Refactor: libcib: new function cib__signon_attempts()

Fix: tools: CIB clients retry signon upon an EAGAIN error

Refactor: libcib: new function cib__signon_attempts()

Fix: tools: CIB clients retry signon upon an EAGAIN error

Fix: libcrmcluster: use local hostname for node name if not configured in…

Test: cts-cli: update expected output for crm_verify to distinguish…

Fix: tools: crm_verify distinguishes configuration warnings and errors

Fix: libcrmcluster: use local hostname for node name if not configured in…

Test: cts-cli: update expected output for crm_verify to distinguish…

Fix: tools: crm_verify distinguishes configuration warnings and errors

fence_sbd: --sbd-path defaults to /usr/sbin/sbd

configure: check additional paths for programs

Fix: libpacemaker: correctly retrieve any existing fail-count for increment

Fix: libpacemaker: correctly retrieve any existing fail-count for increment

Fix: libstonithd: avoid use-after-free when retrieving metadata of Linux-HA…

Fix: libstonithd: free escaped metadata descriptions with g_free()

Fix: libstonithd: free escaped metadata descriptions with g_free()

Fix: libstonithd: avoid use-after-free when retrieving metadata of Linux-HA…

Fix: tools: make crm_mon exit upon loss of the attached pseudo-terminal

Log: scheduler: deprecate Nagios and Upstart resources even if built with…

Log: scheduler: correctly log resource IDs of the deprecated classes

Refactor: fencer: functionize checking whether a fencing action should be done…

Log: fencer: log for an offline target of watchdog fencing correctly

Log: fencer: log the correct value of `stonith-watchdog-timeout`

Fix: fencer: ask an online target of watchdog fencing to acknowledge an `on`…

Fix: fencer: make an `on` action for an offline target of watchdog fencing fail

Fix: fencer: prevent an in-progress watchdog fencing from timing out on query…

Fix: fencer: prevent watchdog fencing from timing out on insufficient `stonith…

Fix: fencer: prevent watchdog fencing from timing out on insufficient per…

Fix: fencer: prevent watchdog fencing targeting an offline node from timing out…

Fix: fencer: respect per-device watchdog fencing timeout from any query result

Log: fencer: log the correct unit of action-specific timeouts

Log: fencer: log the units of fencing timeouts and delays

Test: cts-fencing: test for parsing action-specific timeouts with units

Fix: fencer: correctly parse action-specific timeouts with units

Doc: Pacemaker Explained: action-specific timeouts are `timeout` type now

Fix: libcrmcommon: use uint32_t for 32-bit magic numbers

Fix: libcrmservice: avoid async zombie children by resending ignored SIGCHLD

Doc: fencer: fix pcmk_delay_max description

Fix: libcrmcommon: avoid file descriptor leak in IPC client with async…

Doc: libcrmcommon, Pacemaker Explained: improve descriptions of cluster options

Fix: controller: do not check whether watchdog fencing is enabled for the node…

Test: cts-scheduler: add regression test about a pending node that doesn't have…

Fix: scheduler: Do not fence a pending node that doesn't have an uname in node…

Test: cts-scheduler: update regression test about not fencing a pending node…

Feature: libcrmcommon: bump feature set to 3.18.0 for handling node pending…

Feature: libcrmcluster: add ability to record the timestamps of a node becoming…

Feature: controller: record the timestamps of a node becoming a cluster member…

Feature: scheduler: add `node-pending-timeout` cluster option

Feature: scheduler: ability to determine node status from `in_ccm` and `crmd`…

Test: cts-scheduler: add regression test for fencing a pending node that has…

Feature: controller: read `node-pending-timeout` cluster option

Feature: scheduler: fence a pending node that has reached `node-pending-timeout`

Feature: controller: trigger a new transition if a pending node has reached…

Refactor: controller: pass abort_timer to the timeout function…

Refactor: libcrmcluster: ability to search for a node cache entry by uuid…

Refactor: libcrmcluster: internal functions for getting a node cache entry by…

Fix: controller: update node state correctly based on any existing node cache…

Test: cts-fencing: regression test for fencing timeouts taking fencing delays…

Fix: fencer: add correct values of pcmk_delay_base/max to query rely

Fix: fencer: total fencing timeout takes any requested fencing delay into…

Fix: fencer: per-operation fencing timeout takes any requested fencing delay…

Fix: fencer: fencing timeouts take any pcmk_delay_base into account

Fix: fencer: apply requested fencing delay only for the first device

Fix: controller: initial timeout for fencing callback takes any priority…

Log: controller: log fencing timeout consistently in seconds as priority…

Log: controller: use "target" terminology consistently

Fix: libpacemaker: initial timeout for fencing callback takes any requested…

Fix: fencer: fencing timeout sent to peer takes no delay into account

Fix: controller: avoid use-after-free when disconnecting proxy IPCs during…

Test: cts-cli: add tests for crm_attribute -p or OCF_RESOURCE_INSTANCE…

Fix: libcrmcommon: allow crm_attribute to try OCF_RESOURCE_INSTANCE environment…

Test: cts-cli: update tests for crm_attribute -p or OCF_RESOURCE_INSTANCE…

Test: cts-regression: reflect any test failures again with the return code

Test: cts-cli: consolidate transforming of error messages about ENXIO for…

Refactor: libpacemaker: unify bailing out in pcmk__inject_node()

Fix: libpacemaker: avoid assertion failure if a node_state entry doesn't have…

Test: cts-scheduler: add test for preventing a leftover pending monitor from…

Fix: scheduler: prevent a leftover pending monitor from causing unexpected stop…

Fix: libpacemaker: ensure any pending recurring monitor gets updated if it fails

Test: cts-scheduler: update test for preventing a leftover pending monitor from…

Fix: scheduler: prevent inactive instances from starting if probe is unrunnable…

Test: cts-scheduler: update test for preventing inactive instances from…

Log: sbd-inquisitor: ensure a log info only tells the fact about how…

Log: sbd-inquisitor: log a warning if SBD_PACEMAKER is overridden by -P or -PP…

Log: sbd-inquisitor: do not warn about startup syncing if pacemaker integration…

Fix: sbd-inquisitor: fail startup if pacemaker integration is disabled while…

Test: cts-cli: add test for text output of crm_resource --list-operations/-O

Fix: tools: fix syntax on resetting options in crm_resource

Log: libcrmservice: fix the trace logging about finishing reading stdout/stderr

Fix: scheduler: restart a resource configured with extra operation parameters…

Refactor: scheduler: exchange the names of attr_in_string() and…

Fix: scheduler,controller: consistently calculate and record secure digest…

Fix: libpacemaker: correctly find the best operation history entry to calculate…

Test: cts-scheduler: restart a resource configured with extra operation…

Fix: scheduler: calculate digests based on the crm_feature_set version of the…

Log: controller: log an info instead of a warning for a stonith/shutdown that…

Test: cts-scheduler: add missing last-rc-change attributes to the faked…

Fix: scheduler: a failed migrate_from always matters for the source if the…