char *nodes_maint_s = NULL; if (nodes_standby > 0) { nodes_standby_s = crm_strdup_printf(", %d standby node%s", nodes_standby, s_if_plural(nodes_standby)); } if (nodes_maintenance > 0) { - nodes_maint_s = crm_strdup_printf(", %d maintenance node%s", nresources, - s_if_plural(nresources)); + nodes_maint_s = crm_strdup_printf(", %d maintenance node%s", + data_set->ninstances, + s_if_plural(data_set->ninstances)); } - out->info(out, "CLUSTER OK: %dnode%s online%s%s, %d resource%s configured", + out->info(out, "CLUSTER OK: %dnode%s online%s%s, " + "%d resource instance%s configured", nodes_online, s_if_plural(nodes_online), nodes_standby_s != NULL ? nodes_standby_s : "", nodes_maint_s != NULL ? nodes_maint_s : "", - nresources, s_if_plural(nresources)); + data_set->ninstances, s_if_plural(data_set->ninstances)); free(nodes_standby_s); free(nodes_maint_s); } } /*! * \internal * \brief Reduce the stonith-history * for successful actions we keep the last of every action-type & target * for failed actions we record as well who had failed * for actions in progress we keep full track * * \param[in] history List of stonith actions * */ static stonith_history_t * reduce_stonith_history(stonith_history_t *history) { stonith_history_t *new = history, *hp, *np; if (new) { hp = new->next; new->next = NULL; while (hp) { stonith_history_t *hp_next = hp->next; hp->next = NULL; for (np = new; ; np = np->next) { if ((hp->state == st_done) || (hp->state == st_failed)) { /* action not in progress */ if (safe_str_eq(hp->target, np->target) && safe_str_eq(hp->action, np->action) && (hp->state == np->state) && ((hp->state == st_done) || safe_str_eq(hp->delegate, np->delegate))) { /* purge older hp */ stonith_history_free(hp); break; } } if (!np->next) { np->next = hp; break; } } hp = hp_next; } } return new; } static int send_custom_trap(const char *node, const char *rsc, const char *task, int target_rc, int rc, int status, const char *desc) { pid_t pid; /*setenv needs chars, these are ints */ char *rc_s = crm_itoa(rc); char *status_s = crm_itoa(status); char *target_rc_s = crm_itoa(target_rc); crm_debug("Sending external notification to '%s' via '%s'", options.external_recipient, options.external_agent); if(rsc) { setenv("CRM_notify_rsc", rsc, 1); } if (options.external_recipient) { setenv("CRM_notify_recipient", options.external_recipient, 1); } setenv("CRM_notify_node", node, 1); setenv("CRM_notify_task", task, 1); setenv("CRM_notify_desc", desc, 1); setenv("CRM_notify_rc", rc_s, 1); setenv("CRM_notify_target_rc", target_rc_s, 1); setenv("CRM_notify_status", status_s, 1); pid = fork(); if (pid == -1) { crm_perror(LOG_ERR, "notification fork() failed."); } if (pid == 0) { /* crm_debug("notification: I am the child. Executing the nofitication program."); */ execl(options.external_agent, options.external_agent, NULL); exit(CRM_EX_ERROR); } crm_trace("Finished running custom notification program '%s'.", options.external_agent); free(target_rc_s); free(status_s); free(rc_s); return 0; } static void handle_rsc_op(xmlNode * xml, const char *node_id) { int rc = -1; int status = -1; int target_rc = -1; gboolean notify = TRUE; char *rsc = NULL; char *task = NULL; const char *desc = NULL; const char *magic = NULL; const char *id = NULL; const char *node = NULL; xmlNode *n = xml; xmlNode * rsc_op = xml; if(strcmp((const char*)xml->name, XML_LRM_TAG_RSC_OP) != 0) { xmlNode *cIter; for(cIter = xml->children; cIter; cIter = cIter->next) { handle_rsc_op(cIter, node_id); } return; } id = crm_element_value(rsc_op, XML_LRM_ATTR_TASK_KEY); if (id == NULL) { /* Compatibility with <= 1.1.5 */ id = ID(rsc_op); } magic = crm_element_value(rsc_op, XML_ATTR_TRANSITION_MAGIC); if (magic == NULL) { /* non-change */ return; } if (!decode_transition_magic(magic, NULL, NULL, NULL, &status, &rc, &target_rc)) { crm_err("Invalid event %s detected for %s", magic, id); return; } if (parse_op_key(id, &rsc, &task, NULL) == FALSE) { crm_err("Invalid event detected for %s", id); goto bail; } node = crm_element_value(rsc_op, XML_LRM_ATTR_TARGET); while (n != NULL && safe_str_neq(XML_CIB_TAG_STATE, TYPE(n))) { n = n->parent; } if(node == NULL && n) { node = crm_element_value(n, XML_ATTR_UNAME); } if (node == NULL && n) { node = ID(n); } if (node == NULL) { node = node_id; } if (node == NULL) { crm_err("No node detected for event %s (%s)", magic, id); goto bail; } /* look up where we expected it to be? */ desc = pcmk_strerror(pcmk_ok); if (status == PCMK_LRM_OP_DONE && target_rc == rc) { crm_notice("%s of %s on %s completed: %s", task, rsc, node, desc); if (rc == PCMK_OCF_NOT_RUNNING) { notify = FALSE; } } else if (status == PCMK_LRM_OP_DONE) { desc = services_ocf_exitcode_str(rc); crm_warn("%s of %s on %s failed: %s", task, rsc, node, desc); } else { desc = services_lrm_status_str(status); crm_warn("%s of %s on %s failed: %s", task, rsc, node, desc); } if (notify && options.external_agent) { send_custom_trap(node, rsc, task, target_rc, rc, status, desc); } bail: free(rsc); free(task); } static gboolean mon_trigger_refresh(gpointer user_data) { mainloop_set_trigger(refresh_trigger); return FALSE; } #define NODE_PATT "/lrm[@id=" static char * get_node_from_xpath(const char *xpath) { char *nodeid = NULL; char *tmp = strstr(xpath, NODE_PATT); if(tmp) { tmp += strlen(NODE_PATT); tmp += 1; nodeid = strdup(tmp); tmp = strstr(nodeid, "\'"); CRM_ASSERT(tmp); tmp[0] = 0; } return nodeid; } static void crm_diff_update_v2(const char *event, xmlNode * msg) { xmlNode *change = NULL; xmlNode *diff = get_message_xml(msg, F_CIB_UPDATE_RESULT); for (change = __xml_first_child(diff); change != NULL; change = __xml_next(change)) { const char *name = NULL; const char *op = crm_element_value(change, XML_DIFF_OP); const char *xpath = crm_element_value(change, XML_DIFF_PATH); xmlNode *match = NULL; const char *node = NULL; if(op == NULL) { continue; } else if(strcmp(op, "create") == 0) { match = change->children; } else if(strcmp(op, "move") == 0) { continue; } else if(strcmp(op, "delete") == 0) { continue; } else if(strcmp(op, "modify") == 0) { match = first_named_child(change, XML_DIFF_RESULT); if(match) { match = match->children; } } if(match) { name = (const char *)match->name; } crm_trace("Handling %s operation for %s %p, %s", op, xpath, match, name); if(xpath == NULL) { /* Version field, ignore */ } else if(name == NULL) { crm_debug("No result for %s operation to %s", op, xpath); CRM_ASSERT(strcmp(op, "delete") == 0 || strcmp(op, "move") == 0); } else if(strcmp(name, XML_TAG_CIB) == 0) { xmlNode *state = NULL; xmlNode *status = first_named_child(match, XML_CIB_TAG_STATUS); for (state = __xml_first_child_element(status); state != NULL; state = __xml_next_element(state)) { node = crm_element_value(state, XML_ATTR_UNAME); if (node == NULL) { node = ID(state); } handle_rsc_op(state, node); } } else if(strcmp(name, XML_CIB_TAG_STATUS) == 0) { xmlNode *state = NULL; for (state = __xml_first_child_element(match); state != NULL; state = __xml_next_element(state)) { node = crm_element_value(state, XML_ATTR_UNAME); if (node == NULL) { node = ID(state); } handle_rsc_op(state, node); } } else if(strcmp(name, XML_CIB_TAG_STATE) == 0) { node = crm_element_value(match, XML_ATTR_UNAME); if (node == NULL) { node = ID(match); } handle_rsc_op(match, node); } else if(strcmp(name, XML_CIB_TAG_LRM) == 0) { node = ID(match); handle_rsc_op(match, node); } else if(strcmp(name, XML_LRM_TAG_RESOURCES) == 0) { char *local_node = get_node_from_xpath(xpath); handle_rsc_op(match, local_node); free(local_node); } else if(strcmp(name, XML_LRM_TAG_RESOURCE) == 0) { char *local_node = get_node_from_xpath(xpath); handle_rsc_op(match, local_node); free(local_node); } else if(strcmp(name, XML_LRM_TAG_RSC_OP) == 0) { char *local_node = get_node_from_xpath(xpath); handle_rsc_op(match, local_node); free(local_node); } else { crm_trace("Ignoring %s operation for %s %p, %s", op, xpath, match, name); } } } static void crm_diff_update_v1(const char *event, xmlNode * msg) { /* Process operation updates */ xmlXPathObject *xpathObj = xpath_search(msg, "//" F_CIB_UPDATE_RESULT "//" XML_TAG_DIFF_ADDED "//" XML_LRM_TAG_RSC_OP); int lpc = 0, max = numXpathResults(xpathObj); for (lpc = 0; lpc < max; lpc++) { xmlNode *rsc_op = getXpathResult(xpathObj, lpc); handle_rsc_op(rsc_op, NULL); } freeXpathObject(xpathObj); } static void crm_diff_update(const char *event, xmlNode * msg) { int rc = -1; static bool stale = FALSE; gboolean cib_updated = FALSE; xmlNode *diff = get_message_xml(msg, F_CIB_UPDATE_RESULT); print_dot(output_format); if (current_cib != NULL) { rc = xml_apply_patchset(current_cib, diff, TRUE); switch (rc) { case -pcmk_err_diff_resync: case -pcmk_err_diff_failed: crm_notice("[%s] Patch aborted: %s (%d)", event, pcmk_strerror(rc), rc); free_xml(current_cib); current_cib = NULL; break; case pcmk_ok: cib_updated = TRUE; break; default: crm_notice("[%s] ABORTED: %s (%d)", event, pcmk_strerror(rc), rc); free_xml(current_cib); current_cib = NULL; } } if (current_cib == NULL) { crm_trace("Re-requesting the full cib"); cib->cmds->query(cib, NULL, ¤t_cib, cib_scope_local | cib_sync_call); } if (options.external_agent) { int format = 0; crm_element_value_int(diff, "format", &format); switch(format) { case 1: crm_diff_update_v1(event, msg); break; case 2: crm_diff_update_v2(event, msg); break; default: crm_err("Unknown patch format: %d", format); } } if (current_cib == NULL) { if(!stale) { print_as(output_format, "--- Stale data ---"); } stale = TRUE; return; } stale = FALSE; kick_refresh(cib_updated); } static gboolean mon_refresh_display(gpointer user_data) { xmlNode *cib_copy = copy_xml(current_cib); stonith_history_t *stonith_history = NULL; last_refresh = time(NULL); if (cli_config_update(&cib_copy, NULL, FALSE) == FALSE) { if (cib) { cib->cmds->signoff(cib); } print_as(output_format, "Upgrade failed: %s", pcmk_strerror(-pcmk_err_schema_validation)); if (output_format == mon_output_console) { sleep(2); } clean_up(CRM_EX_CONFIG); return FALSE; } /* get the stonith-history if there is evidence we need it */ while (is_set(options.mon_ops, mon_op_fence_history)) { if (st != NULL) { if (st->cmds->history(st, st_opt_sync_call, NULL, &stonith_history, 120)) { fprintf(stderr, "Critical: Unable to get stonith-history\n"); mon_cib_connection_destroy(NULL); } else { stonith_history = stonith__sort_history(stonith_history); if (is_not_set(options.mon_ops, mon_op_fence_full_history) && output_format != mon_output_xml) { stonith_history = reduce_stonith_history(stonith_history); } break; /* all other cases are errors */ } } else { fprintf(stderr, "Critical: No stonith-API\n"); } free_xml(cib_copy); print_as(output_format, "Reading stonith-history failed"); if (output_format == mon_output_console) { sleep(2); } return FALSE; } if (mon_data_set == NULL) { mon_data_set = pe_new_working_set(); CRM_ASSERT(mon_data_set != NULL); } mon_data_set->input = cib_copy; cluster_status(mon_data_set); /* Unpack constraints if any section will need them * (tickets may be referenced in constraints but not granted yet, * and bans need negative location constraints) */ if (show & (mon_show_bans | mon_show_tickets)) { xmlNode *cib_constraints = get_object_root(XML_CIB_TAG_CONSTRAINTS, mon_data_set->input); unpack_constraints(cib_constraints, mon_data_set); } switch (output_format) { case mon_output_html: case mon_output_cgi: if (print_html_status(out, output_format, mon_data_set, stonith_history, options.mon_ops, show, print_neg_location_prefix) != 0) { fprintf(stderr, "Critical: Unable to output html file\n"); clean_up(CRM_EX_CANTCREAT); return FALSE; } break; case mon_output_legacy_xml: case mon_output_xml: print_xml_status(out, mon_data_set, stonith_history, options.mon_ops, show, print_neg_location_prefix); break; case mon_output_monitor: print_simple_status(out, mon_data_set, stonith_history, options.mon_ops); if (is_set(options.mon_ops, mon_op_has_warnings)) { clean_up(MON_STATUS_WARN); return FALSE; } break; case mon_output_console: /* If curses is not enabled, this will just fall through to the plain * text case. */ #if CURSES_ENABLED blank_screen(); print_status(out, output_format, mon_data_set, stonith_history, options.mon_ops, show, print_neg_location_prefix); refresh(); break; #endif case mon_output_plain: print_status(out, output_format, mon_data_set, stonith_history, options.mon_ops, show, print_neg_location_prefix); break; case mon_output_unset: case mon_output_none: break; } stonith_history_free(stonith_history); stonith_history = NULL; pe_reset_working_set(mon_data_set); return TRUE; } static void mon_st_callback_event(stonith_t * st, stonith_event_t * e) { if (st->state == stonith_disconnected) { /* disconnect cib as well and have everything reconnect */ mon_cib_connection_destroy(NULL); } else if (options.external_agent) { char *desc = crm_strdup_printf("Operation %s requested by %s for peer %s: %s (ref=%s)", e->operation, e->origin, e->target, pcmk_strerror(e->result), e->id); send_custom_trap(e->target, NULL, e->operation, pcmk_ok, e->result, 0, desc); free(desc); } } static void kick_refresh(gboolean data_updated) { static int updates = 0; time_t now = time(NULL); if (data_updated) { updates++; } if(refresh_timer == NULL) { refresh_timer = mainloop_timer_add("refresh", 2000, FALSE, mon_trigger_refresh, NULL); } /* Refresh * - immediately if the last update was more than 5s ago * - every 10 cib-updates * - at most 2s after the last update */ if ((now - last_refresh) > (options.reconnect_msec / 1000)) { mainloop_set_trigger(refresh_trigger); mainloop_timer_stop(refresh_timer); updates = 0; } else if(updates >= 10) { mainloop_set_trigger(refresh_trigger); mainloop_timer_stop(refresh_timer); updates = 0; } else { mainloop_timer_start(refresh_timer); } } static void mon_st_callback_display(stonith_t * st, stonith_event_t * e) { if (st->state == stonith_disconnected) { /* disconnect cib as well and have everything reconnect */ mon_cib_connection_destroy(NULL); } else { print_dot(output_format); kick_refresh(TRUE); } } static void clean_up_connections(void) { if (cib != NULL) { cib->cmds->signoff(cib); cib_delete(cib); cib = NULL; } if (st != NULL) { if (st->state != stonith_disconnected) { st->cmds->remove_notification(st, T_STONITH_NOTIFY_DISCONNECT); st->cmds->remove_notification(st, T_STONITH_NOTIFY_FENCE); st->cmds->remove_notification(st, T_STONITH_NOTIFY_HISTORY); st->cmds->disconnect(st); } stonith_api_delete(st); st = NULL; } } static void handle_html_output(crm_exit_t exit_code) { xmlNodePtr html = NULL; out->finish(out, exit_code, false, (void **) &html); pcmk__html_add_header(html, "meta", "http-equiv", "refresh", "content", crm_itoa(options.reconnect_msec/1000), NULL); htmlDocDump(out->dest, html->doc); } /* * De-init ncurses, disconnect from the CIB manager, disconnect fencing, * deallocate memory and show usage-message if requested. * * We don't actually return, but nominally returning crm_exit_t allows a usage * like "return clean_up(exit_code);" which helps static analysis understand the * code flow. */ static crm_exit_t clean_up(crm_exit_t exit_code) { #if CURSES_ENABLED if (output_format == mon_output_console) { output_format = mon_output_plain; echo(); nocbreak(); endwin(); } #endif clean_up_connections(); free(options.pid_file); pe_free_working_set(mon_data_set); mon_data_set = NULL; if (exit_code == CRM_EX_USAGE) { if (output_format == mon_output_cgi) { fprintf(stdout, "Content-Type: text/plain\n" "Status: 500\n\n"); } else { fprintf(stderr, "%s", g_option_context_get_help(context, TRUE, NULL)); } } pcmk__free_arg_context(context); if (out != NULL) { if (output_format == mon_output_cgi || output_format == mon_output_html) { handle_html_output(exit_code); } else { out->finish(out, exit_code, true, NULL); } pcmk__output_free(out); } crm_exit(exit_code); } diff --git a/tools/crm_mon.h b/tools/crm_mon.h index aa334a7eef..985e317e3e 100644 --- a/tools/crm_mon.h +++ b/tools/crm_mon.h @@ -1,129 +1,128 @@ /* * Copyright 2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include /* Never display node attributes whose name starts with one of these prefixes */ #define FILTER_STR { CRM_FAIL_COUNT_PREFIX, CRM_LAST_FAILURE_PREFIX, \ "shutdown", "terminate", "standby", "probe_complete", \ "#", NULL } /* Convenience macro for prettifying output (e.g. "node" vs "nodes") */ #define s_if_plural(i) (((i) == 1)? "" : "s") #if CURSES_ENABLED # define print_dot(output_format) if (output_format == mon_output_console) { \ printw("."); \ clrtoeol(); \ refresh(); \ } else { \ fprintf(stdout, "."); \ } #else # define print_dot(output_format) fprintf(stdout, "."); #endif #if CURSES_ENABLED # define print_as(output_format, fmt, args...) if (output_format == mon_output_console) { \ printw(fmt, ##args); \ clrtoeol(); \ refresh(); \ } else { \ fprintf(stdout, fmt, ##args); \ } #else # define print_as(output_format, fmt, args...) fprintf(stdout, fmt, ##args); #endif typedef enum mon_output_format_e { mon_output_unset, mon_output_none, mon_output_monitor, mon_output_plain, mon_output_console, mon_output_xml, mon_output_legacy_xml, mon_output_html, mon_output_cgi } mon_output_format_t; #define mon_show_times (0x0001U) #define mon_show_stack (0x0002U) #define mon_show_dc (0x0004U) #define mon_show_count (0x0008U) #define mon_show_nodes (0x0010U) #define mon_show_resources (0x0020U) #define mon_show_attributes (0x0040U) #define mon_show_failcounts (0x0080U) #define mon_show_operations (0x0100U) #define mon_show_tickets (0x0200U) #define mon_show_bans (0x0400U) #define mon_show_fence_history (0x0800U) #define mon_show_headers (mon_show_times | mon_show_stack | mon_show_dc \ | mon_show_count) #define mon_show_default (mon_show_headers | mon_show_nodes \ | mon_show_resources) #define mon_show_all (mon_show_default | mon_show_attributes \ | mon_show_failcounts | mon_show_operations \ | mon_show_tickets | mon_show_bans \ | mon_show_fence_history) #define mon_op_group_by_node (0x0001U) #define mon_op_inactive_resources (0x0002U) #define mon_op_one_shot (0x0004U) #define mon_op_has_warnings (0x0008U) #define mon_op_print_timing (0x0010U) #define mon_op_watch_fencing (0x0020U) #define mon_op_fence_history (0x0040U) #define mon_op_fence_full_history (0x0080U) #define mon_op_fence_connect (0x0100U) #define mon_op_print_brief (0x0200U) #define mon_op_print_pending (0x0400U) #define mon_op_print_clone_detail (0x0800U) #define mon_op_default (mon_op_print_pending) void print_status(pcmk__output_t *out, mon_output_format_t output_format, pe_working_set_t *data_set, stonith_history_t *stonith_history, unsigned int mon_ops, unsigned int show, const char *prefix); void print_xml_status(pcmk__output_t *out, pe_working_set_t *data_set, stonith_history_t *stonith_history, unsigned int mon_ops, unsigned int show, const char *prefix); int print_html_status(pcmk__output_t *out, mon_output_format_t output_format, pe_working_set_t *data_set, stonith_history_t *stonith_history, unsigned int mon_ops, unsigned int show, const char *prefix); GList *append_attr_list(GList *attr_list, char *name); void blank_screen(void); -int count_resources(pe_working_set_t *data_set, resource_t *rsc); void crm_mon_get_parameters(resource_t *rsc, pe_working_set_t *data_set); const char *get_cluster_stack(pe_working_set_t *data_set); char *get_node_display_name(node_t *node, unsigned int mon_ops); int get_resource_display_options(unsigned int mon_ops, mon_output_format_t output_format); void crm_mon_register_messages(pcmk__output_t *out); pcmk__output_t *crm_mon_mk_curses_output(char **argv); void curses_indented_printf(pcmk__output_t *out, const char *format, ...) G_GNUC_PRINTF(2, 3); void curses_indented_vprintf(pcmk__output_t *out, const char *format, va_list args) G_GNUC_PRINTF(2, 0); #if CURSES_ENABLED extern GOptionEntry crm_mon_curses_output_entries[]; #define CRM_MON_SUPPORTED_FORMAT_CURSES { "console", crm_mon_mk_curses_output, crm_mon_curses_output_entries } #endif pcmk__output_t *crm_mon_mk_xml_output(char **argv); #define CRM_MON_SUPPORTED_FORMAT_XML { "xml", crm_mon_mk_xml_output, pcmk__xml_output_entries } diff --git a/tools/crm_mon_output.c b/tools/crm_mon_output.c index 822132a108..9a9f3ce74b 100644 --- a/tools/crm_mon_output.c +++ b/tools/crm_mon_output.c @@ -1,1189 +1,1203 @@ /* * Copyright 2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include #include #include "crm_mon.h" static char * time_t_string(time_t when) { crm_time_t *crm_when = crm_time_new(NULL); char *buf = NULL; crm_time_set_timet(crm_when, &when); buf = crm_time_as_string(crm_when, crm_time_log_date | crm_time_log_timeofday | crm_time_log_with_timezone); crm_time_free(crm_when); return buf; } static char * failed_action_string(xmlNodePtr xml_op) { const char *op_key = crm_element_value(xml_op, XML_LRM_ATTR_TASK_KEY); int rc = crm_parse_int(crm_element_value(xml_op, XML_LRM_ATTR_RC), "0"); int status = crm_parse_int(crm_element_value(xml_op, XML_LRM_ATTR_OPSTATUS), "0"); const char *exit_reason = crm_element_value(xml_op, XML_LRM_ATTR_EXIT_REASON); time_t last_change = 0; if (crm_element_value_epoch(xml_op, XML_RSC_OP_LAST_CHANGE, &last_change) == pcmk_ok) { char *time = time_t_string(last_change); char *buf = crm_strdup_printf("%s on %s '%s' (%d): call=%s, status='%s', exitreason='%s', last-rc-change='%s', queued=%sms, exec=%sms", op_key ? op_key : ID(xml_op), crm_element_value(xml_op, XML_ATTR_UNAME), services_ocf_exitcode_str(rc), rc, crm_element_value(xml_op, XML_LRM_ATTR_CALLID), services_lrm_status_str(status), exit_reason ? exit_reason : "none", time, crm_element_value(xml_op, XML_RSC_OP_T_QUEUE), crm_element_value(xml_op, XML_RSC_OP_T_EXEC)); free(time); return buf; } else { return crm_strdup_printf("%s on %s '%s' (%d): call=%s, status=%s, exitreason='%s'", op_key ? op_key : ID(xml_op), crm_element_value(xml_op, XML_ATTR_UNAME), services_ocf_exitcode_str(rc), rc, crm_element_value(xml_op, XML_LRM_ATTR_CALLID), services_lrm_status_str(status), exit_reason ? exit_reason : "none"); } } static char * last_changed_string(const char *last_written, const char *user, const char *client, const char *origin) { if (last_written != NULL || user != NULL || client != NULL || origin != NULL) { return crm_strdup_printf("%s%s%s%s%s%s%s", last_written ? last_written : "", user ? " by " : "", user ? user : "", client ? " via " : "", client ? client : "", origin ? " on " : "", origin ? origin : ""); } else { return strdup(""); } } static char * op_history_string(xmlNode *xml_op, const char *task, const char *interval_ms_s, int rc, unsigned int mon_ops) { const char *call = crm_element_value(xml_op, XML_LRM_ATTR_CALLID); char *interval_str = NULL; char *buf = NULL; if (interval_ms_s && safe_str_neq(interval_ms_s, "0")) { char *pair = pcmk_format_nvpair("interval", interval_ms_s, "ms"); interval_str = crm_strdup_printf(" %s", pair); free(pair); } if (is_set(mon_ops, mon_op_print_timing)) { char *last_change_str = NULL; char *last_run_str = NULL; char *exec_str = NULL; char *queue_str = NULL; const char *value = NULL; time_t epoch = 0; if ((crm_element_value_epoch(xml_op, XML_RSC_OP_LAST_CHANGE, &epoch) == pcmk_ok) && (epoch > 0)) { char *time = pcmk_format_named_time(XML_RSC_OP_LAST_CHANGE, epoch); last_change_str = crm_strdup_printf(" %s", time); free(time); } if ((crm_element_value_epoch(xml_op, XML_RSC_OP_LAST_RUN, &epoch) == pcmk_ok) && (epoch > 0)) { char *time = pcmk_format_named_time(XML_RSC_OP_LAST_RUN, epoch); last_run_str = crm_strdup_printf(" %s", time); free(time); } value = crm_element_value(xml_op, XML_RSC_OP_T_EXEC); if (value) { char *pair = pcmk_format_nvpair(XML_RSC_OP_T_EXEC, value, "ms"); exec_str = crm_strdup_printf(" %s", pair); free(pair); } value = crm_element_value(xml_op, XML_RSC_OP_T_QUEUE); if (value) { char *pair = pcmk_format_nvpair(XML_RSC_OP_T_QUEUE, value, "ms"); queue_str = crm_strdup_printf(" %s", pair); free(pair); } buf = crm_strdup_printf("(%s) %s:%s%s%s%s%s rc=%d (%s)", call, task, interval_str ? interval_str : "", last_change_str ? last_change_str : "", last_run_str ? last_run_str : "", exec_str ? exec_str : "", queue_str ? queue_str : "", rc, services_ocf_exitcode_str(rc)); if (last_change_str) { free(last_change_str); } if (last_run_str) { free(last_run_str); } if (exec_str) { free(exec_str); } if (queue_str) { free(queue_str); } } else { buf = crm_strdup_printf("(%s) %s:%s", call, task, interval_str ? interval_str : ""); } if (interval_str) { free(interval_str); } return buf; } static char * resource_history_string(resource_t *rsc, const char *rsc_id, gboolean all, int failcount, time_t last_failure) { char *buf = NULL; if (rsc == NULL) { buf = crm_strdup_printf("%s: orphan", rsc_id); } else if (all || failcount || last_failure > 0) { char *failcount_s = failcount > 0 ? crm_strdup_printf(" %s=%d", CRM_FAIL_COUNT_PREFIX, failcount) : strdup(""); char *lastfail_s = last_failure > 0 ? crm_strdup_printf(" %s=%s", CRM_LAST_FAILURE_PREFIX, crm_now_string(&last_failure)) : strdup(""); buf = crm_strdup_printf("%s: migration-threshold=%d%s%s", rsc_id, rsc->migration_threshold, failcount_s, lastfail_s); free(failcount_s); free(lastfail_s); } else { buf = crm_strdup_printf("%s:", rsc_id); } return buf; } static int ban_html(pcmk__output_t *out, va_list args) { pe_node_t *pe_node = va_arg(args, pe_node_t *); pe__location_t *location = va_arg(args, pe__location_t *); unsigned int mon_ops = va_arg(args, unsigned int); char *node_name = get_node_display_name(pe_node, mon_ops); char *buf = crm_strdup_printf("%s\tprevents %s from running %son %s", location->id, location->rsc_lh->id, location->role_filter == RSC_ROLE_MASTER ? "as Master " : "", node_name); pcmk__output_create_html_node(out, "li", NULL, NULL, buf); free(node_name); free(buf); return 0; } static int ban_text(pcmk__output_t *out, va_list args) { pe_node_t *pe_node = va_arg(args, pe_node_t *); pe__location_t *location = va_arg(args, pe__location_t *); unsigned int mon_ops = va_arg(args, unsigned int); char *node_name = get_node_display_name(pe_node, mon_ops); out->list_item(out, NULL, "%s\tprevents %s from running %son %s", location->id, location->rsc_lh->id, location->role_filter == RSC_ROLE_MASTER ? "as Master " : "", node_name); free(node_name); return 0; } static int ban_xml(pcmk__output_t *out, va_list args) { xmlNodePtr node = pcmk__output_create_xml_node(out, "ban"); pe_node_t *pe_node = va_arg(args, pe_node_t *); pe__location_t *location = va_arg(args, pe__location_t *); char *weight_s = crm_itoa(pe_node->weight); xmlSetProp(node, (pcmkXmlStr) "id", (pcmkXmlStr) location->id); xmlSetProp(node, (pcmkXmlStr) "resource", (pcmkXmlStr) location->rsc_lh->id); xmlSetProp(node, (pcmkXmlStr) "node", (pcmkXmlStr) pe_node->details->uname); xmlSetProp(node, (pcmkXmlStr) "weight", (pcmkXmlStr) weight_s); xmlSetProp(node, (pcmkXmlStr) "master_only", (pcmkXmlStr) (location->role_filter == RSC_ROLE_MASTER ? "true" : "false")); free(weight_s); return 0; } static int cluster_counts_html(pcmk__output_t *out, va_list args) { xmlNodePtr nodes_node = pcmk__output_create_xml_node(out, "li"); xmlNodePtr resources_node = pcmk__output_create_xml_node(out, "li"); unsigned int nnodes = va_arg(args, unsigned int); unsigned int nresources = va_arg(args, unsigned int); unsigned int ndisabled = va_arg(args, unsigned int); unsigned int nblocked = va_arg(args, unsigned int); char *nnodes_str = crm_strdup_printf("%d node%s configured", nnodes, s_if_plural(nnodes)); pcmk_create_html_node(nodes_node, "span", NULL, NULL, nnodes_str); free(nnodes_str); if (ndisabled && nblocked) { - char *s = crm_strdup_printf("%d resource%s configured (%d ", nresources, s_if_plural(nresources), ndisabled); + char *s = crm_strdup_printf("%d resource instance%s configured (%d ", + nresources, s_if_plural(nresources), + ndisabled); pcmk_create_html_node(resources_node, "span", NULL, NULL, s); free(s); pcmk_create_html_node(resources_node, "span", NULL, "bold", "DISABLED"); s = crm_strdup_printf(", %d ", nblocked); pcmk_create_html_node(resources_node, "span", NULL, NULL, s); free(s); pcmk_create_html_node(resources_node, "span", NULL, "bold", "BLOCKED"); pcmk_create_html_node(resources_node, "span", NULL, NULL, " from starting due to failure)"); } else if (ndisabled && !nblocked) { - char *s = crm_strdup_printf("%d resource%s configured (%d ", nresources, s_if_plural(nresources), ndisabled); + char *s = crm_strdup_printf("%d resource instance%s configured (%d ", + nresources, s_if_plural(nresources), + ndisabled); pcmk_create_html_node(resources_node, "span", NULL, NULL, s); free(s); pcmk_create_html_node(resources_node, "span", NULL, "bold", "DISABLED"); pcmk_create_html_node(resources_node, "span", NULL, NULL, ")"); } else if (!ndisabled && nblocked) { - char *s = crm_strdup_printf("%d resource%s configured (%d ", nresources, s_if_plural(nresources), nblocked); + char *s = crm_strdup_printf("%d resource instance%s configured (%d ", + nresources, s_if_plural(nresources), + nblocked); pcmk_create_html_node(resources_node, "span", NULL, NULL, s); free(s); pcmk_create_html_node(resources_node, "span", NULL, "bold", "BLOCKED"); pcmk_create_html_node(resources_node, "span", NULL, NULL, " from starting due to failure)"); } else { - char *s = crm_strdup_printf("%d resource%s configured", nresources, s_if_plural(nresources)); + char *s = crm_strdup_printf("%d resource instance%s configured", + nresources, s_if_plural(nresources)); pcmk_create_html_node(resources_node, "span", NULL, NULL, s); free(s); } return 0; } static int cluster_counts_text(pcmk__output_t *out, va_list args) { unsigned int nnodes = va_arg(args, unsigned int); unsigned int nresources = va_arg(args, unsigned int); unsigned int ndisabled = va_arg(args, unsigned int); unsigned int nblocked = va_arg(args, unsigned int); out->list_item(out, NULL, "%d node%s configured", nnodes, s_if_plural(nnodes)); if (ndisabled && nblocked) { - out->list_item(out, NULL, "%d resource%s configured (%d DISABLED, %d BLOCKED from starting due to failure", - nresources, s_if_plural(nresources), ndisabled, nblocked); + out->list_item(out, NULL, "%d resource instance%s configured " + "(%d DISABLED, %d BLOCKED from " + "further action due to failure)", + nresources, s_if_plural(nresources), ndisabled, + nblocked); } else if (ndisabled && !nblocked) { - out->list_item(out, NULL, "%d resource%s configured (%d DISABLED)", + out->list_item(out, NULL, "%d resource instance%s configured " + "(%d DISABLED)", nresources, s_if_plural(nresources), ndisabled); } else if (!ndisabled && nblocked) { - out->list_item(out, NULL, "%d resource%s configured (%d BLOCKED from starting due to failure)", + out->list_item(out, NULL, "%d resource instance%s configured " + "(%d BLOCKED from further action " + "due to failure)", nresources, s_if_plural(nresources), nblocked); } else { - out->list_item(out, NULL, "%d resource%s configured", nresources, s_if_plural(nresources)); + out->list_item(out, NULL, "%d resource instance%s configured", + nresources, s_if_plural(nresources)); } return 0; } static int cluster_counts_xml(pcmk__output_t *out, va_list args) { xmlNodePtr nodes_node = pcmk__output_create_xml_node(out, "nodes_configured"); xmlNodePtr resources_node = pcmk__output_create_xml_node(out, "resources_configured"); unsigned int nnodes = va_arg(args, unsigned int); unsigned int nresources = va_arg(args, unsigned int); unsigned int ndisabled = va_arg(args, unsigned int); unsigned int nblocked = va_arg(args, unsigned int); char *s = crm_itoa(nnodes); xmlSetProp(nodes_node, (pcmkXmlStr) "number", (pcmkXmlStr) s); free(s); s = crm_itoa(nresources); xmlSetProp(resources_node, (pcmkXmlStr) "number", (pcmkXmlStr) s); free(s); s = crm_itoa(ndisabled); xmlSetProp(resources_node, (pcmkXmlStr) "disabled", (pcmkXmlStr) s); free(s); s = crm_itoa(nblocked); xmlSetProp(resources_node, (pcmkXmlStr) "blocked", (pcmkXmlStr) s); free(s); return 0; } static int cluster_dc_html(pcmk__output_t *out, va_list args) { xmlNodePtr node = pcmk__output_create_xml_node(out, "li"); node_t *dc = va_arg(args, node_t *); const char *quorum = va_arg(args, const char *); const char *dc_version_s = va_arg(args, const char *); const char *dc_name = va_arg(args, const char *); pcmk_create_html_node(node, "span", NULL, "bold", "Current DC: "); if (dc) { if (crm_is_true(quorum)) { char *buf = crm_strdup_printf("%s (version %s) - partition with quorum", dc_name, dc_version_s ? dc_version_s : "unknown"); pcmk_create_html_node(node, "span", NULL, NULL, buf); free(buf); } else { char *buf = crm_strdup_printf("%s (version %s) - partition", dc_name, dc_version_s ? dc_version_s : "unknown"); pcmk_create_html_node(node, "span", NULL, NULL, buf); free(buf); pcmk_create_html_node(node, "span", NULL, "warning", "WITHOUT"); pcmk_create_html_node(node, "span", NULL, NULL, "quorum"); } } else { pcmk_create_html_node(node ,"span", NULL, "warning", "NONE"); } return 0; } static int cluster_dc_text(pcmk__output_t *out, va_list args) { node_t *dc = va_arg(args, node_t *); const char *quorum = va_arg(args, const char *); const char *dc_version_s = va_arg(args, const char *); const char *dc_name = va_arg(args, const char *); if (dc) { out->list_item(out, "Current DC", "%s (version %s) - partition %s quorum", dc_name, dc_version_s ? dc_version_s : "unknown", crm_is_true(quorum) ? "with" : "WITHOUT"); } else { out->list_item(out, "Current DC", "NONE"); } return 0; } static int cluster_dc_xml(pcmk__output_t *out, va_list args) { xmlNodePtr node = pcmk__output_create_xml_node(out, "current_dc"); node_t *dc = va_arg(args, node_t *); const char *quorum = va_arg(args, const char *); const char *dc_version_s = va_arg(args, const char *); if (dc) { xmlSetProp(node, (pcmkXmlStr) "present", (pcmkXmlStr) "true"); xmlSetProp(node, (pcmkXmlStr) "version", (pcmkXmlStr) (dc_version_s ? dc_version_s : "")); xmlSetProp(node, (pcmkXmlStr) "name", (pcmkXmlStr) dc->details->uname); xmlSetProp(node, (pcmkXmlStr) "id", (pcmkXmlStr) dc->details->id); xmlSetProp(node, (pcmkXmlStr) "with_quorum", (pcmkXmlStr) (crm_is_true(quorum) ? "true" : "false")); } else { xmlSetProp(node, (pcmkXmlStr) "present", (pcmkXmlStr) "false"); } return 0; } static int cluster_options_html(pcmk__output_t *out, va_list args) { pe_working_set_t *data_set = va_arg(args, pe_working_set_t *); /* Kind of a hack - close the list started by print_cluster_summary so we * can put all the options in their own list, but just for HTML output. */ out->end_list(out); /* And then this list will be closed by print_cluster_summary since it * wants to close the list it created unconditionally. */ out->begin_list(out, NULL, NULL, "Config Options"); out->list_item(out, NULL, "STONITH of failed nodes %s", is_set(data_set->flags, pe_flag_stonith_enabled) ? "enabled" : "disabled"); out->list_item(out, NULL, "Cluster is %s", is_set(data_set->flags, pe_flag_symmetric_cluster) ? "symmetric" : "asymmetric"); switch (data_set->no_quorum_policy) { case no_quorum_freeze: out->list_item(out, NULL, "No Quorum policy: Freeze resources"); break; case no_quorum_stop: out->list_item(out, NULL, "No Quorum policy: Stop ALL resources"); break; case no_quorum_ignore: out->list_item(out, NULL, "No Quorum policy: Ignore"); break; case no_quorum_suicide: out->list_item(out, NULL, "No Quorum policy: Suicide"); break; } if (is_set(data_set->flags, pe_flag_maintenance_mode)) { xmlNodePtr node = pcmk__output_create_xml_node(out, "li"); pcmk_create_html_node(node, "span", NULL, "bold", "DISABLED"); pcmk_create_html_node(node, "span", NULL, NULL, " (the cluster will not attempt to start, stop, or recover services)"); } else { out->list_item(out, NULL, "Resource management enabled"); } return 0; } static int cluster_options_text(pcmk__output_t *out, va_list args) { pe_working_set_t *data_set = va_arg(args, pe_working_set_t *); if (is_set(data_set->flags, pe_flag_maintenance_mode)) { fprintf(out->dest, "\n *** Resource management is DISABLED ***"); fprintf(out->dest, "\n The cluster will not attempt to start, stop or recover services"); fprintf(out->dest, "\n"); } return 0; } static int cluster_options_xml(pcmk__output_t *out, va_list args) { xmlNodePtr node = pcmk__output_create_xml_node(out, "cluster_options"); pe_working_set_t *data_set = va_arg(args, pe_working_set_t *); xmlSetProp(node, (pcmkXmlStr) "stonith-enabled", (pcmkXmlStr) (is_set(data_set->flags, pe_flag_stonith_enabled) ? "true" : "false")); xmlSetProp(node, (pcmkXmlStr) "symmetric-cluster", (pcmkXmlStr) (is_set(data_set->flags, pe_flag_symmetric_cluster) ? "true" : "false")); switch (data_set->no_quorum_policy) { case no_quorum_freeze: xmlSetProp(node, (pcmkXmlStr) "no-quorum-policy", (pcmkXmlStr) "freeze"); break; case no_quorum_stop: xmlSetProp(node, (pcmkXmlStr) "no-quorum-policy", (pcmkXmlStr) "stop"); break; case no_quorum_ignore: xmlSetProp(node, (pcmkXmlStr) "no-quorum-policy", (pcmkXmlStr) "ignore"); break; case no_quorum_suicide: xmlSetProp(node, (pcmkXmlStr) "no-quorum-policy", (pcmkXmlStr) "suicide"); break; } xmlSetProp(node, (pcmkXmlStr) "maintenance-mode", (pcmkXmlStr) (is_set(data_set->flags, pe_flag_maintenance_mode) ? "true" : "false")); return 0; } static int cluster_stack_html(pcmk__output_t *out, va_list args) { xmlNodePtr node = pcmk__output_create_xml_node(out, "li"); const char *stack_s = va_arg(args, const char *); pcmk_create_html_node(node, "span", NULL, "bold", "Stack: "); pcmk_create_html_node(node, "span", NULL, NULL, stack_s); return 0; } static int cluster_stack_text(pcmk__output_t *out, va_list args) { const char *stack_s = va_arg(args, const char *); out->list_item(out, "Stack", "%s", stack_s); return 0; } static int cluster_stack_xml(pcmk__output_t *out, va_list args) { xmlNodePtr node = pcmk__output_create_xml_node(out, "stack"); const char *stack_s = va_arg(args, const char *); xmlSetProp(node, (pcmkXmlStr) "type", (pcmkXmlStr) stack_s); return 0; } static int cluster_times_html(pcmk__output_t *out, va_list args) { xmlNodePtr updated_node = pcmk__output_create_xml_node(out, "li"); xmlNodePtr changed_node = pcmk__output_create_xml_node(out, "li"); const char *last_written = va_arg(args, const char *); const char *user = va_arg(args, const char *); const char *client = va_arg(args, const char *); const char *origin = va_arg(args, const char *); char *buf = last_changed_string(last_written, user, client, origin); pcmk_create_html_node(updated_node, "span", NULL, "bold", "Last updated: "); pcmk_create_html_node(updated_node, "span", NULL, NULL, crm_now_string(NULL)); pcmk_create_html_node(changed_node, "span", NULL, "bold", "Last change: "); pcmk_create_html_node(changed_node, "span", NULL, NULL, buf); free(buf); return 0; } static int cluster_times_xml(pcmk__output_t *out, va_list args) { xmlNodePtr updated_node = pcmk__output_create_xml_node(out, "last_update"); xmlNodePtr changed_node = pcmk__output_create_xml_node(out, "last_change"); const char *last_written = va_arg(args, const char *); const char *user = va_arg(args, const char *); const char *client = va_arg(args, const char *); const char *origin = va_arg(args, const char *); xmlSetProp(updated_node, (pcmkXmlStr) "time", (pcmkXmlStr) crm_now_string(NULL)); xmlSetProp(changed_node, (pcmkXmlStr) "time", (pcmkXmlStr) (last_written ? last_written : "")); xmlSetProp(changed_node, (pcmkXmlStr) "user", (pcmkXmlStr) (user ? user : "")); xmlSetProp(changed_node, (pcmkXmlStr) "client", (pcmkXmlStr) (client ? client : "")); xmlSetProp(changed_node, (pcmkXmlStr) "origin", (pcmkXmlStr) (origin ? origin : "")); return 0; } static int cluster_times_text(pcmk__output_t *out, va_list args) { const char *last_written = va_arg(args, const char *); const char *user = va_arg(args, const char *); const char *client = va_arg(args, const char *); const char *origin = va_arg(args, const char *); char *buf = last_changed_string(last_written, user, client, origin); out->list_item(out, "Last updated", "%s", crm_now_string(NULL)); out->list_item(out, "Last change", " %s", buf); free(buf); return 0; } static int failed_action_console(pcmk__output_t *out, va_list args) { xmlNodePtr xml_op = va_arg(args, xmlNodePtr); char *s = failed_action_string(xml_op); curses_indented_printf(out, "%s\n", s); free(s); return 0; } static int failed_action_html(pcmk__output_t *out, va_list args) { xmlNodePtr xml_op = va_arg(args, xmlNodePtr); char *s = failed_action_string(xml_op); pcmk__output_create_html_node(out, "li", NULL, NULL, s); free(s); return 0; } static int failed_action_text(pcmk__output_t *out, va_list args) { xmlNodePtr xml_op = va_arg(args, xmlNodePtr); char *s = failed_action_string(xml_op); pcmk__indented_printf(out, "%s\n", s); free(s); return 0; } static int failed_action_xml(pcmk__output_t *out, va_list args) { xmlNodePtr xml_op = va_arg(args, xmlNodePtr); const char *op_key = crm_element_value(xml_op, XML_LRM_ATTR_TASK_KEY); const char *last = crm_element_value(xml_op, XML_RSC_OP_LAST_CHANGE); int rc = crm_parse_int(crm_element_value(xml_op, XML_LRM_ATTR_RC), "0"); int status = crm_parse_int(crm_element_value(xml_op, XML_LRM_ATTR_OPSTATUS), "0"); const char *exit_reason = crm_element_value(xml_op, XML_LRM_ATTR_EXIT_REASON); char *rc_s = crm_itoa(rc); char *reason_s = crm_xml_escape(exit_reason ? exit_reason : "none"); xmlNodePtr node = pcmk__output_create_xml_node(out, "failure"); xmlSetProp(node, (pcmkXmlStr) (op_key ? "op_key" : "id"), (pcmkXmlStr) (op_key ? op_key : "id")); xmlSetProp(node, (pcmkXmlStr) "node", (pcmkXmlStr) crm_element_value(xml_op, XML_ATTR_UNAME)); xmlSetProp(node, (pcmkXmlStr) "exitstatus", (pcmkXmlStr) services_ocf_exitcode_str(rc)); xmlSetProp(node, (pcmkXmlStr) "exitreason", (pcmkXmlStr) reason_s); xmlSetProp(node, (pcmkXmlStr) "exitcode", (pcmkXmlStr) rc_s); xmlSetProp(node, (pcmkXmlStr) "call", (pcmkXmlStr) crm_element_value(xml_op, XML_LRM_ATTR_CALLID)); xmlSetProp(node, (pcmkXmlStr) "status", (pcmkXmlStr) services_lrm_status_str(status)); if (last) { char *s = crm_itoa(crm_parse_ms(crm_element_value(xml_op, XML_LRM_ATTR_INTERVAL_MS))); char *rc_change = time_t_string(crm_parse_int(last, "0")); xmlSetProp(node, (pcmkXmlStr) "last-rc-change", (pcmkXmlStr) rc_change); xmlSetProp(node, (pcmkXmlStr) "queued", (pcmkXmlStr) crm_element_value(xml_op, XML_RSC_OP_T_QUEUE)); xmlSetProp(node, (pcmkXmlStr) "exec", (pcmkXmlStr) crm_element_value(xml_op, XML_RSC_OP_T_EXEC)); xmlSetProp(node, (pcmkXmlStr) "interval", (pcmkXmlStr) s); xmlSetProp(node, (pcmkXmlStr) "task", (pcmkXmlStr) crm_element_value(xml_op, XML_LRM_ATTR_TASK)); free(s); free(rc_change); } free(reason_s); free(rc_s); return 0; } static int node_html(pcmk__output_t *out, va_list args) { node_t *node = va_arg(args, node_t *); unsigned int mon_ops = va_arg(args, unsigned int); gboolean full = va_arg(args, gboolean); char *node_name = get_node_display_name(node, mon_ops); char *buf = crm_strdup_printf("Node: %s", node_name); int print_opts = get_resource_display_options(mon_ops, mon_output_html); if (full) { xmlNodePtr item_node = pcmk__output_create_xml_node(out, "li"); pcmk_create_html_node(item_node, "span", NULL, NULL, buf); if (node->details->standby_onfail && node->details->online) { pcmk_create_html_node(item_node, "span", NULL, "standby", " standby (on-fail)"); } else if (node->details->standby && node->details->online) { char *s = crm_strdup_printf(" standby%s", node->details->running_rsc ? " (with active resources)" : ""); pcmk_create_html_node(item_node, "span", NULL, " standby", s); free(s); } else if (node->details->standby) { pcmk_create_html_node(item_node, "span", NULL, "offline", " OFFLINE (standby)"); } else if (node->details->maintenance && node->details->online) { pcmk_create_html_node(item_node, "span", NULL, "maint", " maintenance"); } else if (node->details->maintenance) { pcmk_create_html_node(item_node, "span", NULL, "offline", " OFFLINE (maintenance)"); } else if (node->details->online) { pcmk_create_html_node(item_node, "span", NULL, "online", " online"); } else { pcmk_create_html_node(item_node, "span", NULL, "offline", " OFFLINE"); } if (is_set(mon_ops, mon_op_print_brief) && is_set(mon_ops, mon_op_group_by_node)) { out->begin_list(out, NULL, NULL, NULL); pe__rscs_brief_output(out, node->details->running_rsc, print_opts | pe_print_rsconly, FALSE); out->end_list(out); } else if (is_set(mon_ops, mon_op_group_by_node)) { GListPtr lpc2 = NULL; out->begin_list(out, NULL, NULL, NULL); for (lpc2 = node->details->running_rsc; lpc2 != NULL; lpc2 = lpc2->next) { resource_t *rsc = (resource_t *) lpc2->data; out->message(out, crm_map_element_name(rsc->xml), print_opts | pe_print_rsconly, rsc); } out->end_list(out); } } else { out->begin_list(out, NULL, NULL, "%s", buf); } free(buf); free(node_name); return 0; } static int node_text(pcmk__output_t *out, va_list args) { node_t *node = va_arg(args, node_t *); unsigned int mon_ops = va_arg(args, unsigned int); gboolean full = va_arg(args, gboolean); if (full) { const char *node_mode = va_arg(args, const char *); char *node_name = get_node_display_name(node, mon_ops); int print_opts = get_resource_display_options(mon_ops, mon_output_xml); char *buf = NULL; /* Print the node name and status */ if (pe__is_guest_node(node)) { buf = crm_strdup_printf("GuestNode %s: %s", node_name, node_mode); } else if (pe__is_remote_node(node)) { buf = crm_strdup_printf("RemoteNode %s: %s", node_name, node_mode); } else { buf = crm_strdup_printf("Node %s: %s", node_name, node_mode); } /* If we're grouping by node, print its resources */ if (is_set(mon_ops, mon_op_group_by_node)) { out->begin_list(out, NULL, NULL, "%s", buf); out->begin_list(out, NULL, NULL, "Resources"); if (is_set(mon_ops, mon_op_print_brief)) { pe__rscs_brief_output(out, node->details->running_rsc, print_opts | pe_print_rsconly, FALSE); } else { GListPtr gIter2 = NULL; for (gIter2 = node->details->running_rsc; gIter2 != NULL; gIter2 = gIter2->next) { resource_t *rsc = (resource_t *) gIter2->data; out->message(out, crm_map_element_name(rsc->xml), print_opts | pe_print_rsconly, rsc); } } out->end_list(out); out->end_list(out); } else { out->list_item(out, NULL, "%s", buf); } free(buf); free(node_name); } else { out->begin_list(out, NULL, NULL, "Node: %s", get_node_display_name(node, mon_ops)); } return 0; } static int node_xml(pcmk__output_t *out, va_list args) { node_t *node = va_arg(args, node_t *); unsigned int mon_ops G_GNUC_UNUSED = va_arg(args, unsigned int); gboolean full = va_arg(args, gboolean); if (full) { const char *node_type = "unknown"; int print_opts = get_resource_display_options(mon_ops, mon_output_xml); char *length_s = crm_itoa(g_list_length(node->details->running_rsc)); switch (node->details->type) { case node_member: node_type = "member"; break; case node_remote: node_type = "remote"; break; case node_ping: node_type = "ping"; break; } pe__name_and_nvpairs_xml(out, true, "node", 13, "name", node->details->uname, "id", node->details->id, "online", node->details->online ? "true" : "false", "standby", node->details->standby ? "true" : "false", "standby_onfail", node->details->standby_onfail ? "true" : "false", "maintenance", node->details->maintenance ? "true" : "false", "pending", node->details->pending ? "true" : "false", "unclean", node->details->unclean ? "true" : "false", "shutdown", node->details->shutdown ? "true" : "false", "expected_up", node->details->expected_up ? "true" : "false", "is_dc", node->details->is_dc ? "true" : "false", "resources_running", length_s, "type", node_type); if (pe__is_guest_node(node)) { xmlNodePtr xml_node = pcmk__output_xml_peek_parent(out); xmlSetProp(xml_node, (pcmkXmlStr) "id_as_resource", (pcmkXmlStr) node->details->remote_rsc->container->id); } if (is_set(mon_ops, mon_op_group_by_node)) { GListPtr lpc = NULL; for (lpc = node->details->running_rsc; lpc != NULL; lpc = lpc->next) { resource_t *rsc = (resource_t *) lpc->data; out->message(out, crm_map_element_name(rsc->xml), print_opts | pe_print_rsconly, rsc); } } free(length_s); out->end_list(out); } else { xmlNodePtr parent = pcmk__output_xml_create_parent(out, "node"); xmlSetProp(parent, (pcmkXmlStr) "name", (pcmkXmlStr) node->details->uname); } return 0; } static int node_attribute_text(pcmk__output_t *out, va_list args) { const char *name = va_arg(args, const char *); const char *value = va_arg(args, const char *); gboolean add_extra = va_arg(args, gboolean); int expected_score = va_arg(args, int); if (add_extra) { int v = crm_parse_int(value, "0"); if (v <= 0) { out->list_item(out, NULL, "%-32s\t: %-10s\t: Connectivity is lost", name, value); } else if (v < expected_score) { out->list_item(out, NULL, "%-32s\t: %-10s\t: Connectivity is degraded (Expected=%d)", name, value, expected_score); } else { out->list_item(out, NULL, "%-32s\t: %-10s", name, value); } } else { out->list_item(out, NULL, "%-32s\t: %-10s", name, value); } return 0; } static int node_attribute_html(pcmk__output_t *out, va_list args) { const char *name = va_arg(args, const char *); const char *value = va_arg(args, const char *); gboolean add_extra = va_arg(args, gboolean); int expected_score = va_arg(args, int); if (add_extra) { int v = crm_parse_int(value, "0"); char *s = crm_strdup_printf("%s: %s", name, value); xmlNodePtr item_node = pcmk__output_create_xml_node(out, "li"); pcmk_create_html_node(item_node, "span", NULL, NULL, s); free(s); if (v <= 0) { pcmk_create_html_node(item_node, "span", NULL, "bold", "(connectivity is lost)"); } else if (v < expected_score) { char *buf = crm_strdup_printf("(connectivity is degraded -- expected %d", expected_score); pcmk_create_html_node(item_node, "span", NULL, "bold", buf); free(buf); } } else { out->list_item(out, NULL, "%s: %s", name, value); } return 0; } static int node_attribute_xml(pcmk__output_t *out, va_list args) { const char *name = va_arg(args, const char *); const char *value = va_arg(args, const char *); gboolean add_extra = va_arg(args, gboolean); int expected_score = va_arg(args, int); xmlNodePtr node = pcmk__output_create_xml_node(out, "attribute"); xmlSetProp(node, (pcmkXmlStr) "name", (pcmkXmlStr) name); xmlSetProp(node, (pcmkXmlStr) "value", (pcmkXmlStr) value); if (add_extra) { char *buf = crm_itoa(expected_score); xmlSetProp(node, (pcmkXmlStr) "expected", (pcmkXmlStr) buf); free(buf); } return 0; } static int op_history_text(pcmk__output_t *out, va_list args) { xmlNode *xml_op = va_arg(args, xmlNode *); const char *task = va_arg(args, const char *); const char *interval_ms_s = va_arg(args, const char *); int rc = va_arg(args, int); unsigned int mon_ops = va_arg(args, unsigned int); char *buf = op_history_string(xml_op, task, interval_ms_s, rc, mon_ops); out->list_item(out, NULL, "%s", buf); free(buf); return 0; } static int op_history_xml(pcmk__output_t *out, va_list args) { xmlNode *xml_op = va_arg(args, xmlNode *); const char *task = va_arg(args, const char *); const char *interval_ms_s = va_arg(args, const char *); int rc = va_arg(args, int); unsigned int mon_ops = va_arg(args, unsigned int); char *rc_s = NULL; xmlNodePtr node = pcmk__output_create_xml_node(out, "operation_history"); xmlSetProp(node, (pcmkXmlStr) "call", (pcmkXmlStr) crm_element_value(xml_op, XML_LRM_ATTR_CALLID)); xmlSetProp(node, (pcmkXmlStr) "task", (pcmkXmlStr) task); if (interval_ms_s && safe_str_neq(interval_ms_s, "0")) { char *s = crm_strdup_printf("%sms", interval_ms_s); xmlSetProp(node, (pcmkXmlStr) "interval", (pcmkXmlStr) s); free(s); } if (is_set(mon_ops, mon_op_print_timing)) { const char *value = NULL; value = crm_element_value(xml_op, XML_RSC_OP_LAST_CHANGE); if (value) { time_t int_value = (time_t) crm_parse_int(value, NULL); if (int_value > 0) { xmlSetProp(node, (pcmkXmlStr) XML_RSC_OP_LAST_CHANGE, (pcmkXmlStr) crm_now_string(&int_value)); } } value = crm_element_value(xml_op, XML_RSC_OP_LAST_RUN); if (value) { time_t int_value = (time_t) crm_parse_int(value, NULL); if (int_value > 0) { xmlSetProp(node, (pcmkXmlStr) XML_RSC_OP_LAST_RUN, (pcmkXmlStr) crm_now_string(&int_value)); } } value = crm_element_value(xml_op, XML_RSC_OP_T_EXEC); if (value) { char *s = crm_strdup_printf("%sms", value); xmlSetProp(node, (pcmkXmlStr) XML_RSC_OP_T_EXEC, (pcmkXmlStr) s); free(s); } value = crm_element_value(xml_op, XML_RSC_OP_T_QUEUE); if (value) { char *s = crm_strdup_printf("%sms", value); xmlSetProp(node, (pcmkXmlStr) XML_RSC_OP_T_QUEUE, (pcmkXmlStr) s); free(s); } } rc_s = crm_itoa(rc); xmlSetProp(node, (pcmkXmlStr) "rc", (pcmkXmlStr) rc_s); xmlSetProp(node, (pcmkXmlStr) "rc_text", (pcmkXmlStr) services_ocf_exitcode_str(rc)); free(rc_s); return 0; } static int resource_history_text(pcmk__output_t *out, va_list args) { resource_t *rsc = va_arg(args, resource_t *); const char *rsc_id = va_arg(args, const char *); gboolean all = va_arg(args, gboolean); int failcount = va_arg(args, int); time_t last_failure = va_arg(args, int); char *buf = resource_history_string(rsc, rsc_id, all, failcount, last_failure); out->begin_list(out, NULL, NULL, "%s", buf); free(buf); return 0; } static int resource_history_xml(pcmk__output_t *out, va_list args) { resource_t *rsc = va_arg(args, resource_t *); const char *rsc_id = va_arg(args, const char *); gboolean all = va_arg(args, gboolean); int failcount = va_arg(args, int); time_t last_failure = va_arg(args, int); xmlNodePtr node = pcmk__output_xml_create_parent(out, "resource_history"); xmlSetProp(node, (pcmkXmlStr) "id", (pcmkXmlStr) rsc_id); if (rsc == NULL) { xmlSetProp(node, (pcmkXmlStr) "orphan", (pcmkXmlStr) "true"); } else if (all || failcount || last_failure > 0) { char *migration_s = crm_itoa(rsc->migration_threshold); xmlSetProp(node, (pcmkXmlStr) "orphan", (pcmkXmlStr) "false"); xmlSetProp(node, (pcmkXmlStr) "migration-threshold", (pcmkXmlStr) migration_s); free(migration_s); if (failcount > 0) { char *s = crm_itoa(failcount); xmlSetProp(node, (pcmkXmlStr) CRM_FAIL_COUNT_PREFIX, (pcmkXmlStr) s); free(s); } if (last_failure > 0) { xmlSetProp(node, (pcmkXmlStr) CRM_LAST_FAILURE_PREFIX, (pcmkXmlStr) crm_now_string(&last_failure)); } } return 0; } static int stonith_event_console(pcmk__output_t *out, va_list args) { stonith_history_t *event = va_arg(args, stonith_history_t *); int full_history = va_arg(args, int); gboolean later_succeeded = va_arg(args, gboolean); char *buf = NULL; buf = time_t_string(event->completed); switch (event->state) { case st_failed: curses_indented_printf(out, "%s of %s failed: delegate=%s, client=%s, origin=%s, %s='%s %s'\n", stonith_action_str(event->action), event->target, event->delegate ? event->delegate : "", event->client, event->origin, full_history ? "completed" : "last-failed", buf, later_succeeded ? "(a later attempt succeeded)" : ""); break; case st_done: curses_indented_printf(out, "%s of %s successful: delegate=%s, client=%s, origin=%s, %s='%s'\n", stonith_action_str(event->action), event->target, event->delegate ? event->delegate : "", event->client, event->origin, full_history ? "completed" : "last-successful", buf); break; default: curses_indented_printf(out, "%s of %s pending: client=%s, origin=%s\n", stonith_action_str(event->action), event->target, event->client, event->origin); break; } free(buf); return 0; } static int ticket_console(pcmk__output_t *out, va_list args) { ticket_t *ticket = va_arg(args, ticket_t *); if (ticket->last_granted > -1) { char *time = pcmk_format_named_time("last-granted", ticket->last_granted); out->list_item(out, ticket->id, "\t%s%s %s", ticket->granted ? "granted" : "revoked", ticket->standby ? " [standby]" : "", time); free(time); } else { out->list_item(out, ticket->id, "\t%s%s", ticket->granted ? "granted" : "revoked", ticket->standby ? " [standby]" : ""); } return 0; } static pcmk__message_entry_t fmt_functions[] = { { "ban", "console", ban_text }, { "ban", "html", ban_html }, { "ban", "text", ban_text }, { "ban", "xml", ban_xml }, { "bundle", "console", pe__bundle_text }, { "clone", "console", pe__clone_text }, { "cluster-counts", "console", cluster_counts_text }, { "cluster-counts", "html", cluster_counts_html }, { "cluster-counts", "text", cluster_counts_text }, { "cluster-counts", "xml", cluster_counts_xml }, { "cluster-dc", "console", cluster_dc_text }, { "cluster-dc", "html", cluster_dc_html }, { "cluster-dc", "text", cluster_dc_text }, { "cluster-dc", "xml", cluster_dc_xml }, { "cluster-options", "console", cluster_options_text }, { "cluster-options", "html", cluster_options_html }, { "cluster-options", "text", cluster_options_text }, { "cluster-options", "xml", cluster_options_xml }, { "cluster-stack", "console", cluster_stack_text }, { "cluster-stack", "html", cluster_stack_html }, { "cluster-stack", "text", cluster_stack_text }, { "cluster-stack", "xml", cluster_stack_xml }, { "cluster-times", "console", cluster_times_text }, { "cluster-times", "html", cluster_times_html }, { "cluster-times", "text", cluster_times_text }, { "cluster-times", "xml", cluster_times_xml }, { "failed-action", "console", failed_action_console }, { "failed-action", "html", failed_action_html }, { "failed-action", "text", failed_action_text }, { "failed-action", "xml", failed_action_xml }, { "group", "console", pe__group_text }, { "node", "console", node_text }, { "node", "html", node_html }, { "node", "text", node_text }, { "node", "xml", node_xml }, { "node-attribute", "console", node_attribute_text }, { "node-attribute", "html", node_attribute_html }, { "node-attribute", "text", node_attribute_text }, { "node-attribute", "xml", node_attribute_xml }, { "op-history", "console", op_history_text }, { "op-history", "html", op_history_text }, { "op-history", "text", op_history_text }, { "op-history", "xml", op_history_xml }, { "primitive", "console", pe__resource_text }, { "resource-history", "console", resource_history_text }, { "resource-history", "html", resource_history_text }, { "resource-history", "text", resource_history_text }, { "resource-history", "xml", resource_history_xml }, { "stonith-event", "console", stonith_event_console }, { "ticket", "console", ticket_console }, { NULL, NULL, NULL } }; void crm_mon_register_messages(pcmk__output_t *out) { pcmk__register_messages(out, fmt_functions); } diff --git a/tools/crm_mon_print.c b/tools/crm_mon_print.c index ba802740ea..91f6871d44 100644 --- a/tools/crm_mon_print.c +++ b/tools/crm_mon_print.c @@ -1,1263 +1,1263 @@ /* * Copyright 2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "crm_mon.h" static void print_resources_heading(pcmk__output_t *out, unsigned int mon_ops); static void print_resources_closing(pcmk__output_t *out, unsigned int mon_ops); static gboolean print_resources(pcmk__output_t *out, pe_working_set_t *data_set, int print_opts, unsigned int mon_ops, gboolean brief_output, gboolean print_summary); static void print_rsc_history(pcmk__output_t *out, pe_working_set_t *data_set, node_t *node, xmlNode *rsc_entry, unsigned int mon_ops, GListPtr op_list); static void print_node_history(pcmk__output_t *out, pe_working_set_t *data_set, xmlNode *node_state, gboolean operations, unsigned int mon_ops); static gboolean add_extra_info(pcmk__output_t *out, node_t * node, GListPtr rsc_list, const char *attrname, int *expected_score); static void print_node_attribute(gpointer name, gpointer user_data); static gboolean print_node_summary(pcmk__output_t *out, pe_working_set_t * data_set, gboolean operations, unsigned int mon_ops); static gboolean print_cluster_tickets(pcmk__output_t *out, pe_working_set_t * data_set); static gboolean print_neg_locations(pcmk__output_t *out, pe_working_set_t *data_set, unsigned int mon_ops, const char *prefix); static gboolean print_node_attributes(pcmk__output_t *out, pe_working_set_t *data_set, unsigned int mon_ops); static void print_cluster_times(pcmk__output_t *out, pe_working_set_t *data_set); static void print_cluster_dc(pcmk__output_t *out, pe_working_set_t *data_set, unsigned int mon_ops); static void print_cluster_summary(pcmk__output_t *out, pe_working_set_t *data_set, unsigned int mon_ops, unsigned int show); static gboolean print_failed_actions(pcmk__output_t *out, pe_working_set_t *data_set); static gboolean print_failed_stonith_actions(pcmk__output_t *out, stonith_history_t *history, unsigned int mon_ops); static gboolean print_stonith_pending(pcmk__output_t *out, stonith_history_t *history, unsigned int mon_ops); static gboolean print_stonith_history(pcmk__output_t *out, stonith_history_t *history, unsigned int mon_ops); static gboolean print_stonith_history_full(pcmk__output_t *out, stonith_history_t *history, unsigned int mon_ops); /*! * \internal * \brief Print resources section heading appropriate to options * * \param[in] out The output functions structure. * \param[in] mon_ops Bitmask of mon_op_options. */ static void print_resources_heading(pcmk__output_t *out, unsigned int mon_ops) { const char *heading; if (is_set(mon_ops, mon_op_group_by_node)) { /* Active resources have already been printed by node */ heading = is_set(mon_ops, mon_op_inactive_resources) ? "Inactive Resources" : NULL; } else if (is_set(mon_ops, mon_op_inactive_resources)) { heading = "Full List of Resources"; } else { heading = "Active Resources"; } /* Print section heading */ out->begin_list(out, NULL, NULL, "%s", heading); } /*! * \internal * \brief Print whatever resource section closing is appropriate * * \param[in] out The output functions structure. * \param[in] mon_ops Bitmask of mon_op_options. */ static void print_resources_closing(pcmk__output_t *out, unsigned int mon_ops) { const char *heading; /* What type of resources we did or did not display */ if (is_set(mon_ops, mon_op_group_by_node)) { heading = "inactive "; } else if (is_set(mon_ops, mon_op_inactive_resources)) { heading = ""; } else { heading = "active "; } out->list_item(out, NULL, "No %sresources", heading); } /*! * \internal * \brief Print whatever resource section(s) are appropriate * * \param[in] out The output functions structure. * \param[in] data_set Cluster state to display. * \param[in] print_opts Bitmask of pe_print_options. * \param[in] mon_ops Bitmask of mon_op_options. * \param[in] brief_output Whether to display full or brief output. * \param[in] print_summary Whether to display a failure summary. */ static gboolean print_resources(pcmk__output_t *out, pe_working_set_t *data_set, int print_opts, unsigned int mon_ops, gboolean brief_output, gboolean print_summary) { GListPtr rsc_iter; gboolean printed_resource = FALSE; /* If we already showed active resources by node, and * we're not showing inactive resources, we have nothing to do */ if (is_set(mon_ops, mon_op_group_by_node) && is_not_set(mon_ops, mon_op_inactive_resources)) { return FALSE; } print_resources_heading(out, mon_ops); /* If we haven't already printed resources grouped by node, * and brief output was requested, print resource summary */ if (brief_output && is_not_set(mon_ops, mon_op_group_by_node)) { pe__rscs_brief_output(out, data_set->resources, print_opts, is_set(mon_ops, mon_op_inactive_resources)); } /* For each resource, display it if appropriate */ for (rsc_iter = data_set->resources; rsc_iter != NULL; rsc_iter = rsc_iter->next) { resource_t *rsc = (resource_t *) rsc_iter->data; /* Complex resources may have some sub-resources active and some inactive */ gboolean is_active = rsc->fns->active(rsc, TRUE); gboolean partially_active = rsc->fns->active(rsc, FALSE); /* Skip inactive orphans (deleted but still in CIB) */ if (is_set(rsc->flags, pe_rsc_orphan) && !is_active) { continue; /* Skip active resources if we already displayed them by node */ } else if (is_set(mon_ops, mon_op_group_by_node)) { if (is_active) { continue; } /* Skip primitives already counted in a brief summary */ } else if (brief_output && (rsc->variant == pe_native)) { continue; /* Skip resources that aren't at least partially active, * unless we're displaying inactive resources */ } else if (!partially_active && is_not_set(mon_ops, mon_op_inactive_resources)) { continue; } /* Print this resource */ if (printed_resource == FALSE) { printed_resource = TRUE; } out->message(out, crm_map_element_name(rsc->xml), print_opts, rsc); } if (print_summary && !printed_resource) { print_resources_closing(out, mon_ops); } out->end_list(out); return TRUE; } static int failure_count(pe_working_set_t *data_set, node_t *node, resource_t *rsc, time_t *last_failure) { return rsc ? pe_get_failcount(node, rsc, last_failure, pe_fc_default, NULL, data_set) : 0; } static GListPtr get_operation_list(xmlNode *rsc_entry) { GListPtr op_list = NULL; xmlNode *rsc_op = NULL; for (rsc_op = __xml_first_child_element(rsc_entry); rsc_op != NULL; rsc_op = __xml_next_element(rsc_op)) { if (crm_str_eq((const char *) rsc_op->name, XML_LRM_TAG_RSC_OP, TRUE)) { op_list = g_list_append(op_list, rsc_op); } } op_list = g_list_sort(op_list, sort_op_by_callid); return op_list; } /*! * \internal * \brief Print resource operation/failure history * * \param[in] out The output functions structure. * \param[in] data_set Cluster state to display. * \param[in] node Node that ran this resource. * \param[in] rsc_entry Root of XML tree describing resource status. * \param[in] mon_ops Bitmask of mon_op_options. * \param[in] op_list A list of operations to print. */ static void print_rsc_history(pcmk__output_t *out, pe_working_set_t *data_set, node_t *node, xmlNode *rsc_entry, unsigned int mon_ops, GListPtr op_list) { GListPtr gIter = NULL; gboolean printed = FALSE; const char *rsc_id = crm_element_value(rsc_entry, XML_ATTR_ID); resource_t *rsc = pe_find_resource(data_set->resources, rsc_id); /* Print each operation */ for (gIter = op_list; gIter != NULL; gIter = gIter->next) { xmlNode *xml_op = (xmlNode *) gIter->data; const char *task = crm_element_value(xml_op, XML_LRM_ATTR_TASK); const char *interval_ms_s = crm_element_value(xml_op, XML_LRM_ATTR_INTERVAL_MS); const char *op_rc = crm_element_value(xml_op, XML_LRM_ATTR_RC); int rc = crm_parse_int(op_rc, "0"); /* Display 0-interval monitors as "probe" */ if (safe_str_eq(task, CRMD_ACTION_STATUS) && ((interval_ms_s == NULL) || safe_str_eq(interval_ms_s, "0"))) { task = "probe"; } /* Ignore notifies and some probes */ if (safe_str_eq(task, CRMD_ACTION_NOTIFY) || (safe_str_eq(task, "probe") && (rc == 7))) { continue; } /* If this is the first printed operation, print heading for resource */ if (printed == FALSE) { time_t last_failure = 0; int failcount = failure_count(data_set, node, rsc, &last_failure); out->message(out, "resource-history", rsc, rsc_id, TRUE, failcount, last_failure); printed = TRUE; } /* Print the operation */ out->message(out, "op-history", xml_op, task, interval_ms_s, rc, mon_ops); } /* Free the list we created (no need to free the individual items) */ g_list_free(op_list); /* If we printed anything, close the resource */ if (printed) { out->end_list(out); } } /*! * \internal * \brief Print node operation/failure history * * \param[in] out The output functions structure. * \param[in] data_set Cluster state to display. * \param[in] node_state Root of XML tree describing node status. * \param[in] operations Whether to print operations or just failcounts. * \param[in] mon_ops Bitmask of mon_op_options. */ static void print_node_history(pcmk__output_t *out, pe_working_set_t *data_set, xmlNode *node_state, gboolean operations, unsigned int mon_ops) { node_t *node = pe_find_node_id(data_set->nodes, ID(node_state)); xmlNode *lrm_rsc = NULL; xmlNode *rsc_entry = NULL; gboolean printed_header = FALSE; if (node && node->details && node->details->online) { lrm_rsc = find_xml_node(node_state, XML_CIB_TAG_LRM, FALSE); lrm_rsc = find_xml_node(lrm_rsc, XML_LRM_TAG_RESOURCES, FALSE); /* Print history of each of the node's resources */ for (rsc_entry = __xml_first_child_element(lrm_rsc); rsc_entry != NULL; rsc_entry = __xml_next_element(rsc_entry)) { if (crm_str_eq((const char *)rsc_entry->name, XML_LRM_TAG_RESOURCE, TRUE)) { if (operations == FALSE) { const char *rsc_id = crm_element_value(rsc_entry, XML_ATTR_ID); resource_t *rsc = pe_find_resource(data_set->resources, rsc_id); time_t last_failure = 0; int failcount = failure_count(data_set, node, rsc, &last_failure); if (failcount > 0) { if (printed_header == FALSE) { printed_header = TRUE; out->message(out, "node", node, mon_ops, FALSE); } out->message(out, "resource-history", rsc, rsc_id, FALSE, failcount, last_failure); out->end_list(out); } } else { GListPtr op_list = get_operation_list(rsc_entry); if (printed_header == FALSE) { printed_header = TRUE; out->message(out, "node", node, mon_ops, FALSE); } if (g_list_length(op_list) > 0) { print_rsc_history(out, data_set, node, rsc_entry, mon_ops, op_list); } } } } if (printed_header) { out->end_list(out); } } } /*! * \internal * \brief Determine whether extended information about an attribute should be added. * * \param[in] out The output functions structure. * \param[in] node Node that ran this resource. * \param[in] rsc_list The list of resources for this node. * \param[in] attrname The attribute to find. * \param[out] expected_score The expected value for this attribute. * * \return TRUE if extended information should be printed, FALSE otherwise * \note Currently, extended information is only supported for ping/pingd * resources, for which a message will be printed if connectivity is lost * or degraded. */ static gboolean add_extra_info(pcmk__output_t *out, node_t *node, GListPtr rsc_list, const char *attrname, int *expected_score) { GListPtr gIter = NULL; for (gIter = rsc_list; gIter != NULL; gIter = gIter->next) { resource_t *rsc = (resource_t *) gIter->data; const char *type = g_hash_table_lookup(rsc->meta, "type"); const char *name = NULL; if (rsc->children != NULL) { if (add_extra_info(out, node, rsc->children, attrname, expected_score)) { return TRUE; } } if (safe_str_neq(type, "ping") && safe_str_neq(type, "pingd")) { return FALSE; } name = g_hash_table_lookup(rsc->parameters, "name"); if (name == NULL) { name = "pingd"; } /* To identify the resource with the attribute name. */ if (safe_str_eq(name, attrname)) { int host_list_num = 0; /* int value = crm_parse_int(attrvalue, "0"); */ const char *hosts = g_hash_table_lookup(rsc->parameters, "host_list"); const char *multiplier = g_hash_table_lookup(rsc->parameters, "multiplier"); if (hosts) { char **host_list = g_strsplit(hosts, " ", 0); host_list_num = g_strv_length(host_list); g_strfreev(host_list); } /* pingd multiplier is the same as the default value. */ *expected_score = host_list_num * crm_parse_int(multiplier, "1"); return TRUE; } } return FALSE; } /* structure for passing multiple user data to g_list_foreach() */ struct mon_attr_data { pcmk__output_t *out; node_t *node; }; static void print_node_attribute(gpointer name, gpointer user_data) { const char *value = NULL; int expected_score = 0; gboolean add_extra = FALSE; struct mon_attr_data *data = (struct mon_attr_data *) user_data; value = pe_node_attribute_raw(data->node, name); add_extra = add_extra_info(data->out, data->node, data->node->details->running_rsc, name, &expected_score); /* Print attribute name and value */ data->out->message(data->out, "node-attribute", name, value, add_extra, expected_score); } /*! * \internal * \brief Print history for all nodes. * * \param[in] out The output functions structure. * \param[in] data_set Cluster state to display. * \param[in] operations Whether to print operations or just failcounts. * \param[in] mon_ops Bitmask of mon_op_options. */ static gboolean print_node_summary(pcmk__output_t *out, pe_working_set_t * data_set, gboolean operations, unsigned int mon_ops) { xmlNode *node_state = NULL; xmlNode *cib_status = get_object_root(XML_CIB_TAG_STATUS, data_set->input); if (xmlChildElementCount(cib_status) == 0) { return FALSE; } /* Print heading */ if (operations) { out->begin_list(out, NULL, NULL, "Operations"); } else { out->begin_list(out, NULL, NULL, "Migration Summary"); } /* Print each node in the CIB status */ for (node_state = __xml_first_child_element(cib_status); node_state != NULL; node_state = __xml_next_element(node_state)) { if (crm_str_eq((const char *)node_state->name, XML_CIB_TAG_STATE, TRUE)) { print_node_history(out, data_set, node_state, operations, mon_ops); } } /* Close section */ out->end_list(out); return TRUE; } /*! * \internal * \brief Print all tickets. * * \param[in] out The output functions structure. * \param[in] data_set Cluster state to display. */ static gboolean print_cluster_tickets(pcmk__output_t *out, pe_working_set_t * data_set) { GHashTableIter iter; gpointer key, value; if (g_hash_table_size(data_set->tickets) == 0) { return FALSE; } /* Print section heading */ out->begin_list(out, NULL, NULL, "Tickets"); /* Print each ticket */ g_hash_table_iter_init(&iter, data_set->tickets); while (g_hash_table_iter_next(&iter, &key, &value)) { ticket_t *ticket = (ticket_t *) value; out->message(out, "ticket", ticket); } /* Close section */ out->end_list(out); return TRUE; } /*! * \internal * \brief Print section for negative location constraints * * \param[in] out The output functions structure. * \param[in] data_set Cluster state to display. * \param[in] mon_ops Bitmask of mon_op_options. * \param[in] prefix ID prefix to filter results by. */ static gboolean print_neg_locations(pcmk__output_t *out, pe_working_set_t *data_set, unsigned int mon_ops, const char *prefix) { GListPtr gIter, gIter2; gboolean printed_header = FALSE; /* Print each ban */ for (gIter = data_set->placement_constraints; gIter != NULL; gIter = gIter->next) { pe__location_t *location = gIter->data; if (!g_str_has_prefix(location->id, prefix)) continue; for (gIter2 = location->node_list_rh; gIter2 != NULL; gIter2 = gIter2->next) { node_t *node = (node_t *) gIter2->data; if (node->weight < 0) { if (printed_header == FALSE) { printed_header = TRUE; out->begin_list(out, NULL, NULL, "Negative Location Constraints"); } out->message(out, "ban", node, location, mon_ops); } } } if (printed_header) { out->end_list(out); return TRUE; } else { return FALSE; } } /*! * \internal * \brief Print node attributes section * * \param[in] out The output functions structure. * \param[in] data_set Cluster state to display. * \param[in] mon_ops Bitmask of mon_op_options. */ static gboolean print_node_attributes(pcmk__output_t *out, pe_working_set_t *data_set, unsigned int mon_ops) { GListPtr gIter = NULL; gboolean printed_header = FALSE; /* Unpack all resource parameters (it would be more efficient to do this * only when needed for the first time in add_extra_info()) */ for (gIter = data_set->resources; gIter != NULL; gIter = gIter->next) { crm_mon_get_parameters(gIter->data, data_set); } /* Display each node's attributes */ for (gIter = data_set->nodes; gIter != NULL; gIter = gIter->next) { struct mon_attr_data data; data.out = out; data.node = (node_t *) gIter->data; if (data.node && data.node->details && data.node->details->online) { GList *attr_list = NULL; GHashTableIter iter; gpointer key, value; g_hash_table_iter_init(&iter, data.node->details->attrs); while (g_hash_table_iter_next (&iter, &key, &value)) { attr_list = append_attr_list(attr_list, key); } if (g_list_length(attr_list) == 0) { g_list_free(attr_list); continue; } if (printed_header == FALSE) { printed_header = TRUE; out->begin_list(out, NULL, NULL, "Node Attributes"); } out->message(out, "node", data.node, mon_ops, FALSE); g_list_foreach(attr_list, print_node_attribute, &data); g_list_free(attr_list); out->end_list(out); } } /* Print section footer */ if (printed_header) { out->end_list(out); return TRUE; } else { return FALSE; } } /*! * \internal * \brief Print times the display was last updated and CIB last changed * * \param[in] out The output functions structure. * \param[in] data_set Cluster state to display. */ static void print_cluster_times(pcmk__output_t *out, pe_working_set_t *data_set) { const char *last_written = crm_element_value(data_set->input, XML_CIB_ATTR_WRITTEN); const char *user = crm_element_value(data_set->input, XML_ATTR_UPDATE_USER); const char *client = crm_element_value(data_set->input, XML_ATTR_UPDATE_CLIENT); const char *origin = crm_element_value(data_set->input, XML_ATTR_UPDATE_ORIG); out->message(out, "cluster-times", last_written, user, client, origin); } /*! * \internal * \brief Print current DC and its version * * \param[in] out The output functions structure. * \param[in] data_set Cluster state to display. * \param[in] mon_ops Bitmask of mon_op_options. */ static void print_cluster_dc(pcmk__output_t *out, pe_working_set_t *data_set, unsigned int mon_ops) { node_t *dc = data_set->dc_node; xmlNode *dc_version = get_xpath_object("//nvpair[@name='dc-version']", data_set->input, LOG_DEBUG); const char *dc_version_s = dc_version? crm_element_value(dc_version, XML_NVPAIR_ATTR_VALUE) : NULL; const char *quorum = crm_element_value(data_set->input, XML_ATTR_HAVE_QUORUM); char *dc_name = dc? get_node_display_name(dc, mon_ops) : NULL; out->message(out, "cluster-dc", dc, quorum, dc_version_s, dc_name); free(dc_name); } /*! * \internal * \brief Print a summary of cluster-wide information * * \param[in] out The output functions structure. * \param[in] data_set Cluster state to display. * \param[in] mon_ops Bitmask of mon_op_options. * \param[in] show Bitmask of mon_show_options. */ static void print_cluster_summary(pcmk__output_t *out, pe_working_set_t *data_set, unsigned int mon_ops, unsigned int show) { const char *stack_s = get_cluster_stack(data_set); gboolean header_printed = FALSE; if (show & mon_show_stack) { if (header_printed == FALSE) { out->begin_list(out, NULL, NULL, "Cluster Summary"); header_printed = TRUE; } out->message(out, "cluster-stack", stack_s); } /* Always print DC if none, even if not requested */ if ((data_set->dc_node == NULL) || (show & mon_show_dc)) { if (header_printed == FALSE) { out->begin_list(out, NULL, NULL, "Cluster Summary"); header_printed = TRUE; } print_cluster_dc(out, data_set, mon_ops); } if (show & mon_show_times) { if (header_printed == FALSE) { out->begin_list(out, NULL, NULL, "Cluster Summary"); header_printed = TRUE; } print_cluster_times(out, data_set); } if (is_set(data_set->flags, pe_flag_maintenance_mode) || data_set->disabled_resources || data_set->blocked_resources || is_set(show, mon_show_count)) { if (header_printed == FALSE) { out->begin_list(out, NULL, NULL, "Cluster Summary"); header_printed = TRUE; } out->message(out, "cluster-counts", g_list_length(data_set->nodes), - count_resources(data_set, NULL), data_set->disabled_resources, + data_set->ninstances, data_set->disabled_resources, data_set->blocked_resources); } /* There is not a separate option for showing cluster options, so show with * stack for now; a separate option could be added if there is demand */ if (show & mon_show_stack) { out->message(out, "cluster-options", data_set); } if (header_printed) { out->end_list(out); } } /*! * \internal * \brief Print a section for failed actions * * \param[in] out The output functions structure. * \param[in] data_set Cluster state to display. */ static gboolean print_failed_actions(pcmk__output_t *out, pe_working_set_t *data_set) { xmlNode *xml_op = NULL; if (xmlChildElementCount(data_set->failed) == 0) { return FALSE; } /* Print section heading */ out->begin_list(out, NULL, NULL, "Failed Resource Actions"); /* Print each failed action */ for (xml_op = __xml_first_child(data_set->failed); xml_op != NULL; xml_op = __xml_next(xml_op)) { out->message(out, "failed-action", xml_op); out->increment_list(out); } /* End section */ out->end_list(out); return TRUE; } /*! * \internal * \brief Print a section for failed stonith actions * * \note This function should not be called for XML output. * * \param[in] out The output functions structure. * \param[in] history List of stonith actions. * \param[in] mon_ops Bitmask of mon_op_options. */ static gboolean print_failed_stonith_actions(pcmk__output_t *out, stonith_history_t *history, unsigned int mon_ops) { stonith_history_t *hp; for (hp = history; hp; hp = hp->next) { if (hp->state == st_failed) { break; } } if (!hp) { return FALSE; } /* Print section heading */ out->begin_list(out, NULL, NULL, "Failed Fencing Actions"); /* Print each failed stonith action */ for (hp = history; hp; hp = hp->next) { if (hp->state == st_failed) { out->message(out, "stonith-event", hp, mon_ops & mon_op_fence_full_history, stonith__later_succeeded(hp, history)); out->increment_list(out); } } /* End section */ out->end_list(out); return TRUE; } /*! * \internal * \brief Print pending stonith actions * * \note This function should not be called for XML output. * * \param[in] out The output functions structure. * \param[in] history List of stonith actions. * \param[in] mon_ops Bitmask of mon_op_options. */ static gboolean print_stonith_pending(pcmk__output_t *out, stonith_history_t *history, unsigned int mon_ops) { /* xml-output always shows the full history * so we'll never have to show pending-actions * separately */ if (history && (history->state != st_failed) && (history->state != st_done)) { stonith_history_t *hp; /* Print section heading */ out->begin_list(out, NULL, NULL, "Pending Fencing Actions"); for (hp = history; hp; hp = hp->next) { if ((hp->state == st_failed) || (hp->state == st_done)) { break; } out->message(out, "stonith-event", hp, mon_ops & mon_op_fence_full_history, stonith__later_succeeded(hp, history)); out->increment_list(out); } /* End section */ out->end_list(out); return TRUE; } return FALSE; } /*! * \internal * \brief Print fencing history, skipping all failed actions. * * \note This function should not be called for XML output. * * \param[in] out The output functions structure. * \param[in] history List of stonith actions. * \param[in] mon_ops Bitmask of mon_op_options. */ static gboolean print_stonith_history(pcmk__output_t *out, stonith_history_t *history, unsigned int mon_ops) { stonith_history_t *hp; if (history == NULL) { return FALSE; } /* Print section heading */ out->begin_list(out, NULL, NULL, "Fencing History"); for (hp = history; hp; hp = hp->next) { if (hp->state != st_failed) { out->message(out, "stonith-event", hp, mon_ops & mon_op_fence_full_history, stonith__later_succeeded(hp, history)); out->increment_list(out); } } /* End section */ out->end_list(out); return TRUE; } /*! * \internal * \brief Print fencing history, including failed actions. * * \note This function should be called for XML output. It may also be * interesting for other output formats. * * \param[in] out The output functions structure. * \param[in] history List of stonith actions. * \param[in] mon_ops Bitmask of mon_op_options. */ static gboolean print_stonith_history_full(pcmk__output_t *out, stonith_history_t *history, unsigned int mon_ops) { stonith_history_t *hp; if (history == NULL) { return FALSE; } /* Print section heading */ out->begin_list(out, NULL, NULL, "Fencing History"); for (hp = history; hp; hp = hp->next) { out->message(out, "stonith-event", hp, mon_ops & mon_op_fence_full_history, stonith__later_succeeded(hp, history)); out->increment_list(out); } /* End section */ out->end_list(out); return TRUE; } /*! * \internal * \brief Top-level printing function for text/curses output. * * \param[in] out The output functions structure. * \param[in] output_format Is this text or curses output? * \param[in] data_set Cluster state to display. * \param[in] stonith_history List of stonith actions. * \param[in] mon_ops Bitmask of mon_op_options. * \param[in] show Bitmask of mon_show_options. * \param[in] prefix ID prefix to filter results by. */ void print_status(pcmk__output_t *out, mon_output_format_t output_format, pe_working_set_t *data_set, stonith_history_t *stonith_history, unsigned int mon_ops, unsigned int show, const char *prefix) { GListPtr gIter = NULL; int print_opts = get_resource_display_options(mon_ops, output_format); gboolean printed = FALSE; /* space-separated lists of node names */ char *online_nodes = NULL; char *online_remote_nodes = NULL; char *online_guest_nodes = NULL; char *offline_nodes = NULL; char *offline_remote_nodes = NULL; print_cluster_summary(out, data_set, mon_ops, show); if (is_set(show, mon_show_headers)) { out->info(out, "%s", ""); } /* Gather node information (and print if in bad state or grouping by node) */ out->begin_list(out, NULL, NULL, "Node List"); for (gIter = data_set->nodes; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; const char *node_mode = NULL; char *node_name = get_node_display_name(node, mon_ops); /* Get node mode */ if (node->details->unclean) { if (node->details->online) { node_mode = "UNCLEAN (online)"; } else if (node->details->pending) { node_mode = "UNCLEAN (pending)"; } else { node_mode = "UNCLEAN (offline)"; } } else if (node->details->pending) { node_mode = "pending"; } else if (node->details->standby_onfail && node->details->online) { node_mode = "standby (on-fail)"; } else if (node->details->standby) { if (node->details->online) { if (node->details->running_rsc) { node_mode = "standby (with active resources)"; } else { node_mode = "standby"; } } else { node_mode = "OFFLINE (standby)"; } } else if (node->details->maintenance) { if (node->details->online) { node_mode = "maintenance"; } else { node_mode = "OFFLINE (maintenance)"; } } else if (node->details->online) { node_mode = "online"; if (is_not_set(mon_ops, mon_op_group_by_node)) { if (pe__is_guest_node(node)) { online_guest_nodes = add_list_element(online_guest_nodes, node_name); } else if (pe__is_remote_node(node)) { online_remote_nodes = add_list_element(online_remote_nodes, node_name); } else { online_nodes = add_list_element(online_nodes, node_name); } free(node_name); continue; } } else { node_mode = "OFFLINE"; if (is_not_set(mon_ops, mon_op_group_by_node)) { if (pe__is_remote_node(node)) { offline_remote_nodes = add_list_element(offline_remote_nodes, node_name); } else if (pe__is_guest_node(node)) { /* ignore offline guest nodes */ } else { offline_nodes = add_list_element(offline_nodes, node_name); } free(node_name); continue; } } /* If we get here, node is in bad state, or we're grouping by node */ out->message(out, "node", node, mon_ops, TRUE, node_mode); free(node_name); } /* If we're not grouping by node, summarize nodes by status */ if (online_nodes) { out->list_item(out, "Online", "[%s ]", online_nodes); free(online_nodes); } if (offline_nodes) { out->list_item(out, "OFFLINE", "[%s ]", offline_nodes); free(offline_nodes); } if (online_remote_nodes) { out->list_item(out, "RemoteOnline", "[%s ]", online_remote_nodes); free(online_remote_nodes); } if (offline_remote_nodes) { out->list_item(out, "RemoteOFFLINE", "[%s ]", offline_remote_nodes); free(offline_remote_nodes); } if (online_guest_nodes) { out->list_item(out, "GuestOnline", "[%s ]", online_guest_nodes); free(online_guest_nodes); } out->end_list(out); out->info(out, "%s", ""); /* Print resources section, if needed */ printed = print_resources(out, data_set, print_opts, mon_ops, is_set(mon_ops, mon_op_print_brief), TRUE); /* print Node Attributes section if requested */ if (show & mon_show_attributes) { if (printed) { out->info(out, "%s", ""); } printed = print_node_attributes(out, data_set, mon_ops); } /* If requested, print resource operations (which includes failcounts) * or just failcounts */ if (show & (mon_show_operations | mon_show_failcounts)) { if (printed) { out->info(out, "%s", ""); } printed = print_node_summary(out, data_set, ((show & mon_show_operations)? TRUE : FALSE), mon_ops); } /* If there were any failed actions, print them */ if (xml_has_children(data_set->failed)) { if (printed) { out->info(out, "%s", ""); } printed = print_failed_actions(out, data_set); } /* Print failed stonith actions */ if (is_set(mon_ops, mon_op_fence_history)) { if (printed) { out->info(out, "%s", ""); } printed = print_failed_stonith_actions(out, stonith_history, mon_ops); } /* Print tickets if requested */ if (show & mon_show_tickets) { if (printed) { out->info(out, "%s", ""); } printed = print_cluster_tickets(out, data_set); } /* Print negative location constraints if requested */ if (show & mon_show_bans) { if (printed) { out->info(out, "%s", ""); } printed = print_neg_locations(out, data_set, mon_ops, prefix); } /* Print stonith history */ if (is_set(mon_ops, mon_op_fence_history)) { if (printed) { out->info(out, "%s", ""); } if (show & mon_show_fence_history) { print_stonith_history(out, stonith_history, mon_ops); } else { print_stonith_pending(out, stonith_history, mon_ops); } } } /*! * \internal * \brief Top-level printing function for XML output. * * \param[in] out The output functions structure. * \param[in] data_set Cluster state to display. * \param[in] stonith_history List of stonith actions. * \param[in] mon_ops Bitmask of mon_op_options. * \param[in] show Bitmask of mon_show_options. * \param[in] prefix ID prefix to filter results by. */ void print_xml_status(pcmk__output_t *out, pe_working_set_t *data_set, stonith_history_t *stonith_history, unsigned int mon_ops, unsigned int show, const char *prefix) { GListPtr gIter = NULL; int print_opts = get_resource_display_options(mon_ops, mon_output_xml); print_cluster_summary(out, data_set, mon_ops, show); /*** NODES ***/ out->begin_list(out, NULL, NULL, "nodes"); for (gIter = data_set->nodes; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; out->message(out, "node", node, mon_ops, TRUE); } out->end_list(out); /* Print resources section, if needed */ print_resources(out, data_set, print_opts, mon_ops, FALSE, FALSE); /* print Node Attributes section if requested */ if (show & mon_show_attributes) { print_node_attributes(out, data_set, mon_ops); } /* If requested, print resource operations (which includes failcounts) * or just failcounts */ if (show & (mon_show_operations | mon_show_failcounts)) { print_node_summary(out, data_set, ((show & mon_show_operations)? TRUE : FALSE), mon_ops); } /* If there were any failed actions, print them */ if (xml_has_children(data_set->failed)) { print_failed_actions(out, data_set); } /* Print stonith history */ if (is_set(mon_ops, mon_op_fence_history)) { print_stonith_history_full(out, stonith_history, mon_ops); } /* Print tickets if requested */ if (show & mon_show_tickets) { print_cluster_tickets(out, data_set); } /* Print negative location constraints if requested */ if (show & mon_show_bans) { print_neg_locations(out, data_set, mon_ops, prefix); } } /*! * \internal * \brief Top-level printing function for HTML output. * * \param[in] out The output functions structure. * \param[in] output_format Is this HTML or CGI output? * \param[in] data_set Cluster state to display. * \param[in] stonith_history List of stonith actions. * \param[in] mon_ops Bitmask of mon_op_options. * \param[in] show Bitmask of mon_show_options. * \param[in] prefix ID prefix to filter results by. */ int print_html_status(pcmk__output_t *out, mon_output_format_t output_format, pe_working_set_t *data_set, stonith_history_t *stonith_history, unsigned int mon_ops, unsigned int show, const char *prefix) { GListPtr gIter = NULL; int print_opts = get_resource_display_options(mon_ops, output_format); print_cluster_summary(out, data_set, mon_ops, show); /*** NODE LIST ***/ out->begin_list(out, NULL, NULL, "Node List"); for (gIter = data_set->nodes; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; out->message(out, "node", node, mon_ops, TRUE); } out->end_list(out); /* Print resources section, if needed */ print_resources(out, data_set, print_opts, mon_ops, is_set(mon_ops, mon_op_print_brief), TRUE); /* print Node Attributes section if requested */ if (show & mon_show_attributes) { print_node_attributes(out, data_set, mon_ops); } /* If requested, print resource operations (which includes failcounts) * or just failcounts */ if (show & (mon_show_operations | mon_show_failcounts)) { print_node_summary(out, data_set, ((show & mon_show_operations)? TRUE : FALSE), mon_ops); } /* If there were any failed actions, print them */ if (xml_has_children(data_set->failed)) { print_failed_actions(out, data_set); } /* Print failed stonith actions */ if (is_set(mon_ops, mon_op_fence_history)) { print_failed_stonith_actions(out, stonith_history, mon_ops); } /* Print stonith history */ if (is_set(mon_ops, mon_op_fence_history)) { if (show & mon_show_fence_history) { print_stonith_history(out, stonith_history, mon_ops); } else { print_stonith_pending(out, stonith_history, mon_ops); } } /* Print tickets if requested */ if (show & mon_show_tickets) { print_cluster_tickets(out, data_set); } /* Print negative location constraints if requested */ if (show & mon_show_bans) { print_neg_locations(out, data_set, mon_ops, prefix); } return 0; } diff --git a/tools/crm_mon_runtime.c b/tools/crm_mon_runtime.c index 063545d1b7..ec0433dfc3 100644 --- a/tools/crm_mon_runtime.c +++ b/tools/crm_mon_runtime.c @@ -1,226 +1,206 @@ /* * Copyright 2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include #include #include "crm_mon.h" void blank_screen(void) { #if CURSES_ENABLED int lpc = 0; for (lpc = 0; lpc < LINES; lpc++) { move(lpc, 0); clrtoeol(); } move(0, 0); refresh(); #endif } static int compare_attribute(gconstpointer a, gconstpointer b) { int rc; rc = strcmp((const char *)a, (const char *)b); return rc; } GList * append_attr_list(GList *attr_list, char *name) { int i; const char *filt_str[] = FILTER_STR; CRM_CHECK(name != NULL, return attr_list); /* filtering automatic attributes */ for (i = 0; filt_str[i] != NULL; i++) { if (g_str_has_prefix(name, filt_str[i])) { return attr_list; } } return g_list_insert_sorted(attr_list, name, compare_attribute); } -int -count_resources(pe_working_set_t * data_set, resource_t * rsc) -{ - int count = 0; - GListPtr gIter = NULL; - - if (rsc == NULL) { - gIter = data_set->resources; - } else if (rsc->children) { - gIter = rsc->children; - } else { - return is_not_set(rsc->flags, pe_rsc_orphan); - } - - for (; gIter != NULL; gIter = gIter->next) { - count += count_resources(data_set, gIter->data); - } - return count; -} - void crm_mon_get_parameters(resource_t *rsc, pe_working_set_t * data_set) { get_rsc_attributes(rsc->parameters, rsc, NULL, data_set); if(rsc->children) { GListPtr gIter = NULL; for (gIter = rsc->children; gIter != NULL; gIter = gIter->next) { crm_mon_get_parameters(gIter->data, data_set); } } } /*! * \internal * \brief Get the name of the stack in use (or "unknown" if not available) * * \param[in] data_set Working set of CIB state * * \return String representing stack name */ const char * get_cluster_stack(pe_working_set_t *data_set) { xmlNode *stack = get_xpath_object("//nvpair[@name='cluster-infrastructure']", data_set->input, LOG_DEBUG); return stack? crm_element_value(stack, XML_NVPAIR_ATTR_VALUE) : "unknown"; } /*! * \internal * \brief Return human-friendly string representing node name * * The returned string will be in the format * uname[@hostUname] [(nodeID)] * "@hostUname" will be printed if the node is a guest node. * "(nodeID)" will be printed if the node ID is different from the node uname, * and detailed output has been requested. * * \param[in] node Node to represent * \return Newly allocated string with representation of node name * \note It is the caller's responsibility to free the result with free(). */ char * get_node_display_name(node_t *node, unsigned int mon_ops) { char *node_name; const char *node_host = NULL; const char *node_id = NULL; int name_len; CRM_ASSERT((node != NULL) && (node->details != NULL) && (node->details->uname != NULL)); /* Host is displayed only if this is a guest node */ if (pe__is_guest_node(node)) { node_t *host_node = pe__current_node(node->details->remote_rsc); if (host_node && host_node->details) { node_host = host_node->details->uname; } if (node_host == NULL) { node_host = ""; /* so we at least get "uname@" to indicate guest */ } } /* Node ID is displayed if different from uname and detail is requested */ if (is_set(mon_ops, mon_op_print_clone_detail) && safe_str_neq(node->details->uname, node->details->id)) { node_id = node->details->id; } /* Determine name length */ name_len = strlen(node->details->uname) + 1; if (node_host) { name_len += strlen(node_host) + 1; /* "@node_host" */ } if (node_id) { name_len += strlen(node_id) + 3; /* + " (node_id)" */ } /* Allocate and populate display name */ node_name = malloc(name_len); CRM_ASSERT(node_name != NULL); strcpy(node_name, node->details->uname); if (node_host) { strcat(node_name, "@"); strcat(node_name, node_host); } if (node_id) { strcat(node_name, " ("); strcat(node_name, node_id); strcat(node_name, ")"); } return node_name; } /*! * \internal * \brief Return resource display options corresponding to command-line choices * * \return Bitmask of pe_print_options suitable for resource print functions */ int get_resource_display_options(unsigned int mon_ops, mon_output_format_t output_format) { int print_opts; /* Determine basic output format */ switch (output_format) { case mon_output_console: print_opts = pe_print_ncurses; break; case mon_output_html: case mon_output_cgi: print_opts = pe_print_html; break; case mon_output_xml: print_opts = pe_print_xml; break; default: print_opts = pe_print_printf; break; } /* Add optional display elements */ if (is_set(mon_ops, mon_op_print_pending)) { print_opts |= pe_print_pending; } if (is_set(mon_ops, mon_op_print_clone_detail)) { print_opts |= pe_print_clone_details|pe_print_implicit; } if (is_not_set(mon_ops, mon_op_inactive_resources)) { print_opts |= pe_print_clone_active; } if (is_set(mon_ops, mon_op_print_brief)) { print_opts |= pe_print_brief; } return print_opts; } diff --git a/tools/crm_resource.c b/tools/crm_resource.c index 89e8a79305..e1096cc118 100644 --- a/tools/crm_resource.c +++ b/tools/crm_resource.c @@ -1,1417 +1,1418 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include #include #include bool BE_QUIET = FALSE; bool scope_master = FALSE; int cib_options = cib_sync_call; static GMainLoop *mainloop = NULL; #define MESSAGE_TIMEOUT_S 60 static gboolean resource_ipc_timeout(gpointer data) { fprintf(stderr, "Aborting because no messages received in %d seconds\n", MESSAGE_TIMEOUT_S); crm_err("No messages received in %d seconds", MESSAGE_TIMEOUT_S); crm_exit(CRM_EX_TIMEOUT); } static void resource_ipc_connection_destroy(gpointer user_data) { crm_info("Connection to controller was terminated"); crm_exit(CRM_EX_DISCONNECT); } static void start_mainloop(void) { if (crmd_replies_needed == 0) { return; } mainloop = g_main_loop_new(NULL, FALSE); fprintf(stderr, "Waiting for %d repl%s from the controller", crmd_replies_needed, (crmd_replies_needed == 1)? "y" : "ies"); crm_debug("Waiting for %d repl%s from the controller", crmd_replies_needed, (crmd_replies_needed == 1)? "y" : "ies"); g_timeout_add(MESSAGE_TIMEOUT_S * 1000, resource_ipc_timeout, NULL); g_main_loop_run(mainloop); } static int resource_ipc_callback(const char *buffer, ssize_t length, gpointer userdata) { xmlNode *msg = string2xml(buffer); fprintf(stderr, "."); crm_log_xml_trace(msg, "[inbound]"); crmd_replies_needed--; if ((crmd_replies_needed == 0) && mainloop && g_main_loop_is_running(mainloop)) { fprintf(stderr, " OK\n"); crm_debug("Got all the replies we expected"); crm_exit(CRM_EX_OK); } free_xml(msg); return 0; } static int compare_id(gconstpointer a, gconstpointer b) { return strcmp((const char *)a, (const char *)b); } static GListPtr build_constraint_list(xmlNode *root) { GListPtr retval = NULL; xmlNode *cib_constraints = NULL; xmlXPathObjectPtr xpathObj = NULL; int ndx = 0; cib_constraints = get_object_root(XML_CIB_TAG_CONSTRAINTS, root); xpathObj = xpath_search(cib_constraints, "//" XML_CONS_TAG_RSC_LOCATION); for (ndx = 0; ndx < numXpathResults(xpathObj); ndx++) { xmlNode *match = getXpathResult(xpathObj, ndx); retval = g_list_insert_sorted(retval, (gpointer) ID(match), compare_id); } freeXpathObject(xpathObj); return retval; } struct ipc_client_callbacks crm_callbacks = { .dispatch = resource_ipc_callback, .destroy = resource_ipc_connection_destroy, }; /* short option letters still available: eEJkKXyYZ */ /* *INDENT-OFF* */ static struct crm_option long_options[] = { /* Top-level Options */ { "help", no_argument, NULL, '?', "\t\tDisplay this text and exit" }, { "version", no_argument, NULL, '$', "\t\tDisplay version information and exit" }, { "verbose", no_argument, NULL, 'V', "\t\tIncrease debug output (may be specified multiple times)" }, { "quiet", no_argument, NULL, 'Q', "\t\tBe less descriptive in results" }, { "resource", required_argument, NULL, 'r', "\tResource ID" }, { "-spacer-", no_argument, NULL, '-', "\nQueries:" }, { "list", no_argument, NULL, 'L', "\t\tList all cluster resources with status"}, { "list-raw", no_argument, NULL, 'l', "\t\tList IDs of all instantiated resources (individual members rather than groups etc.)" }, { "list-cts", no_argument, NULL, 'c', NULL, pcmk_option_hidden }, { "list-operations", no_argument, NULL, 'O', "\tList active resource operations, optionally filtered by --resource and/or --node" }, { "list-all-operations", no_argument, NULL, 'o', "List all resource operations, optionally filtered by --resource and/or --node" }, { "list-standards", no_argument, NULL, 0, "\tList supported standards" }, { "list-ocf-providers", no_argument, NULL, 0, "List all available OCF providers" }, { "list-agents", required_argument, NULL, 0, "List all agents available for the named standard and/or provider." }, { "list-ocf-alternatives", required_argument, NULL, 0, "List all available providers for the named OCF agent" }, { "show-metadata", required_argument, NULL, 0, "Show the metadata for the named class:provider:agent" }, { "query-xml", no_argument, NULL, 'q', "\tShow XML configuration of resource (after any template expansion)" }, { "query-xml-raw", no_argument, NULL, 'w', "\tShow XML configuration of resource (before any template expansion)" }, { "get-parameter", required_argument, NULL, 'g', "Display named parameter for resource.\n" "\t\t\t\tUse instance attribute unless --meta or --utilization is specified" }, { "get-property", required_argument, NULL, 'G', "Display named property of resource ('class', 'type', or 'provider') (requires --resource)", pcmk_option_hidden }, { "locate", no_argument, NULL, 'W', "\t\tShow node(s) currently running resource" }, { "stack", no_argument, NULL, 'A', "\t\tDisplay the prerequisites and dependents of a resource" }, { "constraints", no_argument, NULL, 'a', "\tDisplay the (co)location constraints that apply to a resource" }, { "why", no_argument, NULL, 'Y', "\t\tShow why resources are not running, optionally filtered by --resource and/or --node" }, { "-spacer-", no_argument, NULL, '-', "\nCommands:" }, { "validate", no_argument, NULL, 0, "\t\tValidate resource configuration by calling agent's validate-all action.\n" "\t\t\t\tThe configuration may be specified either by giving an existing\n" "\t\t\t\tresource name with -r, or by specifying --class, --agent, and\n" "\t\t\t\t--provider arguments, along with any number of --option arguments." }, { "cleanup", no_argument, NULL, 'C', "\t\tIf resource has any past failures, clear its history and fail count.\n" "\t\t\t\tOptionally filtered by --resource, --node, --operation, and --interval (otherwise all).\n" "\t\t\t\t--operation and --interval apply to fail counts, but entire history is always cleared,\n" "\t\t\t\tto allow current state to be rechecked. If the named resource is part of a group, or\n" "\t\t\t\tone numbered instance of a clone or bundled resource, the clean-up applies to the\n" "\t\t\t\twhole collective resource unless --force is given." }, { "refresh", no_argument, NULL, 'R', "\t\tDelete resource's history (including failures) so its current state is rechecked.\n" "\t\t\t\tOptionally filtered by --resource and --node (otherwise all). If the named resource is\n" "\t\t\t\tpart of a group, or one numbered instance of a clone or bundled resource, the clean-up\n" "applies to the whole collective resource unless --force is given." }, { "set-parameter", required_argument, NULL, 'p', "Set named parameter for resource (requires -v).\n" "\t\t\t\tUse instance attribute unless --meta or --utilization is specified." }, { "delete-parameter", required_argument, NULL, 'd', "Delete named parameter for resource.\n" "\t\t\t\tUse instance attribute unless --meta or --utilization is specified." }, { "set-property", required_argument, NULL, 'S', "Set named property of resource ('class', 'type', or 'provider') (requires -r, -t, -v)", pcmk_option_hidden }, { "-spacer-", no_argument, NULL, '-', "\nResource location:" }, { "move", no_argument, NULL, 'M', "\t\tCreate a constraint to move resource. If --node is specified, the constraint\n" "\t\t\t\twill be to move to that node, otherwise it will be to ban the current node.\n" "\t\t\t\tUnless --force is specified, this will return an error if the resource is\n" "\t\t\t\talready running on the specified node. If --force is specified, this will\n" "\t\t\t\talways ban the current node. Optional: --lifetime, --master.\n" "\t\t\t\tNOTE: This may prevent the resource from running on its previous location\n" "\t\t\t\tuntil the implicit constraint expires or is removed with --clear." }, { "ban", no_argument, NULL, 'B', "\t\tCreate a constraint to keep resource off a node. Optional: --node, --lifetime, --master.\n" "\t\t\t\tNOTE: This will prevent the resource from running on the affected node\n" "\t\t\t\tuntil the implicit constraint expires or is removed with --clear.\n" "\t\t\t\tIf --node is not specified, it defaults to the node currently running the resource\n" "\t\t\t\tfor primitives and groups, or the master for promotable clones with promoted-max=1\n" "\t\t\t\t(all other situations result in an error as there is no sane default).\n" }, { "clear", no_argument, NULL, 'U', "\t\tRemove all constraints created by the --ban and/or --move commands.\n" "\t\t\t\tRequires: --resource. Optional: --node, --master, --expired.\n" "\t\t\t\tIf --node is not specified, all constraints created by --ban and --move\n" "\t\t\t\twill be removed for the named resource. If --node and --force are specified,\n" "\t\t\t\tany constraint created by --move will be cleared, even if it is not for the specified node.\n" "\t\t\t\tIf --expired is specified, only those constraints whose lifetimes have expired will\n" "\t\t\t\tbe removed.\n" }, { "expired", no_argument, NULL, 'e', "\t\tModifies the --clear argument to remove constraints with expired lifetimes.\n" }, { "lifetime", required_argument, NULL, 'u', "\tLifespan (as ISO 8601 duration) of created constraints (with -B, -M)\n" "\t\t\t\t(see https://en.wikipedia.org/wiki/ISO_8601#Durations)" }, { "master", no_argument, NULL, 0, "\t\tLimit scope of command to the Master role (with -B, -M, -U).\n" "\t\t\t\tFor -B and -M, the previous master may remain active in the Slave role." }, { "-spacer-", no_argument, NULL, '-', "\nAdvanced Commands:" }, { "delete", no_argument, NULL, 'D', "\t\t(Advanced) Delete a resource from the CIB. Required: -t" }, { "fail", no_argument, NULL, 'F', "\t\t(Advanced) Tell the cluster this resource has failed" }, { "restart", no_argument, NULL, 0, "\t\t(Advanced) Tell the cluster to restart this resource and anything that depends on it" }, { "wait", no_argument, NULL, 0, "\t\t(Advanced) Wait until the cluster settles into a stable state" }, { "force-demote", no_argument, NULL, 0, "\t(Advanced) Bypass the cluster and demote a resource on the local node.\n" "\t\t\t\tUnless --force is specified, this will refuse to do so if the cluster\n" "\t\t\t\tbelieves the resource is a clone instance already running on the local node." }, { "force-stop", no_argument, NULL, 0, "\t(Advanced) Bypass the cluster and stop a resource on the local node." }, { "force-start", no_argument, NULL, 0, "\t(Advanced) Bypass the cluster and start a resource on the local node.\n" "\t\t\t\tUnless --force is specified, this will refuse to do so if the cluster\n" "\t\t\t\tbelieves the resource is a clone instance already running on the local node." }, { "force-promote", no_argument, NULL, 0, "\t(Advanced) Bypass the cluster and promote a resource on the local node.\n" "\t\t\t\tUnless --force is specified, this will refuse to do so if the cluster\n" "\t\t\t\tbelieves the resource is a clone instance already running on the local node." }, { "force-check", no_argument, NULL, 0, "\t(Advanced) Bypass the cluster and check the state of a resource on the local node." }, { "-spacer-", no_argument, NULL, '-', "\nValidate Options:" }, { "class", required_argument, NULL, 0, "\tThe standard the resource agent confirms to (for example, ocf).\n" "\t\t\t\tUse with --agent, --provider, --option, and --validate." }, { "agent", required_argument, NULL, 0, "\tThe agent to use (for example, IPaddr).\n" "\t\t\t\tUse with --class, --provider, --option, and --validate." }, { "provider", required_argument, NULL, 0, "\tThe vendor that supplies the resource agent (for example, heartbeat).\n" "\t\t\t\tuse with --class, --agent, --option, and --validate." }, { "option", required_argument, NULL, 0, "\tSpecify a device configuration parameter as NAME=VALUE\n" "\t\t\t\t(may be specified multiple times). Use with --validate\n" "\t\t\t\tand without the -r option." }, { "-spacer-", no_argument, NULL, '-', "\nAdditional Options:" }, { "node", required_argument, NULL, 'N', "\tNode name" }, { "recursive", no_argument, NULL, 0, "\tFollow colocation chains when using --set-parameter" }, { "resource-type", required_argument, NULL, 't', "Resource XML element (primitive, group, etc.) (with -D)" }, { "parameter-value", required_argument, NULL, 'v', "Value to use with -p" }, { "meta", no_argument, NULL, 'm', "\t\tUse resource meta-attribute instead of instance attribute (with -p, -g, -d)" }, { "utilization", no_argument, NULL, 'z', "\tUse resource utilization attribute instead of instance attribute (with -p, -g, -d)" }, { "operation", required_argument, NULL, 'n', "\tOperation to clear instead of all (with -C -r)" }, { "interval", required_argument, NULL, 'I', "\tInterval of operation to clear (default 0) (with -C -r -n)" }, { "set-name", required_argument, NULL, 's', "\t(Advanced) XML ID of attributes element to use (with -p, -d)" }, { "nvpair", required_argument, NULL, 'i', "\t(Advanced) XML ID of nvpair element to use (with -p, -d)" }, { "timeout", required_argument, NULL, 'T', "\t(Advanced) Abort if command does not finish in this time (with --restart, --wait, --force-*)" }, { "force", no_argument, NULL, 'f', "\t\tIf making CIB changes, do so regardless of quorum.\n" "\t\t\t\tSee help for individual commands for additional behavior.\n" }, { "xml-file", required_argument, NULL, 'x', NULL, pcmk_option_hidden }, /* legacy options */ {"host-uname", required_argument, NULL, 'H', NULL, pcmk_option_hidden}, {"-spacer-", 1, NULL, '-', "\nExamples:", pcmk_option_paragraph}, {"-spacer-", 1, NULL, '-', "List the available OCF agents:", pcmk_option_paragraph}, {"-spacer-", 1, NULL, '-', " crm_resource --list-agents ocf", pcmk_option_example}, {"-spacer-", 1, NULL, '-', "List the available OCF agents from the linux-ha project:", pcmk_option_paragraph}, {"-spacer-", 1, NULL, '-', " crm_resource --list-agents ocf:heartbeat", pcmk_option_example}, {"-spacer-", 1, NULL, '-', "Move 'myResource' to a specific node:", pcmk_option_paragraph}, {"-spacer-", 1, NULL, '-', " crm_resource --resource myResource --move --node altNode", pcmk_option_example}, {"-spacer-", 1, NULL, '-', "Allow (but not force) 'myResource' to move back to its original location:", pcmk_option_paragraph}, {"-spacer-", 1, NULL, '-', " crm_resource --resource myResource --clear", pcmk_option_example}, {"-spacer-", 1, NULL, '-', "Stop 'myResource' (and anything that depends on it):", pcmk_option_paragraph}, {"-spacer-", 1, NULL, '-', " crm_resource --resource myResource --set-parameter target-role --meta --parameter-value Stopped", pcmk_option_example}, {"-spacer-", 1, NULL, '-', "Tell the cluster not to manage 'myResource':", pcmk_option_paragraph}, {"-spacer-", 1, NULL, '-', "The cluster will not attempt to start or stop the resource under any circumstances."}, {"-spacer-", 1, NULL, '-', "Useful when performing maintenance tasks on a resource.", pcmk_option_paragraph}, {"-spacer-", 1, NULL, '-', " crm_resource --resource myResource --set-parameter is-managed --meta --parameter-value false", pcmk_option_example}, {"-spacer-", 1, NULL, '-', "Erase the operation history of 'myResource' on 'aNode':", pcmk_option_paragraph}, {"-spacer-", 1, NULL, '-', "The cluster will 'forget' the existing resource state (including any errors) and attempt to recover the resource."}, {"-spacer-", 1, NULL, '-', "Useful when a resource had failed permanently and has been repaired by an administrator.", pcmk_option_paragraph}, {"-spacer-", 1, NULL, '-', " crm_resource --resource myResource --cleanup --node aNode", pcmk_option_example}, {0, 0, 0, 0} }; /* *INDENT-ON* */ int main(int argc, char **argv) { char rsc_cmd = 'L'; const char *v_class = NULL; const char *v_agent = NULL; const char *v_provider = NULL; char *name = NULL; char *value = NULL; GHashTable *validate_options = NULL; const char *rsc_id = NULL; const char *host_uname = NULL; const char *prop_name = NULL; const char *prop_value = NULL; const char *rsc_type = NULL; const char *prop_id = NULL; const char *prop_set = NULL; const char *rsc_long_cmd = NULL; const char *longname = NULL; const char *operation = NULL; const char *interval_spec = NULL; const char *cib_file = getenv("CIB_file"); GHashTable *override_params = NULL; char *xml_file = NULL; crm_ipc_t *crmd_channel = NULL; pe_working_set_t *data_set = NULL; xmlNode *cib_xml_copy = NULL; cib_t *cib_conn = NULL; resource_t *rsc = NULL; bool recursive = FALSE; char *our_pid = NULL; bool validate_cmdline = FALSE; /* whether we are just validating based on command line options */ bool require_resource = TRUE; /* whether command requires that resource be specified */ bool require_dataset = TRUE; /* whether command requires populated dataset instance */ bool require_crmd = FALSE; // whether command requires controller connection bool clear_expired = FALSE; int rc = pcmk_ok; int is_ocf_rc = 0; int option_index = 0; int timeout_ms = 0; int argerr = 0; int flag; int find_flags = 0; // Flags to use when searching for resource crm_exit_t exit_code = CRM_EX_OK; crm_log_cli_init("crm_resource"); crm_set_options(NULL, "(query|command) [options]", long_options, "Perform tasks related to cluster resources.\nAllows resources to be queried (definition and location), modified, and moved around the cluster.\n"); validate_options = crm_str_table_new(); while (1) { flag = crm_get_option_long(argc, argv, &option_index, &longname); if (flag == -1) break; switch (flag) { case 0: /* long options with no short equivalent */ if (safe_str_eq("master", longname)) { scope_master = TRUE; } else if(safe_str_eq(longname, "recursive")) { recursive = TRUE; } else if (safe_str_eq("wait", longname)) { rsc_cmd = flag; rsc_long_cmd = longname; require_resource = FALSE; require_dataset = FALSE; } else if ( safe_str_eq("validate", longname) || safe_str_eq("restart", longname) || safe_str_eq("force-demote", longname) || safe_str_eq("force-stop", longname) || safe_str_eq("force-start", longname) || safe_str_eq("force-promote", longname) || safe_str_eq("force-check", longname)) { rsc_cmd = flag; rsc_long_cmd = longname; find_flags = pe_find_renamed|pe_find_anon; crm_log_args(argc, argv); } else if (safe_str_eq("list-ocf-providers", longname) || safe_str_eq("list-ocf-alternatives", longname) || safe_str_eq("list-standards", longname)) { const char *text = NULL; lrmd_list_t *list = NULL; lrmd_list_t *iter = NULL; lrmd_t *lrmd_conn = lrmd_api_new(); if (safe_str_eq("list-ocf-providers", longname) || safe_str_eq("list-ocf-alternatives", longname)) { rc = lrmd_conn->cmds->list_ocf_providers(lrmd_conn, optarg, &list); text = "OCF providers"; } else if (safe_str_eq("list-standards", longname)) { rc = lrmd_conn->cmds->list_standards(lrmd_conn, &list); text = "standards"; } if (rc > 0) { for (iter = list; iter != NULL; iter = iter->next) { printf("%s\n", iter->val); } lrmd_list_freeall(list); } else if (optarg) { fprintf(stderr, "No %s found for %s\n", text, optarg); exit_code = CRM_EX_NOSUCH; } else { fprintf(stderr, "No %s found\n", text); exit_code = CRM_EX_NOSUCH; } lrmd_api_delete(lrmd_conn); crm_exit(exit_code); } else if (safe_str_eq("show-metadata", longname)) { char *standard = NULL; char *provider = NULL; char *type = NULL; char *metadata = NULL; lrmd_t *lrmd_conn = lrmd_api_new(); rc = crm_parse_agent_spec(optarg, &standard, &provider, &type); if (rc == pcmk_ok) { rc = lrmd_conn->cmds->get_metadata(lrmd_conn, standard, provider, type, &metadata, 0); } else { fprintf(stderr, "'%s' is not a valid agent specification\n", optarg); rc = -ENXIO; } if (metadata) { printf("%s\n", metadata); } else { fprintf(stderr, "Metadata query for %s failed: %s\n", optarg, pcmk_strerror(rc)); exit_code = crm_errno2exit(rc); } lrmd_api_delete(lrmd_conn); crm_exit(exit_code); } else if (safe_str_eq("list-agents", longname)) { lrmd_list_t *list = NULL; lrmd_list_t *iter = NULL; char *provider = strchr (optarg, ':'); lrmd_t *lrmd_conn = lrmd_api_new(); if (provider) { *provider++ = 0; } rc = lrmd_conn->cmds->list_agents(lrmd_conn, &list, optarg, provider); if (rc > 0) { for (iter = list; iter != NULL; iter = iter->next) { printf("%s\n", iter->val); } lrmd_list_freeall(list); } else { fprintf(stderr, "No agents found for standard=%s, provider=%s\n", optarg, (provider? provider : "*")); exit_code = CRM_EX_NOSUCH; } lrmd_api_delete(lrmd_conn); crm_exit(exit_code); } else if (safe_str_eq("class", longname)) { if (!(pcmk_get_ra_caps(optarg) & pcmk_ra_cap_params)) { if (BE_QUIET == FALSE) { fprintf(stdout, "Standard %s does not support parameters\n", optarg); } crm_exit(exit_code); } else { v_class = optarg; } validate_cmdline = TRUE; require_resource = FALSE; } else if (safe_str_eq("agent", longname)) { validate_cmdline = TRUE; require_resource = FALSE; v_agent = optarg; } else if (safe_str_eq("provider", longname)) { validate_cmdline = TRUE; require_resource = FALSE; v_provider = optarg; } else if (safe_str_eq("option", longname)) { crm_info("Scanning: --option %s", optarg); rc = pcmk_scan_nvpair(optarg, &name, &value); if (rc != 2) { fprintf(stderr, "Invalid option: --option %s: %s", optarg, pcmk_strerror(rc)); argerr++; } else { crm_info("Got: '%s'='%s'", name, value); } g_hash_table_replace(validate_options, name, value); } else { crm_err("Unhandled long option: %s", longname); } break; case 'V': resource_verbose++; crm_bump_log_level(argc, argv); break; case '$': case '?': crm_help(flag, CRM_EX_OK); break; case 'x': xml_file = strdup(optarg); break; case 'Q': BE_QUIET = TRUE; break; case 'm': attr_set_type = XML_TAG_META_SETS; break; case 'z': attr_set_type = XML_TAG_UTILIZATION; break; case 'u': move_lifetime = strdup(optarg); break; case 'f': do_force = TRUE; crm_log_args(argc, argv); break; case 'i': prop_id = optarg; break; case 's': prop_set = optarg; break; case 'r': rsc_id = optarg; break; case 'v': prop_value = optarg; break; case 't': rsc_type = optarg; break; case 'T': timeout_ms = crm_get_msec(optarg); break; case 'e': clear_expired = TRUE; require_resource = FALSE; break; case 'C': case 'R': crm_log_args(argc, argv); require_resource = FALSE; if (cib_file == NULL) { require_crmd = TRUE; } rsc_cmd = flag; find_flags = pe_find_renamed|pe_find_anon; break; case 'n': operation = optarg; break; case 'I': interval_spec = optarg; break; case 'D': require_dataset = FALSE; crm_log_args(argc, argv); rsc_cmd = flag; find_flags = pe_find_renamed|pe_find_any; break; case 'F': require_crmd = TRUE; crm_log_args(argc, argv); rsc_cmd = flag; break; case 'U': case 'B': case 'M': crm_log_args(argc, argv); rsc_cmd = flag; find_flags = pe_find_renamed|pe_find_anon; break; case 'c': case 'L': case 'l': case 'O': case 'o': require_resource = FALSE; rsc_cmd = flag; break; case 'Y': require_resource = FALSE; rsc_cmd = flag; find_flags = pe_find_renamed|pe_find_anon; break; case 'q': case 'w': rsc_cmd = flag; find_flags = pe_find_renamed|pe_find_any; break; case 'W': case 'A': case 'a': rsc_cmd = flag; find_flags = pe_find_renamed|pe_find_anon; break; case 'S': require_dataset = FALSE; crm_log_args(argc, argv); prop_name = optarg; rsc_cmd = flag; find_flags = pe_find_renamed|pe_find_any; break; case 'p': case 'd': crm_log_args(argc, argv); prop_name = optarg; rsc_cmd = flag; find_flags = pe_find_renamed|pe_find_any; break; case 'G': case 'g': prop_name = optarg; rsc_cmd = flag; find_flags = pe_find_renamed|pe_find_any; break; case 'H': case 'N': crm_trace("Option %c => %s", flag, optarg); host_uname = optarg; break; default: CMD_ERR("Argument code 0%o (%c) is not (?yet?) supported", flag, flag); ++argerr; break; } } // Catch the case where the user didn't specify a command if (rsc_cmd == 'L') { require_resource = FALSE; } // --expired without --clear/-U doesn't make sense if (clear_expired == TRUE && rsc_cmd != 'U') { CMD_ERR("--expired requires --clear or -U"); argerr++; } if (optind < argc && argv[optind] != NULL && rsc_cmd == 0 && rsc_long_cmd) { override_params = crm_str_table_new(); while (optind < argc && argv[optind] != NULL) { char *name = calloc(1, strlen(argv[optind])); char *value = calloc(1, strlen(argv[optind])); int rc = sscanf(argv[optind], "%[^=]=%s", name, value); if(rc == 2) { g_hash_table_replace(override_params, name, value); } else { CMD_ERR("Error parsing '%s' as a name=value pair for --%s", argv[optind], rsc_long_cmd); free(value); free(name); argerr++; } optind++; } } else if (optind < argc && argv[optind] != NULL && rsc_cmd == 0) { CMD_ERR("non-option ARGV-elements: "); while (optind < argc && argv[optind] != NULL) { CMD_ERR("[%d of %d] %s ", optind, argc, argv[optind]); optind++; argerr++; } } if (optind > argc) { ++argerr; } // Sanity check validating from command line parameters. If everything checks out, // go ahead and run the validation. This way we don't need a CIB connection. if (validate_cmdline == TRUE) { // -r cannot be used with any of --class, --agent, or --provider if (rsc_id != NULL) { CMD_ERR("--resource cannot be used with --class, --agent, and --provider"); argerr++; // If --class, --agent, or --provider are given, --validate must also be given. } else if (!safe_str_eq(rsc_long_cmd, "validate")) { CMD_ERR("--class, --agent, and --provider require --validate"); argerr++; // Not all of --class, --agent, and --provider need to be given. Not all // classes support the concept of a provider. Check that what we were given // is valid. } else if (crm_str_eq(v_class, "stonith", TRUE)) { if (v_provider != NULL) { CMD_ERR("stonith does not support providers"); argerr++; } else if (stonith_agent_exists(v_agent, 0) == FALSE) { CMD_ERR("%s is not a known stonith agent", v_agent ? v_agent : ""); argerr++; } } else if (resources_agent_exists(v_class, v_provider, v_agent) == FALSE) { CMD_ERR("%s:%s:%s is not a known resource", v_class ? v_class : "", v_provider ? v_provider : "", v_agent ? v_agent : ""); argerr++; } if (argerr == 0) { rc = cli_resource_execute_from_params("test", v_class, v_provider, v_agent, "validate-all", validate_options, override_params, timeout_ms); exit_code = crm_errno2exit(rc); crm_exit(exit_code); } } if (argerr) { CMD_ERR("Invalid option(s) supplied, use --help for valid usage"); crm_exit(CRM_EX_USAGE); } our_pid = crm_getpid_s(); if (do_force) { crm_debug("Forcing..."); cib_options |= cib_quorum_override; } if (require_resource && !rsc_id) { CMD_ERR("Must supply a resource id with -r"); rc = -ENXIO; goto bail; } if (find_flags && rsc_id) { require_dataset = TRUE; } /* Establish a connection to the CIB manager */ cib_conn = cib_new(); rc = cib_conn->cmds->signon(cib_conn, crm_system_name, cib_command); if (rc != pcmk_ok) { CMD_ERR("Error connecting to the CIB manager: %s", pcmk_strerror(rc)); goto bail; } /* Populate working set from XML file if specified or CIB query otherwise */ if (require_dataset) { if (xml_file != NULL) { cib_xml_copy = filename2xml(xml_file); } else { rc = cib_conn->cmds->query(cib_conn, NULL, &cib_xml_copy, cib_scope_local | cib_sync_call); } if(rc != pcmk_ok) { goto bail; } /* Populate the working set instance */ data_set = pe_new_working_set(); if (data_set == NULL) { rc = -ENOMEM; goto bail; } + set_bit(data_set->flags, pe_flag_no_counts); rc = update_working_set_xml(data_set, &cib_xml_copy); if (rc != pcmk_ok) { goto bail; } cluster_status(data_set); } // If command requires that resource exist if specified, find it if (find_flags && rsc_id) { rsc = pe_find_resource_with_flags(data_set->resources, rsc_id, find_flags); if (rsc == NULL) { CMD_ERR("Resource '%s' not found", rsc_id); rc = -ENXIO; goto bail; } } // Establish a connection to the controller if needed if (require_crmd) { xmlNode *xml = NULL; mainloop_io_t *source = mainloop_add_ipc_client(CRM_SYSTEM_CRMD, G_PRIORITY_DEFAULT, 0, NULL, &crm_callbacks); crmd_channel = mainloop_get_ipc_client(source); if (crmd_channel == NULL) { CMD_ERR("Error connecting to the controller"); rc = -ENOTCONN; goto bail; } xml = create_hello_message(our_pid, crm_system_name, "0", "1"); crm_ipc_send(crmd_channel, xml, 0, 0, NULL); free_xml(xml); } /* Handle rsc_cmd appropriately */ if (rsc_cmd == 'L') { rc = pcmk_ok; cli_resource_print_list(data_set, FALSE); } else if (rsc_cmd == 'l') { int found = 0; GListPtr lpc = NULL; rc = pcmk_ok; for (lpc = data_set->resources; lpc != NULL; lpc = lpc->next) { rsc = (resource_t *) lpc->data; found++; cli_resource_print_raw(rsc); } if (found == 0) { printf("NO resources configured\n"); rc = -ENXIO; } } else if (rsc_cmd == 0 && rsc_long_cmd && safe_str_eq(rsc_long_cmd, "restart")) { /* We don't pass data_set because rsc needs to stay valid for the entire * lifetime of cli_resource_restart(), but it will reset and update the * working set multiple times, so it needs to use its own copy. */ rc = cli_resource_restart(rsc, host_uname, timeout_ms, cib_conn); } else if (rsc_cmd == 0 && rsc_long_cmd && safe_str_eq(rsc_long_cmd, "wait")) { rc = wait_till_stable(timeout_ms, cib_conn); } else if (rsc_cmd == 0 && rsc_long_cmd) { // validate, force-(stop|start|demote|promote|check) rc = cli_resource_execute(rsc, rsc_id, rsc_long_cmd, override_params, timeout_ms, cib_conn, data_set); if (rc >= 0) { is_ocf_rc = 1; } } else if (rsc_cmd == 'A' || rsc_cmd == 'a') { GListPtr lpc = NULL; xmlNode *cib_constraints = get_object_root(XML_CIB_TAG_CONSTRAINTS, data_set->input); unpack_constraints(cib_constraints, data_set); // Constraints apply to group/clone, not member/instance rsc = uber_parent(rsc); for (lpc = data_set->resources; lpc != NULL; lpc = lpc->next) { resource_t *r = (resource_t *) lpc->data; clear_bit(r->flags, pe_rsc_allocating); } cli_resource_print_colocation(rsc, TRUE, rsc_cmd == 'A', 1); fprintf(stdout, "* %s\n", rsc->id); cli_resource_print_location(rsc, NULL); for (lpc = data_set->resources; lpc != NULL; lpc = lpc->next) { resource_t *r = (resource_t *) lpc->data; clear_bit(r->flags, pe_rsc_allocating); } cli_resource_print_colocation(rsc, FALSE, rsc_cmd == 'A', 1); } else if (rsc_cmd == 'c') { GListPtr lpc = NULL; rc = pcmk_ok; for (lpc = data_set->resources; lpc != NULL; lpc = lpc->next) { rsc = (resource_t *) lpc->data; cli_resource_print_cts(rsc); } cli_resource_print_cts_constraints(data_set); } else if (rsc_cmd == 'F') { rc = cli_resource_fail(crmd_channel, host_uname, rsc_id, data_set); if (rc == pcmk_ok) { start_mainloop(); } } else if (rsc_cmd == 'O') { rc = cli_resource_print_operations(rsc_id, host_uname, TRUE, data_set); } else if (rsc_cmd == 'o') { rc = cli_resource_print_operations(rsc_id, host_uname, FALSE, data_set); } else if (rsc_cmd == 'W') { rc = cli_resource_search(rsc, rsc_id, data_set); if (rc >= 0) { rc = pcmk_ok; } } else if (rsc_cmd == 'q') { rc = cli_resource_print(rsc, data_set, TRUE); } else if (rsc_cmd == 'w') { rc = cli_resource_print(rsc, data_set, FALSE); } else if (rsc_cmd == 'Y') { node_t *dest = NULL; if (host_uname) { dest = pe_find_node(data_set->nodes, host_uname); if (dest == NULL) { rc = -pcmk_err_node_unknown; goto bail; } } cli_resource_why(cib_conn, data_set->resources, rsc, dest); rc = pcmk_ok; } else if (rsc_cmd == 'U') { GListPtr before = NULL; GListPtr after = NULL; GListPtr remaining = NULL; GListPtr ele = NULL; node_t *dest = NULL; if (BE_QUIET == FALSE) { before = build_constraint_list(data_set->input); } if (clear_expired == TRUE) { rc = cli_resource_clear_all_expired(data_set->input, cib_conn, rsc_id, host_uname, scope_master); } else if (host_uname) { dest = pe_find_node(data_set->nodes, host_uname); if (dest == NULL) { rc = -pcmk_err_node_unknown; if (BE_QUIET == FALSE) { g_list_free(before); } goto bail; } rc = cli_resource_clear(rsc_id, dest->details->uname, NULL, cib_conn, TRUE); } else { rc = cli_resource_clear(rsc_id, NULL, data_set->nodes, cib_conn, TRUE); } if (BE_QUIET == FALSE) { rc = cib_conn->cmds->query(cib_conn, NULL, &cib_xml_copy, cib_scope_local | cib_sync_call); if (rc != pcmk_ok) { CMD_ERR("Could not get modified CIB: %s\n", pcmk_strerror(rc)); g_list_free(before); goto bail; } data_set->input = cib_xml_copy; cluster_status(data_set); after = build_constraint_list(data_set->input); remaining = subtract_lists(before, after, (GCompareFunc) strcmp); for (ele = remaining; ele != NULL; ele = ele->next) { printf("Removing constraint: %s\n", (char *) ele->data); } g_list_free(before); g_list_free(after); g_list_free(remaining); } } else if (rsc_cmd == 'M' && host_uname) { rc = cli_resource_move(rsc, rsc_id, host_uname, cib_conn, data_set); } else if (rsc_cmd == 'B' && host_uname) { node_t *dest = pe_find_node(data_set->nodes, host_uname); if (dest == NULL) { rc = -pcmk_err_node_unknown; goto bail; } rc = cli_resource_ban(rsc_id, dest->details->uname, NULL, cib_conn); } else if (rsc_cmd == 'B' || rsc_cmd == 'M') { pe_node_t *current = NULL; unsigned int nactive = 0; current = pe__find_active_requires(rsc, &nactive); if (nactive == 1) { rc = cli_resource_ban(rsc_id, current->details->uname, NULL, cib_conn); } else if (is_set(rsc->flags, pe_rsc_promotable)) { int count = 0; GListPtr iter = NULL; current = NULL; for(iter = rsc->children; iter; iter = iter->next) { resource_t *child = (resource_t *)iter->data; enum rsc_role_e child_role = child->fns->state(child, TRUE); if(child_role == RSC_ROLE_MASTER) { count++; current = pe__current_node(child); } } if(count == 1 && current) { rc = cli_resource_ban(rsc_id, current->details->uname, NULL, cib_conn); } else { rc = -EINVAL; exit_code = CRM_EX_USAGE; CMD_ERR("Resource '%s' not moved: active in %d locations (promoted in %d).", rsc_id, nactive, count); CMD_ERR("To prevent '%s' from running on a specific location, " "specify a node.", rsc_id); CMD_ERR("To prevent '%s' from being promoted at a specific " "location, specify a node and the master option.", rsc_id); } } else { rc = -EINVAL; exit_code = CRM_EX_USAGE; CMD_ERR("Resource '%s' not moved: active in %d locations.", rsc_id, nactive); CMD_ERR("To prevent '%s' from running on a specific location, " "specify a node.", rsc_id); } } else if (rsc_cmd == 'G') { rc = cli_resource_print_property(rsc, prop_name, data_set); } else if (rsc_cmd == 'S') { xmlNode *msg_data = NULL; if ((rsc_type == NULL) || !strlen(rsc_type)) { CMD_ERR("Must specify -t with resource type"); rc = -ENXIO; goto bail; } else if ((prop_value == NULL) || !strlen(prop_value)) { CMD_ERR("Must supply -v with new value"); rc = -EINVAL; goto bail; } CRM_LOG_ASSERT(prop_name != NULL); msg_data = create_xml_node(NULL, rsc_type); crm_xml_add(msg_data, XML_ATTR_ID, rsc_id); crm_xml_add(msg_data, prop_name, prop_value); rc = cib_conn->cmds->modify(cib_conn, XML_CIB_TAG_RESOURCES, msg_data, cib_options); free_xml(msg_data); } else if (rsc_cmd == 'g') { rc = cli_resource_print_attribute(rsc, prop_name, data_set); } else if (rsc_cmd == 'p') { if (prop_value == NULL || strlen(prop_value) == 0) { CMD_ERR("You need to supply a value with the -v option"); rc = -EINVAL; goto bail; } /* coverity[var_deref_model] False positive */ rc = cli_resource_update_attribute(rsc, rsc_id, prop_set, prop_id, prop_name, prop_value, recursive, cib_conn, data_set); } else if (rsc_cmd == 'd') { /* coverity[var_deref_model] False positive */ rc = cli_resource_delete_attribute(rsc, rsc_id, prop_set, prop_id, prop_name, cib_conn, data_set); } else if ((rsc_cmd == 'C') && rsc) { if (do_force == FALSE) { rsc = uber_parent(rsc); } crmd_replies_needed = 0; crm_debug("Erasing failures of %s (%s requested) on %s", rsc->id, rsc_id, (host_uname? host_uname: "all nodes")); rc = cli_resource_delete(crmd_channel, host_uname, rsc, operation, interval_spec, TRUE, data_set); if ((rc == pcmk_ok) && !BE_QUIET) { // Show any reasons why resource might stay stopped cli_resource_check(cib_conn, rsc); } if (rc == pcmk_ok) { start_mainloop(); } } else if (rsc_cmd == 'C') { rc = cli_cleanup_all(crmd_channel, host_uname, operation, interval_spec, data_set); if (rc == pcmk_ok) { start_mainloop(); } } else if ((rsc_cmd == 'R') && rsc) { if (do_force == FALSE) { rsc = uber_parent(rsc); } crmd_replies_needed = 0; crm_debug("Re-checking the state of %s (%s requested) on %s", rsc->id, rsc_id, (host_uname? host_uname: "all nodes")); rc = cli_resource_delete(crmd_channel, host_uname, rsc, NULL, 0, FALSE, data_set); if ((rc == pcmk_ok) && !BE_QUIET) { // Show any reasons why resource might stay stopped cli_resource_check(cib_conn, rsc); } if (rc == pcmk_ok) { start_mainloop(); } } else if (rsc_cmd == 'R') { const char *router_node = host_uname; xmlNode *msg_data = NULL; xmlNode *cmd = NULL; int attr_options = attrd_opt_none; if (host_uname) { node_t *node = pe_find_node(data_set->nodes, host_uname); if (pe__is_guest_or_remote_node(node)) { node = pe__current_node(node->details->remote_rsc); if (node == NULL) { CMD_ERR("No cluster connection to Pacemaker Remote node %s detected", host_uname); rc = -ENXIO; goto bail; } router_node = node->details->uname; attr_options |= attrd_opt_remote; } } if (crmd_channel == NULL) { printf("Dry run: skipping clean-up of %s due to CIB_file\n", host_uname? host_uname : "all nodes"); rc = pcmk_ok; goto bail; } msg_data = create_xml_node(NULL, "crm-resource-reprobe-op"); crm_xml_add(msg_data, XML_LRM_ATTR_TARGET, host_uname); if (safe_str_neq(router_node, host_uname)) { crm_xml_add(msg_data, XML_LRM_ATTR_ROUTER_NODE, router_node); } cmd = create_request(CRM_OP_REPROBE, msg_data, router_node, CRM_SYSTEM_CRMD, crm_system_name, our_pid); free_xml(msg_data); crm_debug("Re-checking the state of all resources on %s", host_uname?host_uname:"all nodes"); rc = attrd_clear_delegate(NULL, host_uname, NULL, NULL, NULL, NULL, attr_options); if (crm_ipc_send(crmd_channel, cmd, 0, 0, NULL) > 0) { start_mainloop(); } free_xml(cmd); } else if (rsc_cmd == 'D') { xmlNode *msg_data = NULL; if (rsc_type == NULL) { CMD_ERR("You need to specify a resource type with -t"); rc = -ENXIO; goto bail; } msg_data = create_xml_node(NULL, rsc_type); crm_xml_add(msg_data, XML_ATTR_ID, rsc_id); rc = cib_conn->cmds->remove(cib_conn, XML_CIB_TAG_RESOURCES, msg_data, cib_options); free_xml(msg_data); } else { CMD_ERR("Unknown command: %c", rsc_cmd); } bail: free(our_pid); pe_free_working_set(data_set); if (cib_conn != NULL) { cib_conn->cmds->signoff(cib_conn); cib_delete(cib_conn); } if (is_ocf_rc) { exit_code = rc; } else if (rc != pcmk_ok) { CMD_ERR("Error performing operation: %s", pcmk_strerror(rc)); if (rc == -pcmk_err_no_quorum) { CMD_ERR("To ignore quorum, use the force option"); } if (exit_code == CRM_EX_OK) { exit_code = crm_errno2exit(rc); } } crm_exit(exit_code); } diff --git a/tools/crm_resource_runtime.c b/tools/crm_resource_runtime.c index fab2be25ba..01919f67e2 100644 --- a/tools/crm_resource_runtime.c +++ b/tools/crm_resource_runtime.c @@ -1,2101 +1,2103 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include int resource_verbose = 0; bool do_force = FALSE; int crmd_replies_needed = 1; /* The welcome message */ const char *attr_set_type = XML_TAG_ATTR_SETS; static int do_find_resource(const char *rsc, resource_t * the_rsc, pe_working_set_t * data_set) { int found = 0; GListPtr lpc = NULL; for (lpc = the_rsc->running_on; lpc != NULL; lpc = lpc->next) { node_t *node = (node_t *) lpc->data; if (BE_QUIET) { fprintf(stdout, "%s\n", node->details->uname); } else { const char *state = ""; if (!pe_rsc_is_clone(the_rsc) && the_rsc->fns->state(the_rsc, TRUE) == RSC_ROLE_MASTER) { state = "Master"; } fprintf(stdout, "resource %s is running on: %s %s\n", rsc, node->details->uname, state); } found++; } if (BE_QUIET == FALSE && found == 0) { fprintf(stderr, "resource %s is NOT running\n", rsc); } return found; } int cli_resource_search(resource_t *rsc, const char *requested_name, pe_working_set_t *data_set) { int found = 0; resource_t *parent = uber_parent(rsc); if (pe_rsc_is_clone(rsc)) { for (GListPtr iter = rsc->children; iter != NULL; iter = iter->next) { found += do_find_resource(requested_name, iter->data, data_set); } /* The anonymous clone children's common ID is supplied */ } else if (pe_rsc_is_clone(parent) && is_not_set(rsc->flags, pe_rsc_unique) && rsc->clone_name && safe_str_eq(requested_name, rsc->clone_name) && safe_str_neq(requested_name, rsc->id)) { for (GListPtr iter = parent->children; iter; iter = iter->next) { found += do_find_resource(requested_name, iter->data, data_set); } } else { found += do_find_resource(requested_name, rsc, data_set); } return found; } #define XPATH_MAX 1024 static int find_resource_attr(cib_t * the_cib, const char *attr, const char *rsc, const char *set_type, const char *set_name, const char *attr_id, const char *attr_name, char **value) { int offset = 0; int rc = pcmk_ok; xmlNode *xml_search = NULL; char *xpath_string = NULL; if(value) { *value = NULL; } if(the_cib == NULL) { return -ENOTCONN; } xpath_string = calloc(1, XPATH_MAX); offset += snprintf(xpath_string + offset, XPATH_MAX - offset, "%s", get_object_path("resources")); offset += snprintf(xpath_string + offset, XPATH_MAX - offset, "//*[@id=\"%s\"]", rsc); if (set_type) { offset += snprintf(xpath_string + offset, XPATH_MAX - offset, "/%s", set_type); if (set_name) { offset += snprintf(xpath_string + offset, XPATH_MAX - offset, "[@id=\"%s\"]", set_name); } } offset += snprintf(xpath_string + offset, XPATH_MAX - offset, "//nvpair["); if (attr_id) { offset += snprintf(xpath_string + offset, XPATH_MAX - offset, "@id=\"%s\"", attr_id); } if (attr_name) { if (attr_id) { offset += snprintf(xpath_string + offset, XPATH_MAX - offset, " and "); } offset += snprintf(xpath_string + offset, XPATH_MAX - offset, "@name=\"%s\"", attr_name); } offset += snprintf(xpath_string + offset, XPATH_MAX - offset, "]"); CRM_LOG_ASSERT(offset > 0); rc = the_cib->cmds->query(the_cib, xpath_string, &xml_search, cib_sync_call | cib_scope_local | cib_xpath); if (rc != pcmk_ok) { goto bail; } crm_log_xml_debug(xml_search, "Match"); if (xml_has_children(xml_search)) { xmlNode *child = NULL; rc = -EINVAL; printf("Multiple attributes match name=%s\n", attr_name); for (child = __xml_first_child(xml_search); child != NULL; child = __xml_next(child)) { printf(" Value: %s \t(id=%s)\n", crm_element_value(child, XML_NVPAIR_ATTR_VALUE), ID(child)); } } else if(value) { const char *tmp = crm_element_value(xml_search, attr); if (tmp) { *value = strdup(tmp); } } bail: free(xpath_string); free_xml(xml_search); return rc; } /* PRIVATE. Use the find_matching_attr_resources instead. */ static void find_matching_attr_resources_recursive(GList/* */ ** result, resource_t * rsc, const char * rsc_id, const char * attr_set, const char * attr_id, const char * attr_name, cib_t * cib, const char * cmd, int depth) { int rc = pcmk_ok; char *lookup_id = clone_strip(rsc->id); char *local_attr_id = NULL; /* visit the children */ for(GList *gIter = rsc->children; gIter; gIter = gIter->next) { find_matching_attr_resources_recursive(result, (resource_t*)gIter->data, rsc_id, attr_set, attr_id, attr_name, cib, cmd, depth+1); /* do it only once for clones */ if(pe_clone == rsc->variant) { break; } } rc = find_resource_attr(cib, XML_ATTR_ID, lookup_id, attr_set_type, attr_set, attr_id, attr_name, &local_attr_id); /* Post-order traversal. * The root is always on the list and it is the last item. */ if((0 == depth) || (pcmk_ok == rc)) { /* push the head */ *result = g_list_append(*result, rsc); } free(local_attr_id); free(lookup_id); } /* The result is a linearized pre-ordered tree of resources. */ static GList/**/ * find_matching_attr_resources(resource_t * rsc, const char * rsc_id, const char * attr_set, const char * attr_id, const char * attr_name, cib_t * cib, const char * cmd) { int rc = pcmk_ok; char *lookup_id = NULL; char *local_attr_id = NULL; GList * result = NULL; /* If --force is used, update only the requested resource (clone or primitive). * Otherwise, if the primitive has the attribute, use that. * Otherwise use the clone. */ if(do_force == TRUE) { return g_list_append(result, rsc); } if(rsc->parent && pe_clone == rsc->parent->variant) { int rc = pcmk_ok; char *local_attr_id = NULL; rc = find_resource_attr(cib, XML_ATTR_ID, rsc_id, attr_set_type, attr_set, attr_id, attr_name, &local_attr_id); free(local_attr_id); if(rc != pcmk_ok) { rsc = rsc->parent; if (BE_QUIET == FALSE) { printf("Performing %s of '%s' on '%s', the parent of '%s'\n", cmd, attr_name, rsc->id, rsc_id); } } return g_list_append(result, rsc); } else if(rsc->parent == NULL && rsc->children && pe_clone == rsc->variant) { resource_t *child = rsc->children->data; if(child->variant == pe_native) { lookup_id = clone_strip(child->id); /* Could be a cloned group! */ rc = find_resource_attr(cib, XML_ATTR_ID, lookup_id, attr_set_type, attr_set, attr_id, attr_name, &local_attr_id); if(rc == pcmk_ok) { rsc = child; if (BE_QUIET == FALSE) { printf("A value for '%s' already exists in child '%s', performing %s on that instead of '%s'\n", attr_name, lookup_id, cmd, rsc_id); } } free(local_attr_id); free(lookup_id); } return g_list_append(result, rsc); } /* If the resource is a group ==> children inherit the attribute if defined. */ find_matching_attr_resources_recursive(&result, rsc, rsc_id, attr_set, attr_id, attr_name, cib, cmd, 0); return result; } int cli_resource_update_attribute(resource_t *rsc, const char *requested_name, const char *attr_set, const char *attr_id, const char *attr_name, const char *attr_value, bool recursive, cib_t *cib, pe_working_set_t *data_set) { int rc = pcmk_ok; static bool need_init = TRUE; char *local_attr_id = NULL; char *local_attr_set = NULL; GList/**/ *resources = NULL; const char *common_attr_id = attr_id; if(attr_id == NULL && do_force == FALSE && find_resource_attr( cib, XML_ATTR_ID, uber_parent(rsc)->id, NULL, NULL, NULL, attr_name, NULL) == -EINVAL) { printf("\n"); } if (safe_str_eq(attr_set_type, XML_TAG_ATTR_SETS)) { if (do_force == FALSE) { rc = find_resource_attr(cib, XML_ATTR_ID, uber_parent(rsc)->id, XML_TAG_META_SETS, attr_set, attr_id, attr_name, &local_attr_id); if (rc == pcmk_ok && BE_QUIET == FALSE) { printf("WARNING: There is already a meta attribute for '%s' called '%s' (id=%s)\n", uber_parent(rsc)->id, attr_name, local_attr_id); printf(" Delete '%s' first or use the force option to override\n", local_attr_id); } free(local_attr_id); if (rc == pcmk_ok) { return -ENOTUNIQ; } } resources = g_list_append(resources, rsc); } else { resources = find_matching_attr_resources(rsc, requested_name, attr_set, attr_id, attr_name, cib, "update"); } /* If either attr_set or attr_id is specified, * one clearly intends to modify a single resource. * It is the last item on the resource list.*/ for(GList *gIter = (attr_set||attr_id) ? g_list_last(resources) : resources ; gIter; gIter = gIter->next) { char *lookup_id = NULL; xmlNode *xml_top = NULL; xmlNode *xml_obj = NULL; local_attr_id = NULL; local_attr_set = NULL; rsc = (resource_t*)gIter->data; attr_id = common_attr_id; lookup_id = clone_strip(rsc->id); /* Could be a cloned group! */ rc = find_resource_attr(cib, XML_ATTR_ID, lookup_id, attr_set_type, attr_set, attr_id, attr_name, &local_attr_id); if (rc == pcmk_ok) { crm_debug("Found a match for name=%s: id=%s", attr_name, local_attr_id); attr_id = local_attr_id; } else if (rc != -ENXIO) { free(lookup_id); free(local_attr_id); g_list_free(resources); return rc; } else { const char *tag = crm_element_name(rsc->xml); if (attr_set == NULL) { local_attr_set = crm_concat(lookup_id, attr_set_type, '-'); attr_set = local_attr_set; } if (attr_id == NULL) { local_attr_id = crm_concat(attr_set, attr_name, '-'); attr_id = local_attr_id; } xml_top = create_xml_node(NULL, tag); crm_xml_add(xml_top, XML_ATTR_ID, lookup_id); xml_obj = create_xml_node(xml_top, attr_set_type); crm_xml_add(xml_obj, XML_ATTR_ID, attr_set); } xml_obj = crm_create_nvpair_xml(xml_obj, attr_id, attr_name, attr_value); if (xml_top == NULL) { xml_top = xml_obj; } crm_log_xml_debug(xml_top, "Update"); rc = cib->cmds->modify(cib, XML_CIB_TAG_RESOURCES, xml_top, cib_options); if (rc == pcmk_ok && BE_QUIET == FALSE) { printf("Set '%s' option: id=%s%s%s%s%s value=%s\n", lookup_id, local_attr_id, attr_set ? " set=" : "", attr_set ? attr_set : "", attr_name ? " name=" : "", attr_name ? attr_name : "", attr_value); } free_xml(xml_top); free(lookup_id); free(local_attr_id); free(local_attr_set); if(recursive && safe_str_eq(attr_set_type, XML_TAG_META_SETS)) { GListPtr lpc = NULL; if(need_init) { xmlNode *cib_constraints = get_object_root(XML_CIB_TAG_CONSTRAINTS, data_set->input); need_init = FALSE; unpack_constraints(cib_constraints, data_set); for (lpc = data_set->resources; lpc != NULL; lpc = lpc->next) { resource_t *r = (resource_t *) lpc->data; clear_bit(r->flags, pe_rsc_allocating); } } crm_debug("Looking for dependencies %p", rsc->rsc_cons_lhs); set_bit(rsc->flags, pe_rsc_allocating); for (lpc = rsc->rsc_cons_lhs; lpc != NULL; lpc = lpc->next) { rsc_colocation_t *cons = (rsc_colocation_t *) lpc->data; resource_t *peer = cons->rsc_lh; crm_debug("Checking %s %d", cons->id, cons->score); if (cons->score > 0 && is_not_set(peer->flags, pe_rsc_allocating)) { /* Don't get into colocation loops */ crm_debug("Setting %s=%s for dependent resource %s", attr_name, attr_value, peer->id); cli_resource_update_attribute(peer, peer->id, NULL, NULL, attr_name, attr_value, recursive, cib, data_set); } } } } g_list_free(resources); return rc; } int cli_resource_delete_attribute(resource_t *rsc, const char *requested_name, const char *attr_set, const char *attr_id, const char *attr_name, cib_t *cib, pe_working_set_t *data_set) { int rc = pcmk_ok; GList/**/ *resources = NULL; if(attr_id == NULL && do_force == FALSE && find_resource_attr( cib, XML_ATTR_ID, uber_parent(rsc)->id, NULL, NULL, NULL, attr_name, NULL) == -EINVAL) { printf("\n"); } if(safe_str_eq(attr_set_type, XML_TAG_META_SETS)) { resources = find_matching_attr_resources(rsc, requested_name, attr_set, attr_id, attr_name, cib, "delete"); } else { resources = g_list_append(resources, rsc); } for(GList *gIter = resources; gIter; gIter = gIter->next) { char *lookup_id = NULL; xmlNode *xml_obj = NULL; char *local_attr_id = NULL; rsc = (resource_t*)gIter->data; lookup_id = clone_strip(rsc->id); rc = find_resource_attr(cib, XML_ATTR_ID, lookup_id, attr_set_type, attr_set, attr_id, attr_name, &local_attr_id); if (rc == -ENXIO) { free(lookup_id); rc = pcmk_ok; continue; } else if (rc != pcmk_ok) { free(lookup_id); g_list_free(resources); return rc; } if (attr_id == NULL) { attr_id = local_attr_id; } xml_obj = crm_create_nvpair_xml(NULL, attr_id, attr_name, NULL); crm_log_xml_debug(xml_obj, "Delete"); CRM_ASSERT(cib); rc = cib->cmds->remove(cib, XML_CIB_TAG_RESOURCES, xml_obj, cib_options); if (rc == pcmk_ok && BE_QUIET == FALSE) { printf("Deleted '%s' option: id=%s%s%s%s%s\n", lookup_id, local_attr_id, attr_set ? " set=" : "", attr_set ? attr_set : "", attr_name ? " name=" : "", attr_name ? attr_name : ""); } free(lookup_id); free_xml(xml_obj); free(local_attr_id); } g_list_free(resources); return rc; } static int send_lrm_rsc_op(crm_ipc_t * crmd_channel, const char *op, const char *host_uname, const char *rsc_id, bool only_failed, pe_working_set_t * data_set) { char *our_pid = NULL; char *key = NULL; int rc = -ECOMM; xmlNode *cmd = NULL; xmlNode *xml_rsc = NULL; const char *value = NULL; const char *router_node = host_uname; xmlNode *params = NULL; xmlNode *msg_data = NULL; resource_t *rsc = pe_find_resource(data_set->resources, rsc_id); if (rsc == NULL) { CMD_ERR("Resource %s not found", rsc_id); return -ENXIO; } else if (rsc->variant != pe_native) { CMD_ERR("We can only process primitive resources, not %s", rsc_id); return -EINVAL; } else if (host_uname == NULL) { CMD_ERR("Please specify a node name"); return -EINVAL; } else { node_t *node = pe_find_node(data_set->nodes, host_uname); if (pe__is_guest_or_remote_node(node)) { node = pe__current_node(node->details->remote_rsc); if (node == NULL) { CMD_ERR("No cluster connection to Pacemaker Remote node %s detected", host_uname); return -ENXIO; } router_node = node->details->uname; } } key = generate_transition_key(0, getpid(), 0, "xxxxxxxx-xrsc-opxx-xcrm-resourcexxxx"); msg_data = create_xml_node(NULL, XML_GRAPH_TAG_RSC_OP); crm_xml_add(msg_data, XML_ATTR_TRANSITION_KEY, key); free(key); crm_xml_add(msg_data, XML_LRM_ATTR_TARGET, host_uname); if (safe_str_neq(router_node, host_uname)) { crm_xml_add(msg_data, XML_LRM_ATTR_ROUTER_NODE, router_node); } xml_rsc = create_xml_node(msg_data, XML_CIB_TAG_RESOURCE); if (rsc->clone_name) { crm_xml_add(xml_rsc, XML_ATTR_ID, rsc->clone_name); crm_xml_add(xml_rsc, XML_ATTR_ID_LONG, rsc->id); } else { crm_xml_add(xml_rsc, XML_ATTR_ID, rsc->id); } value = crm_copy_xml_element(rsc->xml, xml_rsc, XML_ATTR_TYPE); if (value == NULL) { CMD_ERR("%s has no type! Aborting...", rsc_id); return -ENXIO; } value = crm_copy_xml_element(rsc->xml, xml_rsc, XML_AGENT_ATTR_CLASS); if (value == NULL) { CMD_ERR("%s has no class! Aborting...", rsc_id); return -ENXIO; } crm_copy_xml_element(rsc->xml, xml_rsc, XML_AGENT_ATTR_PROVIDER); params = create_xml_node(msg_data, XML_TAG_ATTRS); crm_xml_add(params, XML_ATTR_CRM_VERSION, CRM_FEATURE_SET); key = crm_meta_name(XML_LRM_ATTR_INTERVAL_MS); crm_xml_add(params, key, "60000"); /* 1 minute */ free(key); our_pid = crm_getpid_s(); cmd = create_request(op, msg_data, router_node, CRM_SYSTEM_CRMD, crm_system_name, our_pid); /* crm_log_xml_warn(cmd, "send_lrm_rsc_op"); */ free_xml(msg_data); if (crm_ipc_send(crmd_channel, cmd, 0, 0, NULL) > 0) { rc = 0; } else { crm_debug("Could not send %s op to the controller", op); rc = -ENOTCONN; } free_xml(cmd); return rc; } /*! * \internal * \brief Get resource name as used in failure-related node attributes * * \param[in] rsc Resource to check * * \return Newly allocated string containing resource's fail name * \note The caller is responsible for freeing the result. */ static inline char * rsc_fail_name(resource_t *rsc) { const char *name = (rsc->clone_name? rsc->clone_name : rsc->id); return is_set(rsc->flags, pe_rsc_unique)? strdup(name) : clone_strip(name); } static int clear_rsc_history(crm_ipc_t *crmd_channel, const char *host_uname, const char *rsc_id, pe_working_set_t *data_set) { int rc = pcmk_ok; /* Erase the resource's entire LRM history in the CIB, even if we're only * clearing a single operation's fail count. If we erased only entries for a * single operation, we might wind up with a wrong idea of the current * resource state, and we might not re-probe the resource. */ rc = send_lrm_rsc_op(crmd_channel, CRM_OP_LRM_DELETE, host_uname, rsc_id, TRUE, data_set); if (rc != pcmk_ok) { return rc; } crmd_replies_needed++; crm_trace("Processing %d mainloop inputs", crmd_replies_needed); while (g_main_context_iteration(NULL, FALSE)) { crm_trace("Processed mainloop input, %d still remaining", crmd_replies_needed); } if (crmd_replies_needed < 0) { crmd_replies_needed = 0; } return rc; } static int clear_rsc_failures(crm_ipc_t *crmd_channel, const char *node_name, const char *rsc_id, const char *operation, const char *interval_spec, pe_working_set_t *data_set) { int rc = pcmk_ok; const char *failed_value = NULL; const char *failed_id = NULL; const char *interval_ms_s = NULL; GHashTable *rscs = NULL; GHashTableIter iter; /* Create a hash table to use as a set of resources to clean. This lets us * clean each resource only once (per node) regardless of how many failed * operations it has. */ rscs = g_hash_table_new_full(crm_str_hash, g_str_equal, NULL, NULL); // Normalize interval to milliseconds for comparison to history entry if (operation) { interval_ms_s = crm_strdup_printf("%u", crm_parse_interval_spec(interval_spec)); } for (xmlNode *xml_op = __xml_first_child(data_set->failed); xml_op != NULL; xml_op = __xml_next(xml_op)) { failed_id = crm_element_value(xml_op, XML_LRM_ATTR_RSCID); if (failed_id == NULL) { // Malformed history entry, should never happen continue; } // No resource specified means all resources match if (rsc_id) { resource_t *fail_rsc = pe_find_resource_with_flags(data_set->resources, failed_id, pe_find_renamed|pe_find_anon); if (!fail_rsc || safe_str_neq(rsc_id, fail_rsc->id)) { continue; } } // Host name should always have been provided by this point failed_value = crm_element_value(xml_op, XML_ATTR_UNAME); if (safe_str_neq(node_name, failed_value)) { continue; } // No operation specified means all operations match if (operation) { failed_value = crm_element_value(xml_op, XML_LRM_ATTR_TASK); if (safe_str_neq(operation, failed_value)) { continue; } // Interval (if operation was specified) defaults to 0 (not all) failed_value = crm_element_value(xml_op, XML_LRM_ATTR_INTERVAL_MS); if (safe_str_neq(interval_ms_s, failed_value)) { continue; } } /* not available until glib 2.32 g_hash_table_add(rscs, (gpointer) failed_id); */ g_hash_table_insert(rscs, (gpointer) failed_id, (gpointer) failed_id); } g_hash_table_iter_init(&iter, rscs); while (g_hash_table_iter_next(&iter, (gpointer *) &failed_id, NULL)) { crm_debug("Erasing failures of %s on %s", failed_id, node_name); rc = clear_rsc_history(crmd_channel, node_name, failed_id, data_set); if (rc != pcmk_ok) { return rc; } } g_hash_table_destroy(rscs); return rc; } static int clear_rsc_fail_attrs(resource_t *rsc, const char *operation, const char *interval_spec, node_t *node) { int rc = pcmk_ok; int attr_options = attrd_opt_none; char *rsc_name = rsc_fail_name(rsc); if (pe__is_guest_or_remote_node(node)) { attr_options |= attrd_opt_remote; } rc = attrd_clear_delegate(NULL, node->details->uname, rsc_name, operation, interval_spec, NULL, attr_options); free(rsc_name); return rc; } int cli_resource_delete(crm_ipc_t *crmd_channel, const char *host_uname, resource_t *rsc, const char *operation, const char *interval_spec, bool just_failures, pe_working_set_t *data_set) { int rc = pcmk_ok; node_t *node = NULL; if (rsc == NULL) { return -ENXIO; } else if (rsc->children) { GListPtr lpc = NULL; for (lpc = rsc->children; lpc != NULL; lpc = lpc->next) { resource_t *child = (resource_t *) lpc->data; rc = cli_resource_delete(crmd_channel, host_uname, child, operation, interval_spec, just_failures, data_set); if (rc != pcmk_ok) { return rc; } } return pcmk_ok; } else if (host_uname == NULL) { GListPtr lpc = NULL; GListPtr nodes = g_hash_table_get_values(rsc->known_on); if(nodes == NULL && do_force) { nodes = node_list_dup(data_set->nodes, FALSE, FALSE); } else if(nodes == NULL && rsc->exclusive_discover) { GHashTableIter iter; pe_node_t *node = NULL; g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void**)&node)) { if(node->weight >= 0) { nodes = g_list_prepend(nodes, node); } } } else if(nodes == NULL) { nodes = g_hash_table_get_values(rsc->allowed_nodes); } for (lpc = nodes; lpc != NULL; lpc = lpc->next) { node = (node_t *) lpc->data; if (node->details->online) { rc = cli_resource_delete(crmd_channel, node->details->uname, rsc, operation, interval_spec, just_failures, data_set); } if (rc != pcmk_ok) { g_list_free(nodes); return rc; } } g_list_free(nodes); return pcmk_ok; } node = pe_find_node(data_set->nodes, host_uname); if (node == NULL) { printf("Unable to clean up %s because node %s not found\n", rsc->id, host_uname); return -ENODEV; } if (!node->details->rsc_discovery_enabled) { printf("Unable to clean up %s because resource discovery disabled on %s\n", rsc->id, host_uname); return -EOPNOTSUPP; } if (crmd_channel == NULL) { printf("Dry run: skipping clean-up of %s on %s due to CIB_file\n", rsc->id, host_uname); return pcmk_ok; } rc = clear_rsc_fail_attrs(rsc, operation, interval_spec, node); if (rc != pcmk_ok) { printf("Unable to clean up %s failures on %s: %s\n", rsc->id, host_uname, pcmk_strerror(rc)); return rc; } if (just_failures) { rc = clear_rsc_failures(crmd_channel, host_uname, rsc->id, operation, interval_spec, data_set); } else { rc = clear_rsc_history(crmd_channel, host_uname, rsc->id, data_set); } if (rc != pcmk_ok) { printf("Cleaned %s failures on %s, but unable to clean history: %s\n", rsc->id, host_uname, pcmk_strerror(rc)); } else { printf("Cleaned up %s on %s\n", rsc->id, host_uname); } return rc; } int cli_cleanup_all(crm_ipc_t *crmd_channel, const char *node_name, const char *operation, const char *interval_spec, pe_working_set_t *data_set) { int rc = pcmk_ok; int attr_options = attrd_opt_none; const char *display_name = node_name? node_name : "all nodes"; if (crmd_channel == NULL) { printf("Dry run: skipping clean-up of %s due to CIB_file\n", display_name); return pcmk_ok; } crmd_replies_needed = 0; if (node_name) { node_t *node = pe_find_node(data_set->nodes, node_name); if (node == NULL) { CMD_ERR("Unknown node: %s", node_name); return -ENXIO; } if (pe__is_guest_or_remote_node(node)) { attr_options |= attrd_opt_remote; } } rc = attrd_clear_delegate(NULL, node_name, NULL, operation, interval_spec, NULL, attr_options); if (rc != pcmk_ok) { printf("Unable to clean up all failures on %s: %s\n", display_name, pcmk_strerror(rc)); return rc; } if (node_name) { rc = clear_rsc_failures(crmd_channel, node_name, NULL, operation, interval_spec, data_set); if (rc != pcmk_ok) { printf("Cleaned all resource failures on %s, but unable to clean history: %s\n", node_name, pcmk_strerror(rc)); return rc; } } else { for (GList *iter = data_set->nodes; iter; iter = iter->next) { pe_node_t *node = (pe_node_t *) iter->data; rc = clear_rsc_failures(crmd_channel, node->details->uname, NULL, operation, interval_spec, data_set); if (rc != pcmk_ok) { printf("Cleaned all resource failures on all nodes, but unable to clean history: %s\n", pcmk_strerror(rc)); return rc; } } } printf("Cleaned up all resources on %s\n", display_name); return pcmk_ok; } void cli_resource_check(cib_t * cib_conn, resource_t *rsc) { int need_nl = 0; char *role_s = NULL; char *managed = NULL; resource_t *parent = uber_parent(rsc); find_resource_attr(cib_conn, XML_NVPAIR_ATTR_VALUE, parent->id, NULL, NULL, NULL, XML_RSC_ATTR_MANAGED, &managed); find_resource_attr(cib_conn, XML_NVPAIR_ATTR_VALUE, parent->id, NULL, NULL, NULL, XML_RSC_ATTR_TARGET_ROLE, &role_s); if(role_s) { enum rsc_role_e role = text2role(role_s); free(role_s); if(role == RSC_ROLE_UNKNOWN) { // Treated as if unset } else if(role == RSC_ROLE_STOPPED) { printf("\n * The configuration specifies that '%s' should remain stopped\n", parent->id); need_nl++; } else if (is_set(parent->flags, pe_rsc_promotable) && (role == RSC_ROLE_SLAVE)) { printf("\n * The configuration specifies that '%s' should not be promoted\n", parent->id); need_nl++; } } if(managed && crm_is_true(managed) == FALSE) { printf("%s * The configuration prevents the cluster from stopping or starting '%s' (unmanaged)\n", need_nl == 0?"\n":"", parent->id); need_nl++; } free(managed); if(need_nl) { printf("\n"); } } int cli_resource_fail(crm_ipc_t * crmd_channel, const char *host_uname, const char *rsc_id, pe_working_set_t * data_set) { crm_warn("Failing: %s", rsc_id); return send_lrm_rsc_op(crmd_channel, CRM_OP_LRM_FAIL, host_uname, rsc_id, FALSE, data_set); } static GHashTable * generate_resource_params(resource_t * rsc, pe_working_set_t * data_set) { GHashTable *params = NULL; GHashTable *meta = NULL; GHashTable *combined = NULL; GHashTableIter iter; if (!rsc) { crm_err("Resource does not exist in config"); return NULL; } params = crm_str_table_new(); meta = crm_str_table_new(); combined = crm_str_table_new(); get_rsc_attributes(params, rsc, NULL /* TODO: Pass in local node */ , data_set); get_meta_attributes(meta, rsc, NULL /* TODO: Pass in local node */ , data_set); if (params) { char *key = NULL; char *value = NULL; g_hash_table_iter_init(&iter, params); while (g_hash_table_iter_next(&iter, (gpointer *) & key, (gpointer *) & value)) { g_hash_table_insert(combined, strdup(key), strdup(value)); } g_hash_table_destroy(params); } if (meta) { char *key = NULL; char *value = NULL; g_hash_table_iter_init(&iter, meta); while (g_hash_table_iter_next(&iter, (gpointer *) & key, (gpointer *) & value)) { char *crm_name = crm_meta_name(key); g_hash_table_insert(combined, crm_name, strdup(value)); } g_hash_table_destroy(meta); } return combined; } static bool resource_is_running_on(resource_t *rsc, const char *host) { bool found = TRUE; GListPtr hIter = NULL; GListPtr hosts = NULL; if(rsc == NULL) { return FALSE; } rsc->fns->location(rsc, &hosts, TRUE); for (hIter = hosts; host != NULL && hIter != NULL; hIter = hIter->next) { pe_node_t *node = (pe_node_t *) hIter->data; if(strcmp(host, node->details->uname) == 0) { crm_trace("Resource %s is running on %s\n", rsc->id, host); goto done; } else if(strcmp(host, node->details->id) == 0) { crm_trace("Resource %s is running on %s\n", rsc->id, host); goto done; } } if(host != NULL) { crm_trace("Resource %s is not running on: %s\n", rsc->id, host); found = FALSE; } else if(host == NULL && hosts == NULL) { crm_trace("Resource %s is not running\n", rsc->id); found = FALSE; } done: g_list_free(hosts); return found; } /*! * \internal * \brief Create a list of all resources active on host from a given list * * \param[in] host Name of host to check whether resources are active * \param[in] rsc_list List of resources to check * * \return New list of resources from list that are active on host */ static GList * get_active_resources(const char *host, GList *rsc_list) { GList *rIter = NULL; GList *active = NULL; for (rIter = rsc_list; rIter != NULL; rIter = rIter->next) { resource_t *rsc = (resource_t *) rIter->data; /* Expand groups to their members, because if we're restarting a member * other than the first, we can't otherwise tell which resources are * stopping and starting. */ if (rsc->variant == pe_group) { active = g_list_concat(active, get_active_resources(host, rsc->children)); } else if (resource_is_running_on(rsc, host)) { active = g_list_append(active, strdup(rsc->id)); } } return active; } GList* subtract_lists(GList *from, GList *items, GCompareFunc cmp) { GList *item = NULL; GList *result = g_list_copy(from); for (item = items; item != NULL; item = item->next) { GList *candidate = NULL; for (candidate = from; candidate != NULL; candidate = candidate->next) { crm_info("Comparing %s with %s", (const char *) candidate->data, (const char *) item->data); if(cmp(candidate->data, item->data) == 0) { result = g_list_remove(result, candidate->data); break; } } } return result; } static void dump_list(GList *items, const char *tag) { int lpc = 0; GList *item = NULL; for (item = items; item != NULL; item = item->next) { crm_trace("%s[%d]: %s", tag, lpc, (char*)item->data); lpc++; } } static void display_list(GList *items, const char *tag) { GList *item = NULL; for (item = items; item != NULL; item = item->next) { fprintf(stdout, "%s%s\n", tag, (const char *)item->data); } } /*! * \internal * \brief Upgrade XML to latest schema version and use it as working set input * * This also updates the working set timestamp to the current time. * * \param[in] data_set Working set instance to update * \param[in] xml XML to use as input * * \return pcmk_ok on success, -ENOKEY if unable to upgrade XML * \note On success, caller is responsible for freeing memory allocated for * data_set->now. * \todo This follows the example of other callers of cli_config_update() * and returns -ENOKEY ("Required key not available") if that fails, * but perhaps -pcmk_err_schema_validation would be better in that case. */ int update_working_set_xml(pe_working_set_t *data_set, xmlNode **xml) { if (cli_config_update(xml, NULL, FALSE) == FALSE) { return -ENOKEY; } data_set->input = *xml; data_set->now = crm_time_new(NULL); return pcmk_ok; } /*! * \internal * \brief Update a working set's XML input based on a CIB query * * \param[in] data_set Data set instance to initialize * \param[in] cib Connection to the CIB manager * * \return pcmk_ok on success, -errno on failure * \note On success, caller is responsible for freeing memory allocated for * data_set->input and data_set->now. */ static int update_working_set_from_cib(pe_working_set_t * data_set, cib_t *cib) { xmlNode *cib_xml_copy = NULL; int rc; rc = cib->cmds->query(cib, NULL, &cib_xml_copy, cib_scope_local | cib_sync_call); if (rc != pcmk_ok) { fprintf(stderr, "Could not obtain the current CIB: %s (%d)\n", pcmk_strerror(rc), rc); return rc; } rc = update_working_set_xml(data_set, &cib_xml_copy); if (rc != pcmk_ok) { fprintf(stderr, "Could not upgrade the current CIB XML\n"); free_xml(cib_xml_copy); return rc; } return pcmk_ok; } static int update_dataset(cib_t *cib, pe_working_set_t * data_set, bool simulate) { char *pid = NULL; char *shadow_file = NULL; cib_t *shadow_cib = NULL; int rc; pe_reset_working_set(data_set); rc = update_working_set_from_cib(data_set, cib); if (rc != pcmk_ok) { return rc; } if(simulate) { pid = crm_getpid_s(); shadow_cib = cib_shadow_new(pid); shadow_file = get_shadow_file(pid); if (shadow_cib == NULL) { fprintf(stderr, "Could not create shadow cib: '%s'\n", pid); rc = -ENXIO; goto cleanup; } rc = write_xml_file(data_set->input, shadow_file, FALSE); if (rc < 0) { fprintf(stderr, "Could not populate shadow cib: %s (%d)\n", pcmk_strerror(rc), rc); goto cleanup; } rc = shadow_cib->cmds->signon(shadow_cib, crm_system_name, cib_command); if(rc != pcmk_ok) { fprintf(stderr, "Could not connect to shadow cib: %s (%d)\n", pcmk_strerror(rc), rc); goto cleanup; } pcmk__schedule_actions(data_set, data_set->input, NULL); run_simulation(data_set, shadow_cib, NULL, TRUE); rc = update_dataset(shadow_cib, data_set, FALSE); } else { cluster_status(data_set); } cleanup: /* Do not free data_set->input here, we need rsc->xml to be valid later on */ cib_delete(shadow_cib); free(pid); if(shadow_file) { unlink(shadow_file); free(shadow_file); } return rc; } static int max_delay_for_resource(pe_working_set_t * data_set, resource_t *rsc) { int delay = 0; int max_delay = 0; if(rsc && rsc->children) { GList *iter = NULL; for(iter = rsc->children; iter; iter = iter->next) { resource_t *child = (resource_t *)iter->data; delay = max_delay_for_resource(data_set, child); if(delay > max_delay) { double seconds = delay / 1000.0; crm_trace("Calculated new delay of %.1fs due to %s", seconds, child->id); max_delay = delay; } } } else if(rsc) { char *key = crm_strdup_printf("%s_%s_0", rsc->id, RSC_STOP); action_t *stop = custom_action(rsc, key, RSC_STOP, NULL, TRUE, FALSE, data_set); const char *value = g_hash_table_lookup(stop->meta, XML_ATTR_TIMEOUT); max_delay = crm_int_helper(value, NULL); pe_free_action(stop); } return max_delay; } static int max_delay_in(pe_working_set_t * data_set, GList *resources) { int max_delay = 0; GList *item = NULL; for (item = resources; item != NULL; item = item->next) { int delay = 0; resource_t *rsc = pe_find_resource(data_set->resources, (const char *)item->data); delay = max_delay_for_resource(data_set, rsc); if(delay > max_delay) { double seconds = delay / 1000.0; crm_trace("Calculated new delay of %.1fs due to %s", seconds, rsc->id); max_delay = delay; } } return 5 + (max_delay / 1000); } #define waiting_for_starts(d, r, h) ((g_list_length(d) > 0) || \ (resource_is_running_on((r), (h)) == FALSE)) /*! * \internal * \brief Restart a resource (on a particular host if requested). * * \param[in] rsc The resource to restart * \param[in] host The host to restart the resource on (or NULL for all) * \param[in] timeout_ms Consider failed if actions do not complete in this time * (specified in milliseconds, but a two-second * granularity is actually used; if 0, a timeout will be * calculated based on the resource timeout) * \param[in] cib Connection to the CIB manager * * \return pcmk_ok on success, -errno on failure (exits on certain failures) */ int cli_resource_restart(pe_resource_t *rsc, const char *host, int timeout_ms, cib_t *cib) { int rc = 0; int lpc = 0; int before = 0; int step_timeout_s = 0; int sleep_interval = 2; int timeout = timeout_ms / 1000; bool stop_via_ban = FALSE; char *rsc_id = NULL; char *orig_target_role = NULL; GList *list_delta = NULL; GList *target_active = NULL; GList *current_active = NULL; GList *restart_target_active = NULL; pe_working_set_t *data_set = NULL; if(resource_is_running_on(rsc, host) == FALSE) { const char *id = rsc->clone_name?rsc->clone_name:rsc->id; if(host) { printf("%s is not running on %s and so cannot be restarted\n", id, host); } else { printf("%s is not running anywhere and so cannot be restarted\n", id); } return -ENXIO; } /* We might set the target-role meta-attribute */ attr_set_type = XML_TAG_META_SETS; rsc_id = strdup(rsc->id); if ((pe_rsc_is_clone(rsc) || pe_bundle_replicas(rsc)) && host) { stop_via_ban = TRUE; } /* grab full cib determine originally active resources disable or ban poll cib and watch for affected resources to get stopped without --timeout, calculate the stop timeout for each step and wait for that if we hit --timeout or the service timeout, re-enable or un-ban, report failure and indicate which resources we couldn't take down if everything stopped, re-enable or un-ban poll cib and watch for affected resources to get started without --timeout, calculate the start timeout for each step and wait for that if we hit --timeout or the service timeout, report (different) failure and indicate which resources we couldn't bring back up report success Optimizations: - use constraints to determine ordered list of affected resources - Allow a --no-deps option (aka. --force-restart) */ data_set = pe_new_working_set(); if (data_set == NULL) { crm_perror(LOG_ERR, "Could not allocate working set"); rc = -ENOMEM; goto done; } + set_bit(data_set->flags, pe_flag_no_counts); rc = update_dataset(cib, data_set, FALSE); if(rc != pcmk_ok) { fprintf(stdout, "Could not get new resource list: %s (%d)\n", pcmk_strerror(rc), rc); goto done; } restart_target_active = get_active_resources(host, data_set->resources); current_active = get_active_resources(host, data_set->resources); dump_list(current_active, "Origin"); if (stop_via_ban) { /* Stop the clone or bundle instance by banning it from the host */ BE_QUIET = TRUE; rc = cli_resource_ban(rsc_id, host, NULL, cib); } else { /* Stop the resource by setting target-role to Stopped. * Remember any existing target-role so we can restore it later * (though it only makes any difference if it's Slave). */ char *lookup_id = clone_strip(rsc->id); find_resource_attr(cib, XML_NVPAIR_ATTR_VALUE, lookup_id, NULL, NULL, NULL, XML_RSC_ATTR_TARGET_ROLE, &orig_target_role); free(lookup_id); rc = cli_resource_update_attribute(rsc, rsc_id, NULL, NULL, XML_RSC_ATTR_TARGET_ROLE, RSC_STOPPED, FALSE, cib, data_set); } if(rc != pcmk_ok) { fprintf(stderr, "Could not set target-role for %s: %s (%d)\n", rsc_id, pcmk_strerror(rc), rc); if (current_active) { g_list_free_full(current_active, free); } if (restart_target_active) { g_list_free_full(restart_target_active, free); } goto done; } rc = update_dataset(cib, data_set, TRUE); if(rc != pcmk_ok) { fprintf(stderr, "Could not determine which resources would be stopped\n"); goto failure; } target_active = get_active_resources(host, data_set->resources); dump_list(target_active, "Target"); list_delta = subtract_lists(current_active, target_active, (GCompareFunc) strcmp); fprintf(stdout, "Waiting for %d resources to stop:\n", g_list_length(list_delta)); display_list(list_delta, " * "); step_timeout_s = timeout / sleep_interval; while(g_list_length(list_delta) > 0) { before = g_list_length(list_delta); if(timeout_ms == 0) { step_timeout_s = max_delay_in(data_set, list_delta) / sleep_interval; } /* We probably don't need the entire step timeout */ for(lpc = 0; lpc < step_timeout_s && g_list_length(list_delta) > 0; lpc++) { sleep(sleep_interval); if(timeout) { timeout -= sleep_interval; crm_trace("%ds remaining", timeout); } rc = update_dataset(cib, data_set, FALSE); if(rc != pcmk_ok) { fprintf(stderr, "Could not determine which resources were stopped\n"); goto failure; } if (current_active) { g_list_free_full(current_active, free); } current_active = get_active_resources(host, data_set->resources); g_list_free(list_delta); list_delta = subtract_lists(current_active, target_active, (GCompareFunc) strcmp); dump_list(current_active, "Current"); dump_list(list_delta, "Delta"); } crm_trace("%d (was %d) resources remaining", g_list_length(list_delta), before); if(before == g_list_length(list_delta)) { /* aborted during stop phase, print the contents of list_delta */ fprintf(stderr, "Could not complete shutdown of %s, %d resources remaining\n", rsc_id, g_list_length(list_delta)); display_list(list_delta, " * "); rc = -ETIME; goto failure; } } if (stop_via_ban) { rc = cli_resource_clear(rsc_id, host, NULL, cib, TRUE); } else if (orig_target_role) { rc = cli_resource_update_attribute(rsc, rsc_id, NULL, NULL, XML_RSC_ATTR_TARGET_ROLE, orig_target_role, FALSE, cib, data_set); free(orig_target_role); orig_target_role = NULL; } else { rc = cli_resource_delete_attribute(rsc, rsc_id, NULL, NULL, XML_RSC_ATTR_TARGET_ROLE, cib, data_set); } if(rc != pcmk_ok) { fprintf(stderr, "Could not unset target-role for %s: %s (%d)\n", rsc_id, pcmk_strerror(rc), rc); goto done; } if (target_active) { g_list_free_full(target_active, free); } target_active = restart_target_active; if (list_delta) { g_list_free(list_delta); } list_delta = subtract_lists(target_active, current_active, (GCompareFunc) strcmp); fprintf(stdout, "Waiting for %d resources to start again:\n", g_list_length(list_delta)); display_list(list_delta, " * "); step_timeout_s = timeout / sleep_interval; while (waiting_for_starts(list_delta, rsc, host)) { before = g_list_length(list_delta); if(timeout_ms == 0) { step_timeout_s = max_delay_in(data_set, list_delta) / sleep_interval; } /* We probably don't need the entire step timeout */ for (lpc = 0; (lpc < step_timeout_s) && waiting_for_starts(list_delta, rsc, host); lpc++) { sleep(sleep_interval); if(timeout) { timeout -= sleep_interval; crm_trace("%ds remaining", timeout); } rc = update_dataset(cib, data_set, FALSE); if(rc != pcmk_ok) { fprintf(stderr, "Could not determine which resources were started\n"); goto failure; } if (current_active) { g_list_free_full(current_active, free); } /* It's OK if dependent resources moved to a different node, * so we check active resources on all nodes. */ current_active = get_active_resources(NULL, data_set->resources); g_list_free(list_delta); list_delta = subtract_lists(target_active, current_active, (GCompareFunc) strcmp); dump_list(current_active, "Current"); dump_list(list_delta, "Delta"); } if(before == g_list_length(list_delta)) { /* aborted during start phase, print the contents of list_delta */ fprintf(stdout, "Could not complete restart of %s, %d resources remaining\n", rsc_id, g_list_length(list_delta)); display_list(list_delta, " * "); rc = -ETIME; goto failure; } } rc = pcmk_ok; goto done; failure: if (stop_via_ban) { cli_resource_clear(rsc_id, host, NULL, cib, TRUE); } else if (orig_target_role) { cli_resource_update_attribute(rsc, rsc_id, NULL, NULL, XML_RSC_ATTR_TARGET_ROLE, orig_target_role, FALSE, cib, data_set); free(orig_target_role); } else { cli_resource_delete_attribute(rsc, rsc_id, NULL, NULL, XML_RSC_ATTR_TARGET_ROLE, cib, data_set); } done: if (list_delta) { g_list_free(list_delta); } if (current_active) { g_list_free_full(current_active, free); } if (target_active && (target_active != restart_target_active)) { g_list_free_full(target_active, free); } if (restart_target_active) { g_list_free_full(restart_target_active, free); } free(rsc_id); pe_free_working_set(data_set); return rc; } static inline int action_is_pending(action_t *action) { if(is_set(action->flags, pe_action_optional)) { return FALSE; } else if(is_set(action->flags, pe_action_runnable) == FALSE) { return FALSE; } else if(is_set(action->flags, pe_action_pseudo)) { return FALSE; } else if(safe_str_eq("notify", action->task)) { return FALSE; } return TRUE; } /*! * \internal * \brief Return TRUE if any actions in a list are pending * * \param[in] actions List of actions to check * * \return TRUE if any actions in the list are pending, FALSE otherwise */ static bool actions_are_pending(GListPtr actions) { GListPtr action; for (action = actions; action != NULL; action = action->next) { action_t *a = (action_t *)action->data; if (action_is_pending(a)) { crm_notice("Waiting for %s (flags=0x%.8x)", a->uuid, a->flags); return TRUE; } } return FALSE; } /*! * \internal * \brief Print pending actions to stderr * * \param[in] actions List of actions to check * * \return void */ static void print_pending_actions(GListPtr actions) { GListPtr action; fprintf(stderr, "Pending actions:\n"); for (action = actions; action != NULL; action = action->next) { action_t *a = (action_t *) action->data; if (action_is_pending(a)) { fprintf(stderr, "\tAction %d: %s", a->id, a->uuid); if (a->node) { fprintf(stderr, "\ton %s", a->node->details->uname); } fprintf(stderr, "\n"); } } } /* For --wait, timeout (in seconds) to use if caller doesn't specify one */ #define WAIT_DEFAULT_TIMEOUT_S (60 * 60) /* For --wait, how long to sleep between cluster state checks */ #define WAIT_SLEEP_S (2) /*! * \internal * \brief Wait until all pending cluster actions are complete * * This waits until either the CIB's transition graph is idle or a timeout is * reached. * * \param[in] timeout_ms Consider failed if actions do not complete in this time * (specified in milliseconds, but one-second granularity * is actually used; if 0, a default will be used) * \param[in] cib Connection to the CIB manager * * \return pcmk_ok on success, -errno on failure */ int wait_till_stable(int timeout_ms, cib_t * cib) { pe_working_set_t *data_set = NULL; int rc = -1; int timeout_s = timeout_ms? ((timeout_ms + 999) / 1000) : WAIT_DEFAULT_TIMEOUT_S; time_t expire_time = time(NULL) + timeout_s; time_t time_diff; bool printed_version_warning = BE_QUIET; // i.e. don't print if quiet data_set = pe_new_working_set(); if (data_set == NULL) { return -ENOMEM; } + set_bit(data_set->flags, pe_flag_no_counts); do { /* Abort if timeout is reached */ time_diff = expire_time - time(NULL); if (time_diff > 0) { crm_info("Waiting up to %ld seconds for cluster actions to complete", time_diff); } else { print_pending_actions(data_set->actions); pe_free_working_set(data_set); return -ETIME; } if (rc == pcmk_ok) { /* this avoids sleep on first loop iteration */ sleep(WAIT_SLEEP_S); } /* Get latest transition graph */ pe_reset_working_set(data_set); rc = update_working_set_from_cib(data_set, cib); if (rc != pcmk_ok) { pe_free_working_set(data_set); return rc; } pcmk__schedule_actions(data_set, data_set->input, NULL); if (!printed_version_warning) { /* If the DC has a different version than the local node, the two * could come to different conclusions about what actions need to be * done. Warn the user in this case. * * @TODO A possible long-term solution would be to reimplement the * wait as a new controller operation that would be forwarded to the * DC. However, that would have potential problems of its own. */ const char *dc_version = g_hash_table_lookup(data_set->config_hash, "dc-version"); if (safe_str_neq(dc_version, PACEMAKER_VERSION "-" BUILD_VERSION)) { printf("warning: wait option may not work properly in " "mixed-version cluster\n"); printed_version_warning = TRUE; } } } while (actions_are_pending(data_set->actions)); pe_free_working_set(data_set); return pcmk_ok; } int cli_resource_execute_from_params(const char *rsc_name, const char *rsc_class, const char *rsc_prov, const char *rsc_type, const char *action, GHashTable *params, GHashTable *override_hash, int timeout_ms) { GHashTable *params_copy = NULL; int rc = pcmk_ok; svc_action_t *op = NULL; if (safe_str_eq(rsc_class, PCMK_RESOURCE_CLASS_STONITH)) { CMD_ERR("Sorry, the %s option doesn't support %s resources yet", action, rsc_class); crm_exit(CRM_EX_UNIMPLEMENT_FEATURE); } /* If no timeout was provided, grab the default. */ if (timeout_ms == 0) { timeout_ms = crm_get_msec(CRM_DEFAULT_OP_TIMEOUT_S); } /* add meta_timeout env needed by some resource agents */ g_hash_table_insert(params, strdup("CRM_meta_timeout"), crm_strdup_printf("%d", timeout_ms)); /* add crm_feature_set env needed by some resource agents */ g_hash_table_insert(params, strdup(XML_ATTR_CRM_VERSION), strdup(CRM_FEATURE_SET)); /* resources_action_create frees the params hash table it's passed, but we * may need to reuse it in a second call to resources_action_create. Thus * we'll make a copy here so that gets freed and the original remains for * reuse. */ params_copy = crm_str_table_dup(params); op = resources_action_create(rsc_name, rsc_class, rsc_prov, rsc_type, action, 0, timeout_ms, params_copy, 0); if (op == NULL) { /* Re-run with stderr enabled so we can display a sane error message */ crm_enable_stderr(TRUE); params_copy = crm_str_table_dup(params); op = resources_action_create(rsc_name, rsc_class, rsc_prov, rsc_type, action, 0, timeout_ms, params_copy, 0); /* Callers of cli_resource_execute expect that the params hash table will * be freed. That function uses this one, so for that reason and for * making the two act the same, we should free the hash table here too. */ g_hash_table_destroy(params); /* We know op will be NULL, but this makes static analysis happy */ services_action_free(op); crm_exit(CRM_EX_DATAERR); } setenv("HA_debug", resource_verbose > 0 ? "1" : "0", 1); if(resource_verbose > 1) { setenv("OCF_TRACE_RA", "1", 1); } /* A resource agent using the standard ocf-shellfuncs library will not print * messages to stderr if it doesn't have a controlling terminal (e.g. if * crm_resource is called via script or ssh). This forces it to do so. */ setenv("OCF_TRACE_FILE", "/dev/stderr", 0); if (override_hash) { GHashTableIter iter; char *name = NULL; char *value = NULL; g_hash_table_iter_init(&iter, override_hash); while (g_hash_table_iter_next(&iter, (gpointer *) & name, (gpointer *) & value)) { printf("Overriding the cluster configuration for '%s' with '%s' = '%s'\n", rsc_name, name, value); g_hash_table_replace(op->params, strdup(name), strdup(value)); } } if (services_action_sync(op)) { int more, lpc, last; char *local_copy = NULL; rc = op->rc; if (op->status == PCMK_LRM_OP_DONE) { printf("Operation %s for %s (%s:%s:%s) returned: '%s' (%d)\n", action, rsc_name, rsc_class, rsc_prov ? rsc_prov : "", rsc_type, services_ocf_exitcode_str(op->rc), op->rc); } else { printf("Operation %s for %s (%s:%s:%s) failed: '%s' (%d)\n", action, rsc_name, rsc_class, rsc_prov ? rsc_prov : "", rsc_type, services_lrm_status_str(op->status), op->status); } /* hide output for validate-all if not in verbose */ if (resource_verbose == 0 && safe_str_eq(action, "validate-all")) goto done; if (op->stdout_data) { local_copy = strdup(op->stdout_data); more = strlen(local_copy); last = 0; for (lpc = 0; lpc < more; lpc++) { if (local_copy[lpc] == '\n' || local_copy[lpc] == 0) { local_copy[lpc] = 0; printf(" > stdout: %s\n", local_copy + last); last = lpc + 1; } } free(local_copy); } if (op->stderr_data) { local_copy = strdup(op->stderr_data); more = strlen(local_copy); last = 0; for (lpc = 0; lpc < more; lpc++) { if (local_copy[lpc] == '\n' || local_copy[lpc] == 0) { local_copy[lpc] = 0; printf(" > stderr: %s\n", local_copy + last); last = lpc + 1; } } free(local_copy); } } else { rc = op->rc == 0 ? pcmk_err_generic : op->rc; } done: services_action_free(op); /* See comment above about why we free params here. */ g_hash_table_destroy(params); return rc; } int cli_resource_execute(resource_t *rsc, const char *requested_name, const char *rsc_action, GHashTable *override_hash, int timeout_ms, cib_t * cib, pe_working_set_t *data_set) { int rc = pcmk_ok; const char *rid = NULL; const char *rtype = NULL; const char *rprov = NULL; const char *rclass = NULL; const char *action = NULL; GHashTable *params = NULL; if (safe_str_eq(rsc_action, "validate")) { action = "validate-all"; } else if (safe_str_eq(rsc_action, "force-check")) { action = "monitor"; } else if (safe_str_eq(rsc_action, "force-stop")) { action = rsc_action+6; } else if (safe_str_eq(rsc_action, "force-start") || safe_str_eq(rsc_action, "force-demote") || safe_str_eq(rsc_action, "force-promote")) { action = rsc_action+6; if(pe_rsc_is_clone(rsc)) { rc = cli_resource_search(rsc, requested_name, data_set); if(rc > 0 && do_force == FALSE) { CMD_ERR("It is not safe to %s %s here: the cluster claims it is already active", action, rsc->id); CMD_ERR("Try setting target-role=Stopped first or specifying " "the force option"); crm_exit(CRM_EX_UNSAFE); } } } else { action = rsc_action; } if(pe_rsc_is_clone(rsc)) { /* Grab the first child resource in the hope it's not a group */ rsc = rsc->children->data; } if(rsc->variant == pe_group) { CMD_ERR("Sorry, the %s option doesn't support group resources", rsc_action); crm_exit(CRM_EX_UNIMPLEMENT_FEATURE); } rclass = crm_element_value(rsc->xml, XML_AGENT_ATTR_CLASS); rprov = crm_element_value(rsc->xml, XML_AGENT_ATTR_PROVIDER); rtype = crm_element_value(rsc->xml, XML_ATTR_TYPE); params = generate_resource_params(rsc, data_set); if (timeout_ms == 0) { timeout_ms = pe_get_configured_timeout(rsc, action, data_set); } rid = pe_rsc_is_anon_clone(rsc->parent)? requested_name : rsc->id; rc = cli_resource_execute_from_params(rid, rclass, rprov, rtype, action, params, override_hash, timeout_ms); return rc; } int cli_resource_move(resource_t *rsc, const char *rsc_id, const char *host_name, cib_t *cib, pe_working_set_t *data_set) { int rc = pcmk_ok; unsigned int count = 0; node_t *current = NULL; node_t *dest = pe_find_node(data_set->nodes, host_name); bool cur_is_dest = FALSE; if (dest == NULL) { return -pcmk_err_node_unknown; } if (scope_master && is_not_set(rsc->flags, pe_rsc_promotable)) { resource_t *p = uber_parent(rsc); if (is_set(p->flags, pe_rsc_promotable)) { CMD_ERR("Using parent '%s' for move instead of '%s'.", rsc->id, rsc_id); rsc_id = p->id; rsc = p; } else { CMD_ERR("Ignoring master option: %s is not promotable", rsc_id); scope_master = FALSE; } } current = pe__find_active_requires(rsc, &count); if (is_set(rsc->flags, pe_rsc_promotable)) { GListPtr iter = NULL; unsigned int master_count = 0; pe_node_t *master_node = NULL; for(iter = rsc->children; iter; iter = iter->next) { resource_t *child = (resource_t *)iter->data; enum rsc_role_e child_role = child->fns->state(child, TRUE); if(child_role == RSC_ROLE_MASTER) { rsc = child; master_node = pe__current_node(child); master_count++; } } if (scope_master || master_count) { count = master_count; current = master_node; } } if (count > 1) { if (pe_rsc_is_clone(rsc)) { current = NULL; } else { return -pcmk_err_multiple; } } if (current && (current->details == dest->details)) { cur_is_dest = TRUE; if (do_force) { crm_info("%s is already %s on %s, reinforcing placement with location constraint.", rsc_id, scope_master?"promoted":"active", dest->details->uname); } else { return -pcmk_err_already; } } /* Clear any previous prefer constraints across all nodes. */ cli_resource_clear(rsc_id, NULL, data_set->nodes, cib, FALSE); /* Clear any previous ban constraints on 'dest'. */ cli_resource_clear(rsc_id, dest->details->uname, data_set->nodes, cib, TRUE); /* Record an explicit preference for 'dest' */ rc = cli_resource_prefer(rsc_id, dest->details->uname, cib); crm_trace("%s%s now prefers node %s%s", rsc->id, scope_master?" (master)":"", dest->details->uname, do_force?"(forced)":""); /* only ban the previous location if current location != destination location. * it is possible to use -M to enforce a location without regard of where the * resource is currently located */ if(do_force && (cur_is_dest == FALSE)) { /* Ban the original location if possible */ if(current) { (void)cli_resource_ban(rsc_id, current->details->uname, NULL, cib); } else if(count > 1) { CMD_ERR("Resource '%s' is currently %s in %d locations. " "One may now move to %s", rsc_id, (scope_master? "promoted" : "active"), count, dest->details->uname); CMD_ERR("To prevent '%s' from being %s at a specific location, " "specify a node.", rsc_id, (scope_master? "promoted" : "active")); } else { crm_trace("Not banning %s from its current location: not active", rsc_id); } } return rc; } static void cli_resource_why_without_rsc_and_host(cib_t *cib_conn,GListPtr resources) { GListPtr lpc = NULL; GListPtr hosts = NULL; for (lpc = resources; lpc != NULL; lpc = lpc->next) { resource_t *rsc = (resource_t *) lpc->data; rsc->fns->location(rsc, &hosts, TRUE); if (hosts == NULL) { printf("Resource %s is not running\n", rsc->id); } else { printf("Resource %s is running\n", rsc->id); } cli_resource_check(cib_conn, rsc); g_list_free(hosts); hosts = NULL; } } static void cli_resource_why_with_rsc_and_host(cib_t *cib_conn, GListPtr resources, resource_t *rsc, const char *host_uname) { if (resource_is_running_on(rsc, host_uname)) { printf("Resource %s is running on host %s\n",rsc->id,host_uname); } else { printf("Resource %s is not running on host %s\n", rsc->id, host_uname); } cli_resource_check(cib_conn, rsc); } static void cli_resource_why_without_rsc_with_host(cib_t *cib_conn,GListPtr resources,node_t *node) { const char* host_uname = node->details->uname; GListPtr allResources = node->details->allocated_rsc; GListPtr activeResources = node->details->running_rsc; GListPtr unactiveResources = subtract_lists(allResources,activeResources,(GCompareFunc) strcmp); GListPtr lpc = NULL; for (lpc = activeResources; lpc != NULL; lpc = lpc->next) { resource_t *rsc = (resource_t *) lpc->data; printf("Resource %s is running on host %s\n",rsc->id,host_uname); cli_resource_check(cib_conn,rsc); } for(lpc = unactiveResources; lpc != NULL; lpc = lpc->next) { resource_t *rsc = (resource_t *) lpc->data; printf("Resource %s is assigned to host %s but not running\n", rsc->id, host_uname); cli_resource_check(cib_conn,rsc); } g_list_free(allResources); g_list_free(activeResources); g_list_free(unactiveResources); } static void cli_resource_why_with_rsc_without_host(cib_t *cib_conn, GListPtr resources, resource_t *rsc) { GListPtr hosts = NULL; rsc->fns->location(rsc, &hosts, TRUE); printf("Resource %s is %srunning\n", rsc->id, (hosts? "" : "not ")); cli_resource_check(cib_conn, rsc); g_list_free(hosts); } void cli_resource_why(cib_t *cib_conn, GListPtr resources, resource_t *rsc, node_t *node) { const char *host_uname = (node == NULL)? NULL : node->details->uname; if ((rsc == NULL) && (host_uname == NULL)) { cli_resource_why_without_rsc_and_host(cib_conn, resources); } else if ((rsc != NULL) && (host_uname != NULL)) { cli_resource_why_with_rsc_and_host(cib_conn, resources, rsc, host_uname); } else if ((rsc == NULL) && (host_uname != NULL)) { cli_resource_why_without_rsc_with_host(cib_conn, resources, node); } else if ((rsc != NULL) && (host_uname == NULL)) { cli_resource_why_with_rsc_without_host(cib_conn, resources, rsc); } } diff --git a/tools/crm_rule.c b/tools/crm_rule.c index 0ff20cdade..5cd46c00ca 100644 --- a/tools/crm_rule.c +++ b/tools/crm_rule.c @@ -1,280 +1,281 @@ /* * Copyright 2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include enum crm_rule_mode { crm_rule_mode_none, crm_rule_mode_check } rule_mode = crm_rule_mode_none; static int crm_rule_check(pe_working_set_t *data_set, const char *rule_id, crm_time_t *effective_date); static struct crm_option long_options[] = { /* Top-level Options */ {"help", no_argument, NULL, '?', "\tThis text"}, {"version", no_argument, NULL, '$', "\tVersion information" }, {"verbose", no_argument, NULL, 'V', "\tIncrease debug output"}, {"-spacer-", required_argument, NULL, '-', "\nModes (mutually exclusive):" }, {"check", no_argument, NULL, 'c', "\tCheck whether a rule is in effect" }, {"-spacer-", required_argument, NULL, '-', "\nAdditional options:" }, {"date", required_argument, NULL, 'd', "Whether the rule is in effect on a given date" }, {"rule", required_argument, NULL, 'r', "The ID of the rule to check" }, {"-spacer-", no_argument, NULL, '-', "\nData:"}, {"xml-text", required_argument, NULL, 'X', "Use argument for XML (or stdin if '-')"}, {"-spacer-", required_argument, NULL, '-', "\n\nThis tool is currently experimental.", pcmk_option_paragraph}, {"-spacer-", required_argument, NULL, '-', "The interface, behavior, and output may " "change with any version of pacemaker.", pcmk_option_paragraph}, {0, 0, 0, 0} }; static int crm_rule_check(pe_working_set_t *data_set, const char *rule_id, crm_time_t *effective_date) { xmlNode *cib_constraints = NULL; xmlNode *match = NULL; xmlXPathObjectPtr xpathObj = NULL; pe_eval_date_result_t result; char *xpath = NULL; int rc = pcmk_ok; int max = 0; /* Rules are under the constraints node in the XML, so first find that. */ cib_constraints = get_object_root(XML_CIB_TAG_CONSTRAINTS, data_set->input); /* Get all rules matching the given ID, which also have only a single date_expression * child whose operation is not 'date_spec'. This is fairly limited, but it's hard * to check expressions more complicated than that. */ xpath = crm_strdup_printf("//rule[@id='%s']/date_expression[@operation!='date_spec']", rule_id); xpathObj = xpath_search(cib_constraints, xpath); max = numXpathResults(xpathObj); if (max == 0) { CMD_ERR("No rule found with ID=%s containing a date_expression", rule_id); rc = -ENXIO; goto bail; } else if (max > 1) { CMD_ERR("More than one date_expression in %s is not supported", rule_id); rc = -EOPNOTSUPP; goto bail; } match = getXpathResult(xpathObj, 0); /* We should have ensured both of these pass with the xpath query above, but * double checking can't hurt. */ CRM_ASSERT(match != NULL); CRM_ASSERT(find_expression_type(match) == time_expr); result = pe_eval_date_expression(match, effective_date, NULL); if (result == pe_date_within_range) { printf("Rule %s is still in effect\n", rule_id); rc = 0; } else if (result == pe_date_after_range) { printf("Rule %s is expired\n", rule_id); rc = 1; } else if (result == pe_date_before_range) { printf("Rule %s has not yet taken effect\n", rule_id); rc = 2; } else { printf("Could not determine whether rule %s is expired\n", rule_id); rc = 3; } bail: free(xpath); freeXpathObject(xpathObj); return rc; } int main(int argc, char **argv) { cib_t *cib_conn = NULL; pe_working_set_t *data_set = NULL; int flag = 0; int option_index = 0; char *rule_id = NULL; crm_time_t *rule_date = NULL; xmlNode *input = NULL; char *input_xml = NULL; int rc = pcmk_ok; crm_exit_t exit_code = CRM_EX_OK; crm_log_cli_init("crm_rule"); crm_set_options(NULL, "[options]", long_options, "Tool for querying the state of rules"); while (flag >= 0) { flag = crm_get_option(argc, argv, &option_index); switch (flag) { case -1: break; case 'V': crm_bump_log_level(argc, argv); break; case '$': case '?': crm_help(flag, CRM_EX_OK); break; case 'c': rule_mode = crm_rule_mode_check; break; case 'd': rule_date = crm_time_new(optarg); if (rule_date == NULL) { exit_code = CRM_EX_DATAERR; goto bail; } break; case 'X': input_xml = optarg; break; case 'r': rule_id = strdup(optarg); break; default: crm_help(flag, CRM_EX_OK); break; } } /* Check command line arguments before opening a connection to * the CIB manager or doing anything else important. */ if (rule_mode == crm_rule_mode_check) { if (rule_id == NULL) { CMD_ERR("--check requires use of --rule=\n"); crm_help(flag, CRM_EX_USAGE); } } /* Set up some defaults. */ if (rule_date == NULL) { rule_date = crm_time_new(NULL); } /* Where does the XML come from? If one of various command line options were * given, use those. Otherwise, connect to the CIB and use that. */ if (safe_str_eq(input_xml, "-")) { input = stdin2xml(); if (input == NULL) { fprintf(stderr, "Couldn't parse input from STDIN\n"); exit_code = CRM_EX_DATAERR; goto bail; } } else if (input_xml != NULL) { input = string2xml(input_xml); if (input == NULL) { fprintf(stderr, "Couldn't parse input string: %s\n", input_xml); exit_code = CRM_EX_DATAERR; goto bail; } } else { /* Establish a connection to the CIB manager */ cib_conn = cib_new(); rc = cib_conn->cmds->signon(cib_conn, crm_system_name, cib_command); if (rc != pcmk_ok) { CMD_ERR("Error connecting to the CIB manager: %s", pcmk_strerror(rc)); exit_code = crm_errno2exit(rc); goto bail; } } /* Populate working set from CIB query */ if (input == NULL) { rc = cib_conn->cmds->query(cib_conn, NULL, &input, cib_scope_local | cib_sync_call); if (rc != pcmk_ok) { exit_code = crm_errno2exit(rc); goto bail; } } /* Populate the working set instance */ data_set = pe_new_working_set(); if (data_set == NULL) { exit_code = crm_errno2exit(ENOMEM); goto bail; } + set_bit(data_set->flags, pe_flag_no_counts); data_set->input = input; data_set->now = rule_date; /* Unpack everything. */ cluster_status(data_set); /* Now do whichever operation mode was asked for. There's only one at the * moment so this looks a little silly, but I expect there will be more * modes in the future. */ switch(rule_mode) { case crm_rule_mode_check: rc = crm_rule_check(data_set, rule_id, rule_date); if (rc < 0) { CMD_ERR("Error checking rule: %s", pcmk_strerror(rc)); exit_code = crm_errno2exit(rc); } else if (rc == 1) { exit_code = CRM_EX_EXPIRED; } else if (rc == 2) { exit_code = CRM_EX_NOT_YET_IN_EFFECT; } else if (rc == 3) { exit_code = CRM_EX_INDETERMINATE; } else { exit_code = rc; } break; default: break; } bail: if (cib_conn != NULL) { cib_conn->cmds->signoff(cib_conn); cib_delete(cib_conn); } pe_free_working_set(data_set); crm_exit(exit_code); } diff --git a/tools/crm_simulate.c b/tools/crm_simulate.c index acb5029d36..73029f0f35 100644 --- a/tools/crm_simulate.c +++ b/tools/crm_simulate.c @@ -1,956 +1,936 @@ /* * Copyright 2009-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include cib_t *global_cib = NULL; GListPtr op_fail = NULL; bool action_numbers = FALSE; gboolean quiet = FALSE; gboolean print_pending = TRUE; char *temp_shadow = NULL; extern gboolean bringing_nodes_online; #define quiet_log(fmt, args...) do { \ if(quiet == FALSE) { \ printf(fmt , ##args); \ } \ } while(0) char *use_date = NULL; static void get_date(pe_working_set_t *data_set, bool print_original) { time_t original_date = 0; crm_element_value_epoch(data_set->input, "execution-date", &original_date); if (use_date) { data_set->now = crm_time_new(use_date); quiet_log(" + Setting effective cluster time: %s", use_date); crm_time_log(LOG_NOTICE, "Pretending 'now' is", data_set->now, crm_time_log_date | crm_time_log_timeofday); } else if (original_date) { data_set->now = crm_time_new(NULL); crm_time_set_timet(data_set->now, &original_date); if (print_original) { char *when = crm_time_as_string(data_set->now, crm_time_log_date|crm_time_log_timeofday); printf("Using the original execution date of: %s\n", when); free(when); } } } static void print_cluster_status(pe_working_set_t * data_set, long options) { char *online_nodes = NULL; char *online_remote_nodes = NULL; char *online_guest_nodes = NULL; char *offline_nodes = NULL; char *offline_remote_nodes = NULL; GListPtr gIter = NULL; for (gIter = data_set->nodes; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; const char *node_mode = NULL; char *node_name = NULL; if (pe__is_guest_node(node)) { node_name = crm_strdup_printf("%s:%s", node->details->uname, node->details->remote_rsc->container->id); } else { node_name = crm_strdup_printf("%s", node->details->uname); } if (node->details->unclean) { if (node->details->online && node->details->unclean) { node_mode = "UNCLEAN (online)"; } else if (node->details->pending) { node_mode = "UNCLEAN (pending)"; } else { node_mode = "UNCLEAN (offline)"; } } else if (node->details->pending) { node_mode = "pending"; } else if (node->details->standby_onfail && node->details->online) { node_mode = "standby (on-fail)"; } else if (node->details->standby) { if (node->details->online) { node_mode = "standby"; } else { node_mode = "OFFLINE (standby)"; } } else if (node->details->maintenance) { if (node->details->online) { node_mode = "maintenance"; } else { node_mode = "OFFLINE (maintenance)"; } } else if (node->details->online) { if (pe__is_guest_node(node)) { online_guest_nodes = add_list_element(online_guest_nodes, node_name); } else if (pe__is_remote_node(node)) { online_remote_nodes = add_list_element(online_remote_nodes, node_name); } else { online_nodes = add_list_element(online_nodes, node_name); } free(node_name); continue; } else { if (pe__is_remote_node(node)) { offline_remote_nodes = add_list_element(offline_remote_nodes, node_name); } else if (pe__is_guest_node(node)) { /* ignore offline container nodes */ } else { offline_nodes = add_list_element(offline_nodes, node_name); } free(node_name); continue; } if (pe__is_guest_node(node)) { printf("GuestNode %s: %s\n", node_name, node_mode); } else if (pe__is_remote_node(node)) { printf("RemoteNode %s: %s\n", node_name, node_mode); } else if (safe_str_eq(node->details->uname, node->details->id)) { printf("Node %s: %s\n", node_name, node_mode); } else { printf("Node %s (%s): %s\n", node_name, node->details->id, node_mode); } free(node_name); } if (online_nodes) { printf("Online: [%s ]\n", online_nodes); free(online_nodes); } if (offline_nodes) { printf("OFFLINE: [%s ]\n", offline_nodes); free(offline_nodes); } if (online_remote_nodes) { printf("RemoteOnline: [%s ]\n", online_remote_nodes); free(online_remote_nodes); } if (offline_remote_nodes) { printf("RemoteOFFLINE: [%s ]\n", offline_remote_nodes); free(offline_remote_nodes); } if (online_guest_nodes) { printf("GuestOnline: [%s ]\n", online_guest_nodes); free(online_guest_nodes); } fprintf(stdout, "\n"); for (gIter = data_set->resources; gIter != NULL; gIter = gIter->next) { resource_t *rsc = (resource_t *) gIter->data; if (is_set(rsc->flags, pe_rsc_orphan) && rsc->role == RSC_ROLE_STOPPED) { continue; } rsc->fns->print(rsc, NULL, pe_print_printf | options, stdout); } fprintf(stdout, "\n"); } static char * create_action_name(pe_action_t *action) { char *action_name = NULL; const char *prefix = ""; const char *action_host = NULL; const char *clone_name = NULL; const char *task = action->task; if (action->node) { action_host = action->node->details->uname; } else if (is_not_set(action->flags, pe_action_pseudo)) { action_host = ""; } if (safe_str_eq(action->task, RSC_CANCEL)) { prefix = "Cancel "; task = action->cancel_task; } if (action->rsc && action->rsc->clone_name) { clone_name = action->rsc->clone_name; } if (clone_name) { char *key = NULL; const char *interval_ms_s = NULL; guint interval_ms = 0; interval_ms_s = g_hash_table_lookup(action->meta, XML_LRM_ATTR_INTERVAL_MS); interval_ms = crm_parse_ms(interval_ms_s); if (safe_str_eq(action->task, RSC_NOTIFY) || safe_str_eq(action->task, RSC_NOTIFIED)) { const char *n_type = g_hash_table_lookup(action->meta, "notify_key_type"); const char *n_task = g_hash_table_lookup(action->meta, "notify_key_operation"); CRM_ASSERT(n_type != NULL); CRM_ASSERT(n_task != NULL); key = generate_notify_key(clone_name, n_type, n_task); } else { key = generate_op_key(clone_name, task, interval_ms); } if (action_host) { action_name = crm_strdup_printf("%s%s %s", prefix, key, action_host); } else { action_name = crm_strdup_printf("%s%s", prefix, key); } free(key); } else if (safe_str_eq(action->task, CRM_OP_FENCE)) { const char *op = g_hash_table_lookup(action->meta, "stonith_action"); action_name = crm_strdup_printf("%s%s '%s' %s", prefix, action->task, op, action_host); } else if (action->rsc && action_host) { action_name = crm_strdup_printf("%s%s %s", prefix, action->uuid, action_host); } else if (action_host) { action_name = crm_strdup_printf("%s%s %s", prefix, action->task, action_host); } else { action_name = crm_strdup_printf("%s", action->uuid); } if (action_numbers) { // i.e. verbose char *with_id = crm_strdup_printf("%s (%d)", action_name, action->id); free(action_name); action_name = with_id; } return action_name; } static void create_dotfile(pe_working_set_t * data_set, const char *dot_file, gboolean all_actions) { GListPtr gIter = NULL; FILE *dot_strm = fopen(dot_file, "w"); if (dot_strm == NULL) { crm_perror(LOG_ERR, "Could not open %s for writing", dot_file); return; } fprintf(dot_strm, " digraph \"g\" {\n"); for (gIter = data_set->actions; gIter != NULL; gIter = gIter->next) { action_t *action = (action_t *) gIter->data; const char *style = "dashed"; const char *font = "black"; const char *color = "black"; char *action_name = create_action_name(action); crm_trace("Action %d: %s %s %p", action->id, action_name, action->uuid, action); if (is_set(action->flags, pe_action_pseudo)) { font = "orange"; } if (is_set(action->flags, pe_action_dumped)) { style = "bold"; color = "green"; } else if (action->rsc != NULL && is_not_set(action->rsc->flags, pe_rsc_managed)) { color = "red"; font = "purple"; if (all_actions == FALSE) { goto do_not_write; } } else if (is_set(action->flags, pe_action_optional)) { color = "blue"; if (all_actions == FALSE) { goto do_not_write; } } else { color = "red"; CRM_CHECK(is_set(action->flags, pe_action_runnable) == FALSE,; ); } set_bit(action->flags, pe_action_dumped); crm_trace("\"%s\" [ style=%s color=\"%s\" fontcolor=\"%s\"]", action_name, style, color, font); fprintf(dot_strm, "\"%s\" [ style=%s color=\"%s\" fontcolor=\"%s\"]\n", action_name, style, color, font); do_not_write: free(action_name); } for (gIter = data_set->actions; gIter != NULL; gIter = gIter->next) { action_t *action = (action_t *) gIter->data; GListPtr gIter2 = NULL; for (gIter2 = action->actions_before; gIter2 != NULL; gIter2 = gIter2->next) { action_wrapper_t *before = (action_wrapper_t *) gIter2->data; char *before_name = NULL; char *after_name = NULL; const char *style = "dashed"; gboolean optional = TRUE; if (before->state == pe_link_dumped) { optional = FALSE; style = "bold"; } else if (is_set(action->flags, pe_action_pseudo) && (before->type & pe_order_stonith_stop)) { continue; } else if (before->type == pe_order_none) { continue; } else if (is_set(before->action->flags, pe_action_dumped) && is_set(action->flags, pe_action_dumped) && before->type != pe_order_load) { optional = FALSE; } if (all_actions || optional == FALSE) { before_name = create_action_name(before->action); after_name = create_action_name(action); crm_trace("\"%s\" -> \"%s\" [ style = %s]", before_name, after_name, style); fprintf(dot_strm, "\"%s\" -> \"%s\" [ style = %s]\n", before_name, after_name, style); free(before_name); free(after_name); } } } fprintf(dot_strm, "}\n"); fflush(dot_strm); fclose(dot_strm); } static void setup_input(const char *input, const char *output) { int rc = pcmk_ok; cib_t *cib_conn = NULL; xmlNode *cib_object = NULL; char *local_output = NULL; if (input == NULL) { /* Use live CIB */ cib_conn = cib_new(); rc = cib_conn->cmds->signon(cib_conn, crm_system_name, cib_command); if (rc == pcmk_ok) { rc = cib_conn->cmds->query(cib_conn, NULL, &cib_object, cib_scope_local | cib_sync_call); } cib_conn->cmds->signoff(cib_conn); cib_delete(cib_conn); cib_conn = NULL; if (rc != pcmk_ok) { fprintf(stderr, "Live CIB query failed: %s (%d)\n", pcmk_strerror(rc), rc); crm_exit(crm_errno2exit(rc)); } else if (cib_object == NULL) { fprintf(stderr, "Live CIB query failed: empty result\n"); crm_exit(CRM_EX_NOINPUT); } } else if (safe_str_eq(input, "-")) { cib_object = filename2xml(NULL); } else { cib_object = filename2xml(input); } if (get_object_root(XML_CIB_TAG_STATUS, cib_object) == NULL) { create_xml_node(cib_object, XML_CIB_TAG_STATUS); } if (cli_config_update(&cib_object, NULL, FALSE) == FALSE) { free_xml(cib_object); crm_exit(CRM_EX_CONFIG); } if (validate_xml(cib_object, NULL, FALSE) != TRUE) { free_xml(cib_object); crm_exit(CRM_EX_CONFIG); } if (output == NULL) { char *pid = crm_getpid_s(); local_output = get_shadow_file(pid); temp_shadow = strdup(local_output); output = local_output; free(pid); } rc = write_xml_file(cib_object, output, FALSE); free_xml(cib_object); cib_object = NULL; if (rc < 0) { fprintf(stderr, "Could not create '%s': %s\n", output, pcmk_strerror(rc)); crm_exit(CRM_EX_CANTCREAT); } setenv("CIB_file", output, 1); free(local_output); } /* *INDENT-OFF* */ static struct crm_option long_options[] = { /* Top-level Options */ {"help", 0, 0, '?', "\tThis text"}, {"version", 0, 0, '$', "\tVersion information" }, {"quiet", 0, 0, 'Q', "\tDisplay only essentialoutput"}, {"verbose", 0, 0, 'V', "\tIncrease debug output"}, {"-spacer-", 0, 0, '-', "\nOperations:"}, {"run", 0, 0, 'R', "\tDetermine the cluster's response to the given configuration and status"}, {"simulate", 0, 0, 'S', "Simulate the transition's execution and display the resulting cluster status"}, {"in-place", 0, 0, 'X', "Simulate the transition's execution and store the result back to the input file"}, {"show-scores", 0, 0, 's', "Show allocation scores"}, {"show-utilization", 0, 0, 'U', "Show utilization information"}, {"profile", 1, 0, 'P', "Run all tests in the named directory to create profiling data"}, {"repeat", 1, 0, 'N', "With --profile, repeat each test N times and print timings"}, {"pending", 0, 0, 'j', "\tDisplay pending state if 'record-pending' is enabled", pcmk_option_hidden}, {"-spacer-", 0, 0, '-', "\nSynthetic Cluster Events:"}, {"node-up", 1, 0, 'u', "\tBring a node online"}, {"node-down", 1, 0, 'd', "\tTake a node offline"}, {"node-fail", 1, 0, 'f', "\tMark a node as failed"}, {"op-inject", 1, 0, 'i', "\tGenerate a failure for the cluster to react to in the simulation"}, {"-spacer-", 0, 0, '-', "\t\tValue is of the form ${resource}_${task}_${interval_in_ms}@${node}=${rc}."}, {"-spacer-", 0, 0, '-', "\t\tEg. memcached_monitor_20000@bart.example.com=7"}, {"-spacer-", 0, 0, '-', "\t\tFor more information on OCF return codes, refer to: https://clusterlabs.org/pacemaker/doc/en-US/Pacemaker/2.0/html/Pacemaker_Administration/s-ocf-return-codes.html"}, {"op-fail", 1, 0, 'F', "\tIf the specified task occurs during the simulation, have it fail with return code ${rc}"}, {"-spacer-", 0, 0, '-', "\t\tValue is of the form ${resource}_${task}_${interval_in_ms}@${node}=${rc}."}, {"-spacer-", 0, 0, '-', "\t\tEg. memcached_stop_0@bart.example.com=1\n"}, {"-spacer-", 0, 0, '-', "\t\tThe transition will normally stop at the failed action. Save the result with --save-output and re-run with --xml-file"}, { "set-datetime", required_argument, NULL, 't', "Set date/time (ISO 8601 format, see https://en.wikipedia.org/wiki/ISO_8601)" }, {"quorum", 1, 0, 'q', "\tSpecify a value for quorum"}, {"watchdog", 1, 0, 'w', "\tAssume a watchdog device is active"}, {"ticket-grant", 1, 0, 'g', "Grant a ticket"}, {"ticket-revoke", 1, 0, 'r', "Revoke a ticket"}, {"ticket-standby", 1, 0, 'b', "Make a ticket standby"}, {"ticket-activate", 1, 0, 'e', "Activate a ticket"}, {"-spacer-", 0, 0, '-', "\nOutput Options:"}, {"save-input", 1, 0, 'I', "\tSave the input configuration to the named file"}, {"save-output", 1, 0, 'O', "Save the output configuration to the named file"}, {"save-graph", 1, 0, 'G', "\tSave the transition graph (XML format) to the named file"}, {"save-dotfile", 1, 0, 'D', "Save the transition graph (DOT format) to the named file"}, {"all-actions", 0, 0, 'a', "\tDisplay all possible actions in the DOT graph - even ones not part of the transition"}, {"-spacer-", 0, 0, '-', "\nData Source:"}, {"live-check", 0, 0, 'L', "\tConnect to the CIB mamager and use the current CIB contents as input"}, {"xml-file", 1, 0, 'x', "\tRetrieve XML from the named file"}, {"xml-pipe", 0, 0, 'p', "\tRetrieve XML from stdin"}, {"-spacer-", 0, 0, '-', "\nExamples:\n"}, {"-spacer-", 0, 0, '-', "Pretend a recurring monitor action found memcached stopped on node fred.example.com and, during recovery, that the memcached stop action failed", pcmk_option_paragraph}, {"-spacer-", 0, 0, '-', " crm_simulate -LS --op-inject memcached:0_monitor_20000@bart.example.com=7 --op-fail memcached:0_stop_0@fred.example.com=1 --save-output /tmp/memcached-test.xml", pcmk_option_example}, {"-spacer-", 0, 0, '-', "Now see what the reaction to the stop failure would be", pcmk_option_paragraph}, {"-spacer-", 0, 0, '-', " crm_simulate -S --xml-file /tmp/memcached-test.xml", pcmk_option_example}, {0, 0, 0, 0} }; /* *INDENT-ON* */ static void profile_one(const char *xml_file, long long repeat, pe_working_set_t *data_set) { xmlNode *cib_object = NULL; clock_t start = 0; printf("* Testing %s ...", xml_file); fflush(stdout); cib_object = filename2xml(xml_file); start = clock(); if (get_object_root(XML_CIB_TAG_STATUS, cib_object) == NULL) { create_xml_node(cib_object, XML_CIB_TAG_STATUS); } if (cli_config_update(&cib_object, NULL, FALSE) == FALSE) { free_xml(cib_object); return; } if (validate_xml(cib_object, NULL, FALSE) != TRUE) { free_xml(cib_object); return; } for (int i = 0; i < repeat; ++i) { xmlNode *input = (repeat == 1)? cib_object : copy_xml(cib_object); data_set->input = input; get_date(data_set, false); pcmk__schedule_actions(data_set, input, NULL); pe_reset_working_set(data_set); } printf(" %.2f secs\n", (clock() - start) / (float) CLOCKS_PER_SEC); } #ifndef FILENAME_MAX # define FILENAME_MAX 512 #endif static void profile_all(const char *dir, long long repeat, pe_working_set_t *data_set) { struct dirent **namelist; int file_num = scandir(dir, &namelist, 0, alphasort); if (file_num > 0) { struct stat prop; char buffer[FILENAME_MAX]; while (file_num--) { if ('.' == namelist[file_num]->d_name[0]) { free(namelist[file_num]); continue; } else if (!crm_ends_with_ext(namelist[file_num]->d_name, ".xml")) { free(namelist[file_num]); continue; } snprintf(buffer, sizeof(buffer), "%s/%s", dir, namelist[file_num]->d_name); if (stat(buffer, &prop) == 0 && S_ISREG(prop.st_mode)) { profile_one(buffer, repeat, data_set); } free(namelist[file_num]); } free(namelist); } } -static int -count_resources(pe_working_set_t * data_set, resource_t * rsc) -{ - int count = 0; - GListPtr gIter = NULL; - - if (rsc == NULL) { - gIter = data_set->resources; - } else if (rsc->children) { - gIter = rsc->children; - } else { - return is_not_set(rsc->flags, pe_rsc_orphan); - } - - for (; gIter != NULL; gIter = gIter->next) { - count += count_resources(data_set, gIter->data); - } - return count; -} - int main(int argc, char **argv) { int rc = pcmk_ok; guint modified = 0; gboolean store = FALSE; gboolean process = FALSE; gboolean simulate = FALSE; gboolean all_actions = FALSE; gboolean have_stdout = FALSE; pe_working_set_t *data_set = NULL; const char *xml_file = "-"; const char *quorum = NULL; const char *watchdog = NULL; const char *test_dir = NULL; const char *dot_file = NULL; const char *graph_file = NULL; const char *input_file = NULL; const char *output_file = NULL; const char *repeat_s = NULL; int flag = 0; int index = 0; int argerr = 0; long long repeat = 1; GListPtr node_up = NULL; GListPtr node_down = NULL; GListPtr node_fail = NULL; GListPtr op_inject = NULL; GListPtr ticket_grant = NULL; GListPtr ticket_revoke = NULL; GListPtr ticket_standby = NULL; GListPtr ticket_activate = NULL; xmlNode *input = NULL; crm_log_cli_init("crm_simulate"); crm_set_options(NULL, "datasource operation [additional options]", long_options, "Tool for simulating the cluster's response to events"); if (argc < 2) { crm_help('?', CRM_EX_USAGE); } while (1) { flag = crm_get_option(argc, argv, &index); if (flag == -1) break; switch (flag) { case 'V': if (have_stdout == FALSE) { /* Redirect stderr to stdout so we can grep the output */ have_stdout = TRUE; close(STDERR_FILENO); dup2(STDOUT_FILENO, STDERR_FILENO); } crm_bump_log_level(argc, argv); action_numbers = TRUE; break; case '?': case '$': crm_help(flag, CRM_EX_OK); break; case 'p': xml_file = "-"; break; case 'Q': quiet = TRUE; break; case 'L': xml_file = NULL; break; case 'x': xml_file = optarg; break; case 'u': modified++; bringing_nodes_online = TRUE; node_up = g_list_append(node_up, optarg); break; case 'd': modified++; node_down = g_list_append(node_down, optarg); break; case 'f': modified++; node_fail = g_list_append(node_fail, optarg); break; case 't': use_date = strdup(optarg); break; case 'i': modified++; op_inject = g_list_append(op_inject, optarg); break; case 'F': process = TRUE; simulate = TRUE; op_fail = g_list_append(op_fail, optarg); break; case 'w': modified++; watchdog = optarg; break; case 'q': modified++; quorum = optarg; break; case 'g': modified++; ticket_grant = g_list_append(ticket_grant, optarg); break; case 'r': modified++; ticket_revoke = g_list_append(ticket_revoke, optarg); break; case 'b': modified++; ticket_standby = g_list_append(ticket_standby, optarg); break; case 'e': modified++; ticket_activate = g_list_append(ticket_activate, optarg); break; case 'a': all_actions = TRUE; break; case 's': process = TRUE; show_scores = TRUE; break; case 'U': process = TRUE; show_utilization = TRUE; break; case 'j': print_pending = TRUE; break; case 'S': process = TRUE; simulate = TRUE; break; case 'X': store = TRUE; process = TRUE; simulate = TRUE; break; case 'R': process = TRUE; break; case 'D': process = TRUE; dot_file = optarg; break; case 'G': process = TRUE; graph_file = optarg; break; case 'I': input_file = optarg; break; case 'O': output_file = optarg; break; case 'P': test_dir = optarg; break; case 'N': repeat_s = optarg; break; default: ++argerr; break; } } if (optind > argc) { ++argerr; } if (argerr) { crm_help('?', CRM_EX_USAGE); } data_set = pe_new_working_set(); if (data_set == NULL) { crm_perror(LOG_ERR, "Could not allocate working set"); rc = -ENOMEM; goto done; } if (test_dir != NULL) { if (repeat_s != NULL) { repeat = crm_int_helper(repeat_s, NULL); if (errno || (repeat < 1)) { fprintf(stderr, "--repeat must be positive integer, not '%s' -- using 1", repeat_s); repeat = 1; } } profile_all(test_dir, repeat, data_set); return CRM_EX_OK; } setup_input(xml_file, store ? xml_file : output_file); global_cib = cib_new(); rc = global_cib->cmds->signon(global_cib, crm_system_name, cib_command); if (rc != pcmk_ok) { fprintf(stderr, "Could not connect to the CIB manager: %s\n", pcmk_strerror(rc)); goto done; } rc = global_cib->cmds->query(global_cib, NULL, &input, cib_sync_call | cib_scope_local); if (rc != pcmk_ok) { fprintf(stderr, "Could not get local CIB: %s\n", pcmk_strerror(rc)); goto done; } data_set->input = input; get_date(data_set, true); if(xml_file) { set_bit(data_set->flags, pe_flag_sanitized); } set_bit(data_set->flags, pe_flag_stdout); cluster_status(data_set); if (quiet == FALSE) { int options = print_pending ? pe_print_pending : 0; if (is_set(data_set->flags, pe_flag_maintenance_mode)) { quiet_log("\n *** Resource management is DISABLED ***"); quiet_log("\n The cluster will not attempt to start, stop or recover services"); quiet_log("\n"); } if (data_set->disabled_resources || data_set->blocked_resources) { - quiet_log("%d of %d resources DISABLED and %d BLOCKED from being started due to failures\n", - data_set->disabled_resources, - count_resources(data_set, NULL), + quiet_log("%d of %d resource instances DISABLED and %d BLOCKED " + "from further action due to failure\n", + data_set->disabled_resources, data_set->ninstances, data_set->blocked_resources); } quiet_log("\nCurrent cluster status:\n"); print_cluster_status(data_set, options); } if (modified) { quiet_log("Performing requested modifications\n"); modify_configuration(data_set, global_cib, quorum, watchdog, node_up, node_down, node_fail, op_inject, ticket_grant, ticket_revoke, ticket_standby, ticket_activate); rc = global_cib->cmds->query(global_cib, NULL, &input, cib_sync_call); if (rc != pcmk_ok) { fprintf(stderr, "Could not get modified CIB: %s\n", pcmk_strerror(rc)); goto done; } cleanup_calculations(data_set); data_set->input = input; get_date(data_set, true); if(xml_file) { set_bit(data_set->flags, pe_flag_sanitized); } set_bit(data_set->flags, pe_flag_stdout); cluster_status(data_set); } if (input_file != NULL) { rc = write_xml_file(input, input_file, FALSE); if (rc < 0) { fprintf(stderr, "Could not create '%s': %s\n", input_file, pcmk_strerror(rc)); goto done; } } if (process || simulate) { crm_time_t *local_date = NULL; if (show_scores && show_utilization) { printf("Allocation scores and utilization information:\n"); } else if (show_scores) { fprintf(stdout, "Allocation scores:\n"); } else if (show_utilization) { printf("Utilization information:\n"); } pcmk__schedule_actions(data_set, input, local_date); input = NULL; /* Don't try and free it twice */ if (graph_file != NULL) { write_xml_file(data_set->graph, graph_file, FALSE); } if (dot_file != NULL) { create_dotfile(data_set, dot_file, all_actions); } if (quiet == FALSE) { GListPtr gIter = NULL; quiet_log("%sTransition Summary:\n", show_scores || show_utilization || modified ? "\n" : ""); fflush(stdout); LogNodeActions(data_set, TRUE); for (gIter = data_set->resources; gIter != NULL; gIter = gIter->next) { resource_t *rsc = (resource_t *) gIter->data; LogActions(rsc, data_set, TRUE); } } } rc = pcmk_ok; if (simulate) { if (run_simulation(data_set, global_cib, op_fail, quiet) != pcmk_ok) { rc = pcmk_err_generic; } if(quiet == FALSE) { get_date(data_set, true); quiet_log("\nRevised cluster status:\n"); set_bit(data_set->flags, pe_flag_stdout); cluster_status(data_set); print_cluster_status(data_set, 0); } } done: pe_free_working_set(data_set); global_cib->cmds->signoff(global_cib); cib_delete(global_cib); free(use_date); fflush(stderr); if (temp_shadow) { unlink(temp_shadow); free(temp_shadow); } crm_exit(crm_errno2exit(rc)); } diff --git a/tools/crm_ticket.c b/tools/crm_ticket.c index fbf90b87ee..11388a2c4e 100644 --- a/tools/crm_ticket.c +++ b/tools/crm_ticket.c @@ -1,904 +1,905 @@ /* * Copyright 2012-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include gboolean do_force = FALSE; gboolean BE_QUIET = FALSE; const char *ticket_id = NULL; const char *get_attr_name = NULL; const char *attr_name = NULL; const char *attr_value = NULL; const char *attr_id = NULL; const char *set_name = NULL; const char *attr_default = NULL; char ticket_cmd = 'S'; char *xml_file = NULL; int cib_options = cib_sync_call; static ticket_t * find_ticket(const char *ticket_id, pe_working_set_t * data_set) { ticket_t *ticket = NULL; ticket = g_hash_table_lookup(data_set->tickets, ticket_id); return ticket; } static void print_date(time_t time) { int lpc = 0; char date_str[26]; asctime_r(localtime(&time), date_str); for (; lpc < 26; lpc++) { if (date_str[lpc] == '\n') { date_str[lpc] = 0; } } fprintf(stdout, "'%s'", date_str); } static int print_ticket(ticket_t * ticket, gboolean raw, gboolean details) { if (raw) { fprintf(stdout, "%s\n", ticket->id); return pcmk_ok; } fprintf(stdout, "%s\t%s %s", ticket->id, ticket->granted ? "granted" : "revoked", ticket->standby ? "[standby]" : " "); if (details && g_hash_table_size(ticket->state) > 0) { GHashTableIter iter; const char *name = NULL; const char *value = NULL; int lpc = 0; fprintf(stdout, " ("); g_hash_table_iter_init(&iter, ticket->state); while (g_hash_table_iter_next(&iter, (void **)&name, (void **)&value)) { if (lpc > 0) { fprintf(stdout, ", "); } fprintf(stdout, "%s=", name); if (crm_str_eq(name, "last-granted", TRUE) || crm_str_eq(name, "expires", TRUE)) { print_date(crm_parse_int(value, 0)); } else { fprintf(stdout, "%s", value); } lpc++; } fprintf(stdout, ")\n"); } else { if (ticket->last_granted > -1) { fprintf(stdout, " last-granted="); print_date(ticket->last_granted); } fprintf(stdout, "\n"); } return pcmk_ok; } static int print_ticket_list(pe_working_set_t * data_set, gboolean raw, gboolean details) { GHashTableIter iter; ticket_t *ticket = NULL; g_hash_table_iter_init(&iter, data_set->tickets); while (g_hash_table_iter_next(&iter, NULL, (void **)&ticket)) { print_ticket(ticket, raw, details); } return pcmk_ok; } #define XPATH_MAX 1024 static int find_ticket_state(cib_t * the_cib, const char *ticket_id, xmlNode ** ticket_state_xml) { int offset = 0; int rc = pcmk_ok; xmlNode *xml_search = NULL; char *xpath_string = NULL; CRM_ASSERT(ticket_state_xml != NULL); *ticket_state_xml = NULL; xpath_string = calloc(1, XPATH_MAX); offset += snprintf(xpath_string + offset, XPATH_MAX - offset, "%s", "/cib/status/tickets"); if (ticket_id) { offset += snprintf(xpath_string + offset, XPATH_MAX - offset, "/%s[@id=\"%s\"]", XML_CIB_TAG_TICKET_STATE, ticket_id); } CRM_LOG_ASSERT(offset > 0); rc = the_cib->cmds->query(the_cib, xpath_string, &xml_search, cib_sync_call | cib_scope_local | cib_xpath); if (rc != pcmk_ok) { goto bail; } crm_log_xml_debug(xml_search, "Match"); if (xml_has_children(xml_search)) { if (ticket_id) { fprintf(stdout, "Multiple ticket_states match ticket_id=%s\n", ticket_id); } *ticket_state_xml = xml_search; } else { *ticket_state_xml = xml_search; } bail: free(xpath_string); return rc; } static int find_ticket_constraints(cib_t * the_cib, const char *ticket_id, xmlNode ** ticket_cons_xml) { int offset = 0; int rc = pcmk_ok; xmlNode *xml_search = NULL; char *xpath_string = NULL; CRM_ASSERT(ticket_cons_xml != NULL); *ticket_cons_xml = NULL; xpath_string = calloc(1, XPATH_MAX); offset += snprintf(xpath_string + offset, XPATH_MAX - offset, "%s/%s", get_object_path(XML_CIB_TAG_CONSTRAINTS), XML_CONS_TAG_RSC_TICKET); if (ticket_id) { offset += snprintf(xpath_string + offset, XPATH_MAX - offset, "[@ticket=\"%s\"]", ticket_id); } CRM_LOG_ASSERT(offset > 0); rc = the_cib->cmds->query(the_cib, xpath_string, &xml_search, cib_sync_call | cib_scope_local | cib_xpath); if (rc != pcmk_ok) { goto bail; } crm_log_xml_debug(xml_search, "Match"); *ticket_cons_xml = xml_search; bail: free(xpath_string); return rc; } static int dump_ticket_xml(cib_t * the_cib, const char *ticket_id) { int rc = pcmk_ok; xmlNode *state_xml = NULL; rc = find_ticket_state(the_cib, ticket_id, &state_xml); if (state_xml == NULL) { return rc; } fprintf(stdout, "State XML:\n"); if (state_xml) { char *state_xml_str = NULL; state_xml_str = dump_xml_formatted(state_xml); fprintf(stdout, "\n%s\n", crm_str(state_xml_str)); free_xml(state_xml); free(state_xml_str); } return pcmk_ok; } static int dump_constraints(cib_t * the_cib, const char *ticket_id) { int rc = pcmk_ok; xmlNode *cons_xml = NULL; char *cons_xml_str = NULL; rc = find_ticket_constraints(the_cib, ticket_id, &cons_xml); if (cons_xml == NULL) { return rc; } cons_xml_str = dump_xml_formatted(cons_xml); fprintf(stdout, "Constraints XML:\n\n%s\n", crm_str(cons_xml_str)); free_xml(cons_xml); free(cons_xml_str); return pcmk_ok; } static int get_ticket_state_attr(const char *ticket_id, const char *attr_name, const char **attr_value, pe_working_set_t * data_set) { ticket_t *ticket = NULL; CRM_ASSERT(attr_value != NULL); *attr_value = NULL; ticket = g_hash_table_lookup(data_set->tickets, ticket_id); if (ticket == NULL) { return -ENXIO; } *attr_value = g_hash_table_lookup(ticket->state, attr_name); if (*attr_value == NULL) { return -ENXIO; } return pcmk_ok; } static gboolean ticket_warning(const char *ticket_id, const char *action) { gboolean rc = FALSE; int offset = 0; static int text_max = 1024; char *warning = NULL; const char *word = NULL; warning = calloc(1, text_max); if (safe_str_eq(action, "grant")) { offset += snprintf(warning + offset, text_max - offset, "This command cannot help you verify whether '%s' has been already granted elsewhere.\n", ticket_id); word = "to"; } else { offset += snprintf(warning + offset, text_max - offset, "Revoking '%s' can trigger the specified 'loss-policy'(s) relating to '%s'.\n\n", ticket_id, ticket_id); offset += snprintf(warning + offset, text_max - offset, "You can check that with:\ncrm_ticket --ticket %s --constraints\n\n", ticket_id); offset += snprintf(warning + offset, text_max - offset, "Otherwise before revoking '%s', you may want to make '%s' standby with:\ncrm_ticket --ticket %s --standby\n\n", ticket_id, ticket_id, ticket_id); word = "from"; } offset += snprintf(warning + offset, text_max - offset, "If you really want to %s '%s' %s this site now, and you know what you are doing,\n", action, ticket_id, word); offset += snprintf(warning + offset, text_max - offset, "please specify --force."); CRM_LOG_ASSERT(offset > 0); fprintf(stdout, "%s\n", warning); free(warning); return rc; } static gboolean allow_modification(const char *ticket_id, GListPtr attr_delete, GHashTable *attr_set) { const char *value = NULL; GListPtr list_iter = NULL; if (do_force) { return TRUE; } if (g_hash_table_lookup_extended(attr_set, "granted", NULL, (gpointer *) & value)) { if (crm_is_true(value)) { ticket_warning(ticket_id, "grant"); return FALSE; } else { ticket_warning(ticket_id, "revoke"); return FALSE; } } for(list_iter = attr_delete; list_iter; list_iter = list_iter->next) { const char *key = (const char *)list_iter->data; if (safe_str_eq(key, "granted")) { ticket_warning(ticket_id, "revoke"); return FALSE; } } return TRUE; } static int modify_ticket_state(const char * ticket_id, GListPtr attr_delete, GHashTable * attr_set, cib_t * cib, pe_working_set_t * data_set) { int rc = pcmk_ok; xmlNode *xml_top = NULL; xmlNode *ticket_state_xml = NULL; gboolean found = FALSE; GListPtr list_iter = NULL; GHashTableIter hash_iter; char *key = NULL; char *value = NULL; ticket_t *ticket = NULL; rc = find_ticket_state(cib, ticket_id, &ticket_state_xml); if (rc == pcmk_ok) { crm_debug("Found a match state for ticket: id=%s", ticket_id); xml_top = ticket_state_xml; found = TRUE; } else if (rc != -ENXIO) { return rc; } else if (g_hash_table_size(attr_set) == 0){ return pcmk_ok; } else { xmlNode *xml_obj = NULL; xml_top = create_xml_node(NULL, XML_CIB_TAG_STATUS); xml_obj = create_xml_node(xml_top, XML_CIB_TAG_TICKETS); ticket_state_xml = create_xml_node(xml_obj, XML_CIB_TAG_TICKET_STATE); crm_xml_add(ticket_state_xml, XML_ATTR_ID, ticket_id); } for(list_iter = attr_delete; list_iter; list_iter = list_iter->next) { const char *key = (const char *)list_iter->data; xml_remove_prop(ticket_state_xml, key); } ticket = find_ticket(ticket_id, data_set); g_hash_table_iter_init(&hash_iter, attr_set); while (g_hash_table_iter_next(&hash_iter, (gpointer *) & key, (gpointer *) & value)) { crm_xml_add(ticket_state_xml, key, value); if (safe_str_eq(key, "granted") && (ticket == NULL || ticket->granted == FALSE) && crm_is_true(value)) { char *now = crm_ttoa(time(NULL)); crm_xml_add(ticket_state_xml, "last-granted", now); free(now); } } if (found && g_list_length(attr_delete)) { crm_log_xml_debug(xml_top, "Replace"); rc = cib->cmds->replace(cib, XML_CIB_TAG_STATUS, ticket_state_xml, cib_options); } else { crm_log_xml_debug(xml_top, "Update"); rc = cib->cmds->modify(cib, XML_CIB_TAG_STATUS, xml_top, cib_options); } free_xml(xml_top); return rc; } static int delete_ticket_state(const char *ticket_id, cib_t * cib) { xmlNode *ticket_state_xml = NULL; int rc = pcmk_ok; rc = find_ticket_state(cib, ticket_id, &ticket_state_xml); if (rc == -ENXIO) { return pcmk_ok; } else if (rc != pcmk_ok) { return rc; } crm_log_xml_debug(ticket_state_xml, "Delete"); rc = cib->cmds->remove(cib, XML_CIB_TAG_STATUS, ticket_state_xml, cib_options); if (rc == pcmk_ok) { fprintf(stdout, "Cleaned up %s\n", ticket_id); } free_xml(ticket_state_xml); return rc; } /* *INDENT-OFF* */ static struct crm_option long_options[] = { /* Top-level Options */ {"help", 0, 0, '?', "\t\tThis text"}, {"version", 0, 0, '$', "\t\tVersion information" }, {"verbose", 0, 0, 'V', "\t\tIncrease debug output"}, {"quiet", 0, 0, 'Q', "\t\tPrint only the value on stdout\n"}, {"ticket", 1, 0, 't', "\tTicket ID" }, {"-spacer-", 1, 0, '-', "\nQueries:"}, {"info", 0, 0, 'l', "\t\tDisplay the information of ticket(s)"}, {"details", 0, 0, 'L', "\t\tDisplay the details of ticket(s)"}, {"raw", 0, 0, 'w', "\t\tDisplay the IDs of ticket(s)"}, {"query-xml", 0, 0, 'q', "\tQuery the XML of ticket(s)"}, {"constraints",0, 0, 'c', "\tDisplay the rsc_ticket constraints that apply to ticket(s)"}, {"-spacer-", 1, 0, '-', "\nCommands:"}, {"grant", 0, 0, 'g', "\t\tGrant a ticket to this cluster site"}, {"revoke", 0, 0, 'r', "\t\tRevoke a ticket from this cluster site"}, {"standby", 0, 0, 's', "\t\tTell this cluster site this ticket is standby"}, {"activate", 0, 0, 'a', "\tTell this cluster site this ticket is active"}, {"-spacer-", 1, 0, '-', "\nAdvanced Commands:"}, {"get-attr", 1, 0, 'G', "\tDisplay the named attribute for a ticket"}, {"set-attr", 1, 0, 'S', "\tSet the named attribute for a ticket"}, {"delete-attr",1, 0, 'D', "\tDelete the named attribute for a ticket"}, {"cleanup", 0, 0, 'C', "\t\tDelete all state of a ticket at this cluster site"}, {"-spacer-", 1, 0, '-', "\nAdditional Options:"}, {"attr-value", 1, 0, 'v', "\tAttribute value to use with -S"}, {"default", 1, 0, 'd', "\t(Advanced) The default attribute value to display if none is found. For use with -G"}, {"force", 0, 0, 'f', "\t\t(Advanced) Force the action to be performed"}, {"xml-file", 1, 0, 'x', NULL, 1}, /* legacy options */ {"set-name", 1, 0, 'n', "\t(Advanced) ID of the instance_attributes object to change"}, {"nvpair", 1, 0, 'i', "\t(Advanced) ID of the nvpair object to change/delete"}, {"-spacer-", 1, 0, '-', "\nExamples:", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', "Display the info of tickets:", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_ticket --info", pcmk_option_example}, {"-spacer-", 1, 0, '-', "Display the detailed info of tickets:", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_ticket --details", pcmk_option_example}, {"-spacer-", 1, 0, '-', "Display the XML of 'ticketA':", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_ticket --ticket ticketA --query-xml", pcmk_option_example}, {"-spacer-", 1, 0, '-', "Display the rsc_ticket constraints that apply to 'ticketA':", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_ticket --ticket ticketA --constraints", pcmk_option_example}, {"-spacer-", 1, 0, '-', "Grant 'ticketA' to this cluster site:", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_ticket --ticket ticketA --grant", pcmk_option_example}, {"-spacer-", 1, 0, '-', "Revoke 'ticketA' from this cluster site:", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_ticket --ticket ticketA --revoke", pcmk_option_example}, {"-spacer-", 1, 0, '-', "Make 'ticketA' standby:", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', "The cluster site will treat a granted 'ticketA' as 'standby'."}, {"-spacer-", 1, 0, '-', "The dependent resources will be stopped or demoted gracefully without triggering loss-policies", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_ticket --ticket ticketA --standby", pcmk_option_example}, {"-spacer-", 1, 0, '-', "Activate 'ticketA' from being standby:", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_ticket --ticket ticketA --activate", pcmk_option_example}, {"-spacer-", 1, 0, '-', "Get the value of the 'granted' attribute for 'ticketA':", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_ticket --ticket ticketA --get-attr granted", pcmk_option_example}, {"-spacer-", 1, 0, '-', "Set the value of the 'standby' attribute for 'ticketA':", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_ticket --ticket ticketA --set-attr standby --attr-value true", pcmk_option_example}, {"-spacer-", 1, 0, '-', "Delete the 'granted' attribute for 'ticketA':", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_ticket --ticket ticketA --delete-attr granted", pcmk_option_example}, {"-spacer-", 1, 0, '-', "Erase the operation history of 'ticketA' at this cluster site:", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', "The cluster site will 'forget' the existing ticket state.", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_ticket --ticket ticketA --cleanup", pcmk_option_example}, {0, 0, 0, 0} }; /* *INDENT-ON* */ int main(int argc, char **argv) { pe_working_set_t *data_set = NULL; xmlNode *cib_xml_copy = NULL; xmlNode *cib_constraints = NULL; cib_t *cib_conn = NULL; crm_exit_t exit_code = CRM_EX_OK; int rc = pcmk_ok; int option_index = 0; int argerr = 0; int flag; guint modified = 0; GListPtr attr_delete = NULL; GHashTable *attr_set = crm_str_table_new(); crm_log_init(NULL, LOG_CRIT, FALSE, FALSE, argc, argv, FALSE); crm_set_options(NULL, "(query|command) [options]", long_options, "Perform tasks related to cluster tickets.\nAllows ticket attributes to be queried, modified and deleted.\n"); if (argc < 2) { crm_help('?', CRM_EX_USAGE); } while (1) { flag = crm_get_option(argc, argv, &option_index); if (flag == -1) break; switch (flag) { case 'V': crm_bump_log_level(argc, argv); break; case '$': case '?': crm_help(flag, CRM_EX_OK); break; case 'Q': BE_QUIET = TRUE; break; case 't': ticket_id = optarg; break; case 'l': case 'L': case 'w': case 'q': case 'c': ticket_cmd = flag; break; case 'g': g_hash_table_insert(attr_set, strdup("granted"), strdup("true")); modified++; break; case 'r': g_hash_table_insert(attr_set, strdup("granted"), strdup("false")); modified++; break; case 's': g_hash_table_insert(attr_set, strdup("standby"), strdup("true")); modified++; break; case 'a': g_hash_table_insert(attr_set, strdup("standby"), strdup("false")); modified++; break; case 'G': get_attr_name = optarg; ticket_cmd = flag; break; case 'S': attr_name = optarg; if (attr_name && attr_value) { g_hash_table_insert(attr_set, strdup(attr_name), strdup(attr_value)); attr_name = NULL; attr_value = NULL; modified++; } break; case 'D': attr_delete = g_list_append(attr_delete, optarg); modified++; break; case 'C': ticket_cmd = flag; break; case 'v': attr_value = optarg; if (attr_name && attr_value) { g_hash_table_insert(attr_set, strdup(attr_name), strdup(attr_value)); attr_name = NULL; attr_value = NULL; modified++; } break; case 'd': attr_default = optarg; break; case 'f': do_force = TRUE; break; case 'x': xml_file = strdup(optarg); break; case 'n': set_name = optarg; break; case 'i': attr_id = optarg; break; default: CMD_ERR("Argument code 0%o (%c) is not (?yet?) supported", flag, flag); ++argerr; break; } } if (optind < argc && argv[optind] != NULL) { CMD_ERR("non-option ARGV-elements:"); while (optind < argc && argv[optind] != NULL) { CMD_ERR("%s", argv[optind++]); ++argerr; } } if (optind > argc) { ++argerr; } if (argerr) { crm_help('?', CRM_EX_USAGE); } data_set = pe_new_working_set(); if (data_set == NULL) { crm_perror(LOG_CRIT, "Could not allocate working set"); exit_code = CRM_EX_OSERR; goto bail; } + set_bit(data_set->flags, pe_flag_no_counts); cib_conn = cib_new(); if (cib_conn == NULL) { CMD_ERR("Could not connect to the CIB manager"); exit_code = CRM_EX_DISCONNECT; goto bail; } rc = cib_conn->cmds->signon(cib_conn, crm_system_name, cib_command); if (rc != pcmk_ok) { CMD_ERR("Could not connect to the CIB manager: %s", pcmk_strerror(rc)); exit_code = crm_errno2exit(rc); goto bail; } if (xml_file != NULL) { cib_xml_copy = filename2xml(xml_file); } else { rc = cib_conn->cmds->query(cib_conn, NULL, &cib_xml_copy, cib_scope_local | cib_sync_call); if (rc != pcmk_ok) { CMD_ERR("Could not get local CIB: %s", pcmk_strerror(rc)); exit_code = crm_errno2exit(rc); goto bail; } } if (cli_config_update(&cib_xml_copy, NULL, FALSE) == FALSE) { CMD_ERR("Could not update local CIB to latest schema version"); exit_code = CRM_EX_CONFIG; goto bail; } data_set->input = cib_xml_copy; data_set->now = crm_time_new(NULL); cluster_status(data_set); /* For recording the tickets that are referenced in rsc_ticket constraints * but have never been granted yet. */ cib_constraints = get_object_root(XML_CIB_TAG_CONSTRAINTS, data_set->input); unpack_constraints(cib_constraints, data_set); if (ticket_cmd == 'l' || ticket_cmd == 'L' || ticket_cmd == 'w') { gboolean raw = FALSE; gboolean details = FALSE; if (ticket_cmd == 'L') { details = TRUE; } else if (ticket_cmd == 'w') { raw = TRUE; } if (ticket_id) { ticket_t *ticket = find_ticket(ticket_id, data_set); if (ticket == NULL) { CMD_ERR("No such ticket '%s'", ticket_id); exit_code = CRM_EX_NOSUCH; goto bail; } rc = print_ticket(ticket, raw, details); } else { rc = print_ticket_list(data_set, raw, details); } if (rc != pcmk_ok) { CMD_ERR("Could not print ticket: %s", pcmk_strerror(rc)); } exit_code = crm_errno2exit(rc); } else if (ticket_cmd == 'q') { rc = dump_ticket_xml(cib_conn, ticket_id); if (rc != pcmk_ok) { CMD_ERR("Could not query ticket XML: %s", pcmk_strerror(rc)); } exit_code = crm_errno2exit(rc); } else if (ticket_cmd == 'c') { rc = dump_constraints(cib_conn, ticket_id); if (rc != pcmk_ok) { CMD_ERR("Could not show ticket constraints: %s", pcmk_strerror(rc)); } exit_code = crm_errno2exit(rc); } else if (ticket_cmd == 'G') { const char *value = NULL; if (ticket_id == NULL) { CMD_ERR("Must supply ticket ID with -t"); exit_code = CRM_EX_NOSUCH; goto bail; } rc = get_ticket_state_attr(ticket_id, get_attr_name, &value, data_set); if (rc == pcmk_ok) { fprintf(stdout, "%s\n", value); } else if (rc == -ENXIO && attr_default) { fprintf(stdout, "%s\n", attr_default); rc = pcmk_ok; } exit_code = crm_errno2exit(rc); } else if (ticket_cmd == 'C') { if (ticket_id == NULL) { CMD_ERR("Must supply ticket ID with -t"); exit_code = CRM_EX_USAGE; goto bail; } if (do_force == FALSE) { ticket_t *ticket = NULL; ticket = find_ticket(ticket_id, data_set); if (ticket == NULL) { CMD_ERR("No such ticket '%s'", ticket_id); exit_code = CRM_EX_NOSUCH; goto bail; } if (ticket->granted) { ticket_warning(ticket_id, "revoke"); exit_code = CRM_EX_INSUFFICIENT_PRIV; goto bail; } } rc = delete_ticket_state(ticket_id, cib_conn); if (rc != pcmk_ok) { CMD_ERR("Could not clean up ticket: %s", pcmk_strerror(rc)); } exit_code = crm_errno2exit(rc); } else if (modified) { if (ticket_id == NULL) { CMD_ERR("Must supply ticket ID with -t"); exit_code = CRM_EX_USAGE; goto bail; } if (attr_value && (attr_name == NULL || strlen(attr_name) == 0)) { CMD_ERR("Must supply attribute name with -S for -v %s", attr_value); exit_code = CRM_EX_USAGE; goto bail; } if (attr_name && (attr_value == NULL || strlen(attr_value) == 0)) { CMD_ERR("Must supply attribute value with -v for -S %s", attr_name); exit_code = CRM_EX_USAGE; goto bail; } if (allow_modification(ticket_id, attr_delete, attr_set) == FALSE) { CMD_ERR("Ticket modification not allowed"); exit_code = CRM_EX_INSUFFICIENT_PRIV; goto bail; } rc = modify_ticket_state(ticket_id, attr_delete, attr_set, cib_conn, data_set); if (rc != pcmk_ok) { CMD_ERR("Could not modify ticket: %s", pcmk_strerror(rc)); } exit_code = crm_errno2exit(rc); } else if (ticket_cmd == 'S') { /* Correct usage was handled in the "if (modified)" block above, so * this is just for reporting usage errors */ if (attr_name == NULL || strlen(attr_name) == 0) { // We only get here if ticket_cmd was left as default CMD_ERR("Must supply a command"); exit_code = CRM_EX_USAGE; goto bail; } if (ticket_id == NULL) { CMD_ERR("Must supply ticket ID with -t"); exit_code = CRM_EX_USAGE; goto bail; } if (attr_value == NULL || strlen(attr_value) == 0) { CMD_ERR("Must supply value with -v for -S %s", attr_name); exit_code = CRM_EX_USAGE; goto bail; } } else { CMD_ERR("Unknown command: %c", ticket_cmd); exit_code = CRM_EX_USAGE; } bail: if (attr_set) { g_hash_table_destroy(attr_set); } attr_set = NULL; if (attr_delete) { g_list_free(attr_delete); } attr_delete = NULL; pe_free_working_set(data_set); data_set = NULL; if (cib_conn != NULL) { cib_conn->cmds->signoff(cib_conn); cib_delete(cib_conn); } if (rc == -pcmk_err_no_quorum) { CMD_ERR("Use --force to ignore quorum"); } crm_exit(exit_code); } diff --git a/tools/crm_verify.c b/tools/crm_verify.c index 870a95ca2b..3e38814e88 100644 --- a/tools/crm_verify.c +++ b/tools/crm_verify.c @@ -1,270 +1,271 @@ /* * Copyright 2004-2019 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include gboolean USE_LIVE_CIB = FALSE; char *cib_save = NULL; extern gboolean stage0(pe_working_set_t * data_set); /* *INDENT-OFF* */ static struct crm_option long_options[] = { /* Top-level Options */ {"help", 0, 0, '?', "\tThis text"}, {"version", 0, 0, '$', "\tVersion information" }, {"verbose", 0, 0, 'V', "\tSpecify multiple times to increase debug output\n"}, {"-spacer-", 1, 0, '-', "\nData sources:"}, {"live-check", 0, 0, 'L', "Check the configuration used by the running cluster\n"}, {"xml-file", 1, 0, 'x', "Check the configuration in the named file"}, {"xml-text", 1, 0, 'X', "Check the configuration in the supplied string"}, {"xml-pipe", 0, 0, 'p', "Check the configuration piped in via stdin"}, {"-spacer-", 1, 0, '-', "\nAdditional Options:"}, {"save-xml", 1, 0, 'S', "Save the verified XML to the named file. Most useful with -L"}, {"-spacer-", 1, 0, '-', "\nExamples:", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', "Check the consistency of the configuration in the running cluster:", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_verify --live-check", pcmk_option_example}, {"-spacer-", 1, 0, '-', "Check the consistency of the configuration in a given file and produce verbose output:", pcmk_option_paragraph}, {"-spacer-", 1, 0, '-', " crm_verify --xml-file file.xml --verbose", pcmk_option_example}, {0, 0, 0, 0} }; /* *INDENT-ON* */ int main(int argc, char **argv) { xmlNode *cib_object = NULL; xmlNode *status = NULL; int argerr = 0; int flag; int option_index = 0; pe_working_set_t *data_set = NULL; cib_t *cib_conn = NULL; int rc = pcmk_ok; bool verbose = FALSE; gboolean xml_stdin = FALSE; const char *xml_tag = NULL; const char *xml_file = NULL; const char *xml_string = NULL; crm_log_cli_init("crm_verify"); crm_set_options(NULL, "[modifiers] data_source", long_options, "check a Pacemaker configuration for errors" "\n\nCheck the well-formedness of a complete Pacemaker XML configuration," "\n\nits conformance to the configured schema, and the presence of common" "\n\nmisconfigurations. Problems reported as errors must be fixed before the" "\n\ncluster will work properly. It is left to the administrator to decide" "\n\nwhether to fix problems reported as warnings."); while (1) { flag = crm_get_option(argc, argv, &option_index); if (flag == -1) break; switch (flag) { case 'X': crm_trace("Option %c => %s", flag, optarg); xml_string = optarg; break; case 'x': crm_trace("Option %c => %s", flag, optarg); xml_file = optarg; break; case 'p': xml_stdin = TRUE; break; case 'S': cib_save = optarg; break; case 'V': verbose = TRUE; crm_bump_log_level(argc, argv); break; case 'L': USE_LIVE_CIB = TRUE; break; case '$': case '?': crm_help(flag, CRM_EX_OK); break; default: fprintf(stderr, "Option -%c is not yet supported\n", flag); ++argerr; break; } } if (optind < argc) { printf("non-option ARGV-elements: "); while (optind < argc) { printf("%s ", argv[optind++]); } printf("\n"); } if (optind > argc) { ++argerr; } if (argerr) { crm_err("%d errors in option parsing", argerr); crm_help(flag, CRM_EX_USAGE); } crm_info("=#=#=#=#= Getting XML =#=#=#=#="); if (USE_LIVE_CIB) { cib_conn = cib_new(); rc = cib_conn->cmds->signon(cib_conn, crm_system_name, cib_command); } if (USE_LIVE_CIB) { if (rc == pcmk_ok) { int options = cib_scope_local | cib_sync_call; crm_info("Reading XML from: live cluster"); rc = cib_conn->cmds->query(cib_conn, NULL, &cib_object, options); } if (rc != pcmk_ok) { fprintf(stderr, "Live CIB query failed: %s\n", pcmk_strerror(rc)); goto done; } if (cib_object == NULL) { fprintf(stderr, "Live CIB query failed: empty result\n"); rc = -ENOMSG; goto done; } } else if (xml_file != NULL) { cib_object = filename2xml(xml_file); if (cib_object == NULL) { fprintf(stderr, "Couldn't parse input file: %s\n", xml_file); rc = -ENODATA; goto done; } } else if (xml_string != NULL) { cib_object = string2xml(xml_string); if (cib_object == NULL) { fprintf(stderr, "Couldn't parse input string: %s\n", xml_string); rc = -ENODATA; goto done; } } else if (xml_stdin) { cib_object = stdin2xml(); if (cib_object == NULL) { fprintf(stderr, "Couldn't parse input from STDIN.\n"); rc = -ENODATA; goto done; } } else { fprintf(stderr, "No configuration source specified." " Use --help for usage information.\n"); rc = -ENODATA; goto done; } xml_tag = crm_element_name(cib_object); if (safe_str_neq(xml_tag, XML_TAG_CIB)) { fprintf(stderr, "This tool can only check complete configurations (i.e. those starting with ).\n"); rc = -EBADMSG; goto done; } if (cib_save != NULL) { write_xml_file(cib_object, cib_save, FALSE); } status = get_object_root(XML_CIB_TAG_STATUS, cib_object); if (status == NULL) { create_xml_node(cib_object, XML_CIB_TAG_STATUS); } if (validate_xml(cib_object, NULL, FALSE) == FALSE) { crm_config_err("CIB did not pass schema validation"); free_xml(cib_object); cib_object = NULL; } else if (cli_config_update(&cib_object, NULL, FALSE) == FALSE) { crm_config_error = TRUE; free_xml(cib_object); cib_object = NULL; fprintf(stderr, "The cluster will NOT be able to use this configuration.\n"); fprintf(stderr, "Please manually update the configuration to conform to the %s syntax.\n", xml_latest_schema()); } data_set = pe_new_working_set(); if (data_set == NULL) { rc = -errno; crm_perror(LOG_CRIT, "Unable to allocate working set"); goto done; } + set_bit(data_set->flags, pe_flag_no_counts); if (cib_object == NULL) { } else if (status != NULL || USE_LIVE_CIB) { /* live queries will always have a status section and can do a full simulation */ pcmk__schedule_actions(data_set, cib_object, NULL); } else { data_set->now = crm_time_new(NULL); data_set->input = cib_object; stage0(data_set); } pe_free_working_set(data_set); if (crm_config_error) { fprintf(stderr, "Errors found during check: config not valid\n"); if (verbose == FALSE) { fprintf(stderr, " -V may provide more details\n"); } rc = -pcmk_err_schema_validation; } else if (crm_config_warning) { fprintf(stderr, "Warnings found during check: config may not be valid\n"); if (verbose == FALSE) { fprintf(stderr, " Use -V -V for more details\n"); } rc = -pcmk_err_schema_validation; } if (USE_LIVE_CIB && cib_conn) { cib_conn->cmds->signoff(cib_conn); cib_delete(cib_conn); } done: crm_exit(crm_errno2exit(rc)); }