diff --git a/extra/resources/ClusterMon.in b/extra/resources/ClusterMon.in index ed8bc76732..c533d2356f 100755 --- a/extra/resources/ClusterMon.in +++ b/extra/resources/ClusterMon.in @@ -1,260 +1,262 @@ #!@BASH_PATH@ # # ocf:pacemaker:ClusterMon resource agent # # Original copyright 2004 SUSE LINUX AG, Lars Marowsky-Bre # Later changes copyright 2008-2019 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 # (GPLv2) WITHOUT ANY WARRANTY. # # Starts crm_mon in background which logs cluster status as # html to the specified file. # # OCF instance parameters: # OCF_RESKEY_user # OCF_RESKEY_pidfile # OCF_RESKEY_update # OCF_RESKEY_extra_options # OCF_RESKEY_htmlfile ####################################################################### # Initialization: : ${OCF_FUNCTIONS=${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs} . ${OCF_FUNCTIONS} : ${__OCF_ACTION=$1} ####################################################################### meta_data() { cat < 1.0 This is a ClusterMon Resource Agent. It outputs current cluster status to the html. Runs crm_mon in the background, recording the cluster status to an HTML file The user we want to run crm_mon as The user we want to run crm_mon as How frequently should we update the cluster status (in milliseconds). For compatibility with old documentation, values less than 1000 will be treated as seconds. Update interval in milliseconds Additional options to pass to crm_mon. Eg. -n -r Extra options PID file location to ensure only one instance is running PID file Location to write HTML output to. HTML output END } ####################################################################### ClusterMon_usage() { cat </dev/null | \ grep -qE "[c]rm_mon.*${OCF_RESKEY_pidfile}" rc=$? case $rc in 0) exit $OCF_SUCCESS;; 1) exit $OCF_NOT_RUNNING;; *) exit $OCF_ERR_GENERIC;; esac fi fi exit $OCF_NOT_RUNNING } CheckOptions() { while getopts Vi:nrh:cdp: OPTION do case $OPTION in V|n|r|c|d);; i) ocf_log warn "You should not have specified the -i option, since OCF_RESKEY_update is set already!";; h) ocf_log warn "You should not have specified the -h option, since OCF_RESKEY_htmlfile is set already!";; p) ocf_log warn "You should not have specified the -p option, since OCF_RESKEY_pidfile is set already!";; *) return $OCF_ERR_ARGS;; esac done if [ $? -ne 0 ]; then return $OCF_ERR_ARGS fi # We should have eaten all options at this stage shift $(($OPTIND -1)) if [ $# -gt 0 ]; then false else true fi } ClusterMon_validate() { # Existence of the user if [ ! -z $OCF_RESKEY_user ]; then getent passwd "$OCF_RESKEY_user" >/dev/null if [ $? -eq 0 ]; then : Yes, user exists. We can further check his permission on crm_mon if necessary else ocf_log err "The user $OCF_RESKEY_user does not exist!" exit $OCF_ERR_ARGS fi fi # Pidfile better be an absolute path case $OCF_RESKEY_pidfile in /*) ;; *) ocf_log warn "You should have pidfile($OCF_RESKEY_pidfile) of absolute path!" ;; esac # Check the update interval if ocf_is_decimal "$OCF_RESKEY_update" && [ $OCF_RESKEY_update -gt 0 ]; then : else ocf_log err "Invalid update interval $OCF_RESKEY_update. It should be positive integer!" exit $OCF_ERR_ARGS fi if CheckOptions $OCF_RESKEY_extra_options; then : else ocf_log err "Invalid options $OCF_RESKEY_extra_options!" exit $OCF_ERR_ARGS fi # Htmlfile better be an absolute path case $OCF_RESKEY_htmlfile in /*) ;; *) ocf_log warn "You should have htmlfile($OCF_RESKEY_htmlfile) of absolute path!" ;; esac echo "Validate OK" return $OCF_SUCCESS } if [ $# -ne 1 ]; then ClusterMon_usage exit $OCF_ERR_ARGS fi : ${OCF_RESKEY_update:="15000"} : ${OCF_RESKEY_pidfile:="/tmp/ClusterMon_${OCF_RESOURCE_INSTANCE}.pid"} : ${OCF_RESKEY_htmlfile:="/tmp/ClusterMon_${OCF_RESOURCE_INSTANCE}.html"} if [ ${OCF_RESKEY_update} -ge 1000 ]; then OCF_RESKEY_update=$(( $OCF_RESKEY_update / 1000 )) fi CMON_CMD="${HA_SBIN_DIR}/crm_mon -p $OCF_RESKEY_pidfile -d -i $OCF_RESKEY_update $OCF_RESKEY_extra_options -h $OCF_RESKEY_htmlfile" case $__OCF_ACTION in meta-data) meta_data exit $OCF_SUCCESS ;; start) ClusterMon_start ;; stop) ClusterMon_stop ;; monitor) ClusterMon_monitor ;; validate-all) ClusterMon_validate ;; usage|help) ClusterMon_usage exit $OCF_SUCCESS ;; *) ClusterMon_usage exit $OCF_ERR_UNIMPLEMENTED ;; esac exit $? + +# vim: set filetype=sh expandtab tabstop=4 softtabstop=4 shiftwidth=4 textwidth=80: diff --git a/extra/resources/Dummy b/extra/resources/Dummy index 643d1afa65..33d892b611 100755 --- a/extra/resources/Dummy +++ b/extra/resources/Dummy @@ -1,278 +1,280 @@ #!/bin/sh # # ocf:pacemaker:Dummy resource agent # # Original copyright 2004 SUSE LINUX AG, Lars Marowsky-Bre -# Later changes copyright 2008-2018 the Pacemaker project contributors +# Later changes copyright 2008-2019 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 # (GPLv2) WITHOUT ANY WARRANTY. # # Does nothing but wait a few seconds, can be configured to fail occassionally # ####################################################################### # Initialization: : ${OCF_FUNCTIONS=${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs} . ${OCF_FUNCTIONS} : ${__OCF_ACTION=$1} ####################################################################### meta_data() { cat < 1.0 This is a Dummy Resource Agent. It does absolutely nothing except keep track of whether its running or not. Its purpose in life is for testing and to serve as a template for RA writers. NB: Please pay attention to the timeouts specified in the actions section below. They should be meaningful for the kind of resource the agent manages. They should be the minimum advised timeouts, but they shouldn't/cannot cover _all_ possible resource instances. So, try to be neither overly generous nor too stingy, but moderate. The minimum timeouts should never be below 10 seconds. Example stateless resource agent Location to store the resource state in. State file Fake password field Password Fake attribute that can be changed to cause a reload Fake attribute that can be changed to cause a reload Number of seconds to sleep during operations. This can be used to test how the cluster reacts to operation timeouts. Operation sleep duration in seconds. Start actions will return failure if running on the host specified here, but the resource will start successfully anyway (future monitor calls will find it running). This can be used to test on-fail=ignore. Report bogus start failure on specified host If this is set, the environment will be dumped to this file for every call. Environment dump file END } ####################################################################### # don't exit on TERM, to test that pacemaker-execd makes sure that we do exit trap sigterm_handler TERM sigterm_handler() { ocf_log info "They use TERM to bring us down. No such luck." # Since we're likely going to get KILLed, clean up any monitor # serialization in progress, so the next probe doesn't return an error. rm -f "${VERIFY_SERIALIZED_FILE}" return } dummy_usage() { cat <> "${OCF_RESKEY_envfile}" fi } dummy_start() { dummy_monitor DS_RETVAL=$? if [ $DS_RETVAL -eq $OCF_SUCCESS ]; then if [ "$(uname -n)" = "${OCF_RESKEY_fail_start_on}" ]; then DS_RETVAL=$OCF_ERR_GENERIC fi return $DS_RETVAL fi touch "${OCF_RESKEY_state}" DS_RETVAL=$? if [ "$(uname -n)" = "${OCF_RESKEY_fail_start_on}" ]; then DS_RETVAL=$OCF_ERR_GENERIC fi return $DS_RETVAL } dummy_stop() { dummy_monitor --force if [ $? -eq $OCF_SUCCESS ]; then rm ${OCF_RESKEY_state} fi rm -f "${VERIFY_SERIALIZED_FILE}" return $OCF_SUCCESS } dummy_monitor() { # Monitor _MUST!_ differentiate correctly between running # (SUCCESS), failed (ERROR) or _cleanly_ stopped (NOT RUNNING). # That is THREE states, not just yes/no. if [ $OCF_RESKEY_op_sleep -ne 0 ]; then if [ "$1" = "" ] && [ -f "${VERIFY_SERIALIZED_FILE}" ]; then # two monitor ops have occurred at the same time. # This verifies a condition in pacemaker-execd regression tests. ocf_log err "$VERIFY_SERIALIZED_FILE exists already" ocf_exit_reason "alternate universe collision" return $OCF_ERR_GENERIC fi touch "${VERIFY_SERIALIZED_FILE}" sleep ${OCF_RESKEY_op_sleep} rm "${VERIFY_SERIALIZED_FILE}" fi if [ -f "${OCF_RESKEY_state}" ]; then # Multiple monitor levels are defined to support various tests case "$OCF_CHECK_LEVEL" in 10) # monitor level with delay, useful for testing timeouts sleep 30 ;; 20) # monitor level that fails intermittently n=$(expr "$(dd if=/dev/urandom bs=1 count=1 2>/dev/null | od | head -1 | cut -f2 -d' ')" % 5) if [ $n -eq 1 ]; then ocf_exit_reason "smoke detected near CPU fan" return $OCF_ERR_GENERIC fi ;; 30) # monitor level that always fails ocf_exit_reason "hyperdrive quota reached" return $OCF_ERR_GENERIC ;; *) ;; esac return $OCF_SUCCESS fi return $OCF_NOT_RUNNING } dummy_validate() { # Is the state directory writable? state_dir=`dirname "$OCF_RESKEY_state"` touch "$state_dir/$$" if [ $? -ne 0 ]; then return $OCF_ERR_ARGS fi rm "$state_dir/$$" return $OCF_SUCCESS } : ${OCF_RESKEY_fake=dummy} : ${OCF_RESKEY_op_sleep=0} : ${OCF_RESKEY_CRM_meta_interval=0} : ${OCF_RESKEY_CRM_meta_globally_unique:="false"} if [ -z "$OCF_RESKEY_state" ]; then OCF_RESKEY_state="${HA_VARRUN%%/}/Dummy-${OCF_RESOURCE_INSTANCE}.state" if [ ${OCF_RESKEY_CRM_meta_globally_unique} = "false" ]; then # Strip off the trailing clone marker (note + is not portable in sed) OCF_RESKEY_state=`echo $OCF_RESKEY_state | sed s/:[0-9][0-9]*\.state/.state/` fi fi VERIFY_SERIALIZED_FILE="${OCF_RESKEY_state}.serialized" dump_env case $__OCF_ACTION in meta-data) meta_data exit $OCF_SUCCESS ;; start) dummy_start;; stop) dummy_stop;; monitor) dummy_monitor;; migrate_to) ocf_log info "Migrating ${OCF_RESOURCE_INSTANCE} to ${OCF_RESKEY_CRM_meta_migrate_target}." dummy_stop ;; migrate_from) ocf_log info "Migrating ${OCF_RESOURCE_INSTANCE} from ${OCF_RESKEY_CRM_meta_migrate_source}." dummy_start ;; reload) ocf_log err "Reloading..." dummy_start ;; validate-all) dummy_validate;; usage|help) dummy_usage exit $OCF_SUCCESS ;; *) dummy_usage exit $OCF_ERR_UNIMPLEMENTED ;; esac rc=$? ocf_log debug "${OCF_RESOURCE_INSTANCE} $__OCF_ACTION : $rc" exit $rc + +# vim: set filetype=sh expandtab tabstop=4 softtabstop=4 shiftwidth=4 textwidth=80: diff --git a/extra/resources/HealthCPU b/extra/resources/HealthCPU index 767b154516..1f5cdcb211 100755 --- a/extra/resources/HealthCPU +++ b/extra/resources/HealthCPU @@ -1,197 +1,199 @@ #!/bin/sh # # ocf:pacemaker:HealthCPU resource agent # # 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 # (GPLv2) WITHOUT ANY WARRANTY. # # # Measures CPUs idling and writes #health-cpu status into the CIB # ################################ # # TODO: Enter default values # Error handling in getting uptime # ################################## ####################################################################### # Initialization: : ${OCF_FUNCTIONS=${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs} . ${OCF_FUNCTIONS} : ${__OCF_ACTION=$1} ####################################################################### meta_data() { cat < 1.0 Systhem health agent that measures the CPU idling and updates the #health-cpu attribute. System health CPU usage Location to store the resource state in. State file Lower (!) limit of idle percentage to switch the health attribute to yellow. I.e. the #health-cpu will go yellow if the %idle of the CPU falls below 50%. Lower limit for yellow health attribute Lower (!) limit of idle percentage to switch the health attribute to red. I.e. the #health-cpu will go red if the %idle of the CPU falls below 10%. Lower limit for red health attribute END } ####################################################################### dummy_usage() { cat < 1.0 Systhem health agent that measures the CPU iowait via top and updates the #health-iowait attribute. System health based on CPU iowait measurement Location to store the resource state in. State file Upper limit of iowait percentage to switch the health attribute to yellow. I.e. the #health-iowait will go yellow if the %iowait of the CPU gets higher than 10%. Upper limit for yellow health attribute Upper limit of iowait percentage to switch the health attribute to red. I.e. the #health-iowait will go red if the %iowait of the CPU get higher than 15%. Upper limit for red health attribute END } ####################################################################### agent_usage() { cat < 1.0 Systhem health agent that checks the S.M.A.R.T. status of the given drives and updates the #health-smart attribute. SMART health status Location to store the resource state in. State file The drive(s) to check as a SPACE separated list. Enter the full path to the device, e.g. "/dev/sda". Drives to check The device type(s) to assume for the drive(s) being tested as a SPACE separated list. Device types Lower limit of the temperature in deg C of the drive(s). Below this limit the status will be red. Lower limit for the red smart attribute Upper limit of the temperature if deg C of the drives(s). If the drive reports a temperature higher than this value the status of #health-smart will be red. Upper limit for red smart attribute Number of deg C below/above the upper/lower temp limits at which point the status of #health-smart will change to yellow. Deg C below/above the upper limits for yellow smart attribute END } ####################################################################### check_temperature() { if [ $1 -lt ${lower_red_limit} ] ; then ocf_log info "Drive ${DRIVE} ${DEVICE} too cold: ${1} C" $ATTRDUP -n "#health-smart" -U "red" -d "5s" return 1 fi if [ $1 -gt ${upper_red_limit} ] ; then ocf_log info "Drive ${DRIVE} ${DEVICE} too hot: ${1} C" $ATTRDUP -n "#health-smart" -U "red" -d "5s" return 1 fi if [ $1 -lt ${lower_yellow_limit} ] ; then ocf_log info "Drive ${DRIVE} ${DEVICE} quite cold: ${1} C" $ATTRDUP -n "#health-smart" -U "yellow" -d "5s" return 1 fi if [ $1 -gt ${upper_yellow_limit} ] ; then ocf_log info "Drive ${DRIVE} ${DEVICE} quite hot: ${1} C" $ATTRDUP -n "#health-smart" -U "yellow" -d "5s" return 1 fi } init_smart() { #Set temperature defaults if [ -z ${OCF_RESKEY_temp_warning} ]; then yellow_threshold=5 else yellow_threshold=${OCF_RESKEY_temp_warning} fi if [ -z ${OCF_RESKEY_temp_lower_limit} ] ; then lower_red_limit=0 else lower_red_limit=${OCF_RESKEY_temp_lower_limit} fi lower_yellow_limit=$((${lower_red_limit}+${yellow_threshold})) if [ -z ${OCF_RESKEY_temp_upper_limit} ] ; then upper_red_limit=60 else upper_red_limit=${OCF_RESKEY_temp_upper_limit} fi upper_yellow_limit=$((${upper_red_limit}-${yellow_threshold})) #Set disk defaults if [ -z "${OCF_RESKEY_drives}" ] ; then DRIVES="/dev/sda" else DRIVES=${OCF_RESKEY_drives} fi #Test for presence of smartctl if [ ! -x $SMARTCTL ] ; then ocf_log err "${SMARTCTL} not installed." exit $OCF_ERR_INSTALLED fi for DRIVE in $DRIVES; do if [ "${OCF_RESKEY_devices}" ]; then for DEVICE in ${OCF_RESKEY_devices}; do $SMARTCTL -d $DEVICE -i ${DRIVE} | grep -q "SMART support is: Enabled" if [ $? -ne "0" ] ; then ocf_log err "S.M.A.R.T. not enabled for drive "${DRIVE} exit $OCF_ERR_INSTALLED fi done else $SMARTCTL -i ${DRIVE} | grep -q "SMART support is: Enabled" if [ $? -ne "0" ] ; then ocf_log err "S.M.A.R.T. not enabled for drive "${DRIVE} exit $OCF_ERR_INSTALLED fi fi done } HealthSMART_usage() { cat < 1.0 This is an example resource agent that implements two states Example stateful resource agent Location to store the resource state in State file If this is set, the environment will be dumped to this file for every call. Environment dump file The notify action will sleep for this many seconds before returning, to simulate a long-running notify. Notify delay in seconds END exit $OCF_SUCCESS } ####################################################################### stateful_usage() { cat < ${OCF_RESKEY_state} } stateful_check_state() { target=$1 if [ -f ${OCF_RESKEY_state} ]; then state=`cat ${OCF_RESKEY_state}` if [ "x$target" = "x$state" ]; then return 0 fi else if [ "x$target" = "x" ]; then return 0 fi fi return 1 } dump_env() { if [ "${OCF_RESKEY_envfile}" != "" ]; then echo "### ${__OCF_ACTION} @ $(date) ### $(env | sort) ###" >> "${OCF_RESKEY_envfile}" fi } stateful_start() { stateful_check_state master if [ $? = 0 ]; then # CRM Error - Should never happen return $OCF_RUNNING_MASTER fi stateful_update slave $CRM_MASTER -v ${slave_score} return 0 } stateful_demote() { stateful_check_state if [ $? = 0 ]; then # CRM Error - Should never happen return $OCF_NOT_RUNNING fi stateful_update slave $CRM_MASTER -v ${slave_score} return 0 } stateful_promote() { stateful_check_state if [ $? = 0 ]; then return $OCF_NOT_RUNNING fi stateful_update master $CRM_MASTER -v ${master_score} return 0 } stateful_stop() { $CRM_MASTER -D stateful_check_state master if [ $? = 0 ]; then # CRM Error - Should never happen return $OCF_RUNNING_MASTER fi if [ -f ${OCF_RESKEY_state} ]; then rm ${OCF_RESKEY_state} fi return 0 } stateful_monitor() { stateful_check_state "master" if [ $? = 0 ]; then if [ $OCF_RESKEY_CRM_meta_interval = 0 ]; then # Restore the master setting during probes $CRM_MASTER -v ${master_score} fi return $OCF_RUNNING_MASTER fi stateful_check_state "slave" if [ $? = 0 ]; then if [ $OCF_RESKEY_CRM_meta_interval = 0 ]; then # Restore the master setting during probes $CRM_MASTER -v ${slave_score} fi return $OCF_SUCCESS fi if [ -f ${OCF_RESKEY_state} ]; then echo "File '${OCF_RESKEY_state}' exists but contains unexpected contents" cat ${OCF_RESKEY_state} return $OCF_ERR_GENERIC fi return 7 } stateful_notify() { if [ "${OCF_RESKEY_notify_delay}" != "0" ]; then sleep "${OCF_RESKEY_notify_delay}" fi return $OCF_SUCCESS } stateful_validate() { exit $OCF_SUCCESS } : ${slave_score=5} : ${master_score=10} : ${OCF_RESKEY_CRM_meta_interval=0} : ${OCF_RESKEY_notify_delay=0} : ${OCF_RESKEY_CRM_meta_globally_unique:="false"} if [ "x$OCF_RESKEY_state" = "x" ]; then if [ ${OCF_RESKEY_CRM_meta_globally_unique} = "false" ]; then state="${HA_VARRUN%%/}/Stateful-${OCF_RESOURCE_INSTANCE}.state" # Strip off the trailing clone marker OCF_RESKEY_state=`echo $state | sed s/:[0-9][0-9]*\.state/.state/` else OCF_RESKEY_state="${HA_VARRUN%%/}/Stateful-${OCF_RESOURCE_INSTANCE}.state" fi fi dump_env case $__OCF_ACTION in meta-data) meta_data;; start) stateful_start;; promote) stateful_promote;; demote) stateful_demote;; notify) stateful_notify ;; stop) stateful_stop;; monitor) stateful_monitor;; validate-all) stateful_validate;; usage|help) stateful_usage $OCF_SUCCESS;; *) stateful_usage $OCF_ERR_UNIMPLEMENTED;; esac exit $? + +# vim: set filetype=sh expandtab tabstop=4 softtabstop=4 shiftwidth=4 textwidth=80: diff --git a/extra/resources/SysInfo.in b/extra/resources/SysInfo.in index 77888048f7..8de271462d 100755 --- a/extra/resources/SysInfo.in +++ b/extra/resources/SysInfo.in @@ -1,370 +1,372 @@ #!@BASH_PATH@ # # ocf:pacemaker:SysInfo resource agent # # Original copyright 2004 SUSE LINUX AG, Lars Marowsky-Bre # Later changes copyright 2008-2019 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 # (GPLv2) WITHOUT ANY WARRANTY. # # # This agent records (in the CIB) various attributes of a node # ####################################################################### # Initialization: : ${OCF_FUNCTIONS=${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs} . ${OCF_FUNCTIONS} : ${__OCF_ACTION=$1} ####################################################################### meta_data() { cat < 1.0 This is a SysInfo Resource Agent. It records (in the CIB) various attributes of a node Sample Linux output: arch: i686 os: Linux-2.4.26-gentoo-r14 free_swap: 1999 cpu_info: Intel(R) Celeron(R) CPU 2.40GHz cpu_speed: 4771.02 cpu_cores: 1 cpu_load: 0.00 ram_total: 513 ram_free: 117 root_free: 2.4 #health_disk: red Sample Darwin output: arch: i386 os: Darwin-8.6.2 cpu_info: Intel Core Duo cpu_speed: 2.16 cpu_cores: 2 cpu_load: 0.18 ram_total: 2016 ram_free: 787 root_free: 13 #health_disk: green Units: free_swap: MB ram_*: MB cpu_speed (Linux): bogomips cpu_speed (Darwin): GHz *_free: GB (or user-defined: disk_unit) SysInfo resource agent PID file PID file Interval to allow values to stabilize Dampening Delay Filesystems or Paths to be queried for free disk space as a SPACE separated list - e.g "/dev/sda1 /tmp". Results will be written to an attribute with leading slashes removed, and other slashes replaced with underscore, and the word 'free' appended - e.g for /dev/sda1 it would be 'dev_sda1_free'. Note: The root filesystem '/' is always queried to an attribute named 'root_free' List of Filesytems/Paths to query for free disk space Unit to report disk free space in. Can be one of: B, K, M, G, T, P (case-insensitive) Unit to report disk free space in The amount of free space required in monitored disks. If any of the monitored disks has less than this amount of free space, , with the node attribute "#health_disk" changing to "red", all resources will move away from the node. Set the node-health-strategy property appropriately for this to take effect. If the unit is not specified, it defaults to disk_unit. minimum disk free space required END } ####################################################################### UpdateStat() { name=$1; shift value="$*" printf "%s:\t%s\n" "$name" "$value" if [ "$__OCF_ACTION" = "start" ] ; then ${HA_SBIN_DIR}/attrd_updater ${OCF_RESKEY_delay} -S status -n $name -B "$value" else ${HA_SBIN_DIR}/attrd_updater ${OCF_RESKEY_delay} -S status -n $name -v "$value" fi } SysInfoStats() { UpdateStat arch "`uname -m`" UpdateStat os "`uname -s`-`uname -r`" case `uname -s` in "Darwin") mem=`top -l 1 | grep Mem: | awk '{print $10}'` mem_used=`top -l 1 | grep Mem: | awk '{print $8}'` mem=`SysInfo_mem_units $mem` mem_used=`SysInfo_mem_units $mem_used` mem_total=`expr $mem_used + $mem` cpu_type=`system_profiler SPHardwareDataType | awk -F': ' '/^CPU Type/ {print $2; exit}'` cpu_speed=`system_profiler SPHardwareDataType | awk -F': ' '/^CPU Speed/ {print $2; exit}'` cpu_cores=`system_profiler SPHardwareDataType | awk -F': ' '/^Number Of/ {print $2; exit}'` cpu_load=`uptime | awk '{ print $10 }'` ;; "FreeBSD") cpu_type=`sysctl -in hw.model` cpu_speed=`sysctl -in dev.cpu.0.freq` cpu_cores=`sysctl -in hw.ncpu` cpu_load=`sysctl -in vm.loadavg | awk '{ print $4 }'` free_pages=`sysctl -in vm.stats.vm.v_free_count` page_count=`sysctl -in vm.stats.vm.v_page_count` page_size=`sysctl -in vm.stats.vm.v_page_size` mem=`expr $free_pages \* $page_size / 1024 / 1024`M mem_total=`expr $page_count \* $page_size / 1024 / 1024`M ;; "Linux") if [ -f /proc/cpuinfo ]; then cpu_type=`awk -F': ' '/model name/ {print $2; exit}' /proc/cpuinfo` cpu_speed=`awk -F': ' '/bogomips/ {print $2; exit}' /proc/cpuinfo` cpu_cores=`grep "^processor" /proc/cpuinfo | wc -l` fi cpu_load=`uptime | awk '{ print $10 }'` if [ -f /proc/meminfo ]; then # meminfo results are in kB mem=`grep "SwapFree" /proc/meminfo | awk '{print $2"k"}'` if [ ! -z $mem ]; then UpdateStat free_swap "$(SysInfo_mem_units "$mem")" fi mem=`grep "Inactive" /proc/meminfo | awk '{print $2"k"}'` mem_total=`grep "MemTotal" /proc/meminfo | awk '{print $2"k"}'` else mem=`top -n 1 | grep Mem: | awk '{print $7}'` fi ;; *) esac if [ x != x"$cpu_type" ]; then UpdateStat cpu_info "$cpu_type" fi if [ x != x"$cpu_speed" ]; then UpdateStat cpu_speed "$cpu_speed" fi if [ x != x"$cpu_cores" ]; then UpdateStat cpu_cores "$cpu_cores" fi if [ x != x"$cpu_load" ]; then UpdateStat cpu_load "$cpu_load" fi if [ ! -z "$mem" ]; then # Massage the memory values UpdateStat ram_total "$(SysInfo_mem_units "$mem_total")" UpdateStat ram_free "$(SysInfo_mem_units "$mem")" fi # Portability notes: # o tail: explicit "-n" not available in Solaris; instead simplify # 'tail -n ' to the equivalent 'tail -'. for disk in "/" ${OCF_RESKEY_disks}; do unset disk_free disk_label disk_free=`df -h ${disk} | tail -1 | awk '{print $4}'` if [ x != x"$disk_free" ]; then disk_label=`echo $disk | sed -e 's#^/$#root#;s#^/*##;s#/#_#g'` disk_free=`SysInfo_hdd_units $disk_free` UpdateStat ${disk_label}_free $disk_free if [ -n "$MIN_FREE" ]; then if [ $disk_free -le $MIN_FREE ]; then UpdateStat "#health_disk" "red" else UpdateStat "#health_disk" "green" fi fi fi done } SysInfo_megabytes() { # Size in megabytes echo $1 | awk '{ n = $0; sub(/[0-9]+(.[0-9]+)?/, ""); split(n, a, $0); n=a[1]; if ($0 == "G" || $0 == "") { n *= 1024 }; if (/^kB?/) { n /= 1024 }; printf "%d\n", n }' # Intentionally round to an integer } SysInfo_mem_units() { mem=$1 if [ -z $1 ]; then return fi mem=$(SysInfo_megabytes "$1") # Round to the next multiple of 50 r=$(($mem % 50)) if [ $r != 0 ]; then mem=$(($mem + 50 - $r)) fi echo $mem } SysInfo_hdd_units() { # Defauts to size in gigabytes case $OCF_RESKEY_disk_unit in [Pp]) echo $(($(SysInfo_megabytes "$1") / 1024 / 1024 / 1024));; [Tt]) echo $(($(SysInfo_megabytes "$1") / 1024 / 1024));; [Gg]) echo $(($(SysInfo_megabytes "$1") / 1024));; [Mm]) echo "$(SysInfo_megabytes "$1")" ;; [Kk]) echo $(($(SysInfo_megabytes "$1") * 1024));; [Bb]) echo $(($(SysInfo_megabytes "$1") * 1024 * 1024));; *) ocf_log err "Invalid value for disk_unit: $OCF_RESKEY_disk_unit" echo $(($(SysInfo_megabytes "$1") / 1024));; esac } SysInfo_usage() { cat < $OCF_RESKEY_pidfile SysInfoStats exit $OCF_SUCCESS } SysInfo_stop() { rm $OCF_RESKEY_pidfile exit $OCF_SUCCESS } SysInfo_monitor() { if [ -f $OCF_RESKEY_pidfile ]; then clone=`cat $OCF_RESKEY_pidfile` fi if [ x$clone = x ]; then rm $OCF_RESKEY_pidfile exit $OCF_NOT_RUNNING elif [ $clone = $OCF_RESKEY_clone ]; then SysInfoStats exit $OCF_SUCCESS elif ocf_is_true "$OCF_RESKEY_CRM_meta_globally_unique"; then SysInfoStats exit $OCF_SUCCESS fi exit $OCF_NOT_RUNNING } SysInfo_validate() { return $OCF_SUCCESS } if [ $# -ne 1 ]; then SysInfo_usage exit $OCF_ERR_ARGS fi : ${OCF_RESKEY_pidfile:="${HA_VARRUN%%/}/SysInfo-${OCF_RESOURCE_INSTANCE}"} : ${OCF_RESKEY_disk_unit:="G"} : ${OCF_RESKEY_clone:="0"} if [ x != x${OCF_RESKEY_delay} ]; then OCF_RESKEY_delay="-d ${OCF_RESKEY_delay}" else OCF_RESKEY_delay="-d 0" fi MIN_FREE="" if [ -n "$OCF_RESKEY_min_disk_free" ]; then ocf_is_decimal "$OCF_RESKEY_min_disk_free" && OCF_RESKEY_min_disk_free="$OCF_RESKEY_min_disk_free$OCF_RESKEY_disk_unit" MIN_FREE=`SysInfo_hdd_units $OCF_RESKEY_min_disk_free` fi case $__OCF_ACTION in meta-data) meta_data exit $OCF_SUCCESS ;; start) SysInfo_start ;; stop) SysInfo_stop ;; monitor) SysInfo_monitor ;; validate-all) SysInfo_validate ;; usage|help) SysInfo_usage exit $OCF_SUCCESS ;; *) SysInfo_usage exit $OCF_ERR_UNIMPLEMENTED ;; esac exit $? + +# vim: set filetype=sh expandtab tabstop=4 softtabstop=4 shiftwidth=4 textwidth=80: diff --git a/extra/resources/SystemHealth b/extra/resources/SystemHealth index de4b056ff4..e23ad4ec4b 100755 --- a/extra/resources/SystemHealth +++ b/extra/resources/SystemHealth @@ -1,237 +1,239 @@ #!/bin/sh # # ocf:pacemaker:SystemHealth resource agent # # 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 # (GPLv2) WITHOUT ANY WARRANTY. # ####################################################################### # Initialization: : ${OCF_FUNCTIONS=${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs} . ${OCF_FUNCTIONS} : ${__OCF_ACTION=$1} ####################################################################### meta_data() { cat < 1.0 This is a SystemHealth Resource Agent. It is used to monitor the health of a system via IPMI. SystemHealth resource agent END } ####################################################################### SystemHealth_usage() { cat < /dev/null 2>&1 RC=$? if [ $RC != 0 ]; then ocf_log err "servicelog_notify not found!" return $OCF_ERR_INSTALLED fi which ipmiservicelogd > /dev/null 2>&1 RC=$? if [ $RC != 0 ]; then ocf_log err "ipmiservicelogd not found!" return $OCF_ERR_INSTALLED fi test -x $OCF_RESKEY_program RC=$? if [ $RC != 0 ]; then ocf_log err "$OCF_RESKEY_program not found!" return $OCF_ERR_INSTALLED fi } SystemHealth_start() { SystemHealth_monitor RC=$? if [ $RC = $OCF_ERR_GENERIC ]; then return $OCF_ERR_GENERIC elif [ $RC = $OCF_SUCCESS ]; then ocf_log warn "starting an already started SystemHealth" return $OCF_SUCCESS fi service ipmi start > /dev/null 2>&1 RC=$? if [ $RC != 0 ]; then ocf_log err "Could not start service IPMI!" return $OCF_ERR_GENERIC fi ipmiservicelogd smi 0 > /dev/null 2>&1 & RC=$? if [ $RC != 0 ]; then ocf_log err "Could not start ipmiservicelogd!" return $OCF_ERR_GENERIC fi servicelog_notify --add --type=EVENT --command="$OCF_RESKEY_program" --method=num_arg --match='type=4' > /dev/null 2>&1 RC=$? if [ $RC != 0 ]; then ocf_log err "servicelog_notify register handler failed!" return $OCF_ERR_GENERIC fi return $OCF_SUCCESS } SystemHealth_stop() { SystemHealth_monitor RC=$? if [ $RC = $OCF_ERR_GENERIC ]; then return $OCF_ERR_GENERIC elif [ $RC = $OCF_SUCCESS ]; then killall ipmiservicelogd RC1=$? if [ $RC1 != 0 ]; then ocf_log err "Could not stop ipmiservicelogd!" fi servicelog_notify --remove --command="$OCF_RESKEY_program" > /dev/null 2>&1 RC2=$? if [ $RC2 != 0 ]; then ocf_log err "servicelog_notify remove handler failed!" fi if [ $RC1 -eq 0 ] && [ $RC2 -eq 0 ]; then return $OCF_SUCCESS else return $OCF_ERR_GENERIC fi elif [ $RC = $OCF_NOT_RUNNING ]; then ocf_log warn "stopping an already stopped SystemHealth" return $OCF_SUCCESS else ocf_log err "SystemHealth_stop: should not be here!" return $OCF_ERR_GENERIC fi } SystemHealth_monitor() { # Monitor _MUST!_ differentiate correctly between running # (SUCCESS), failed (ERROR) or _cleanly_ stopped (NOT RUNNING). # That is THREE states, not just yes/no. if [ ! -f /var/run/ipmiservicelogd.pid0 ]; then ocf_log debug "ipmiservicelogd is not running!" return $OCF_NOT_RUNNING fi ps -p "$(cat /var/run/ipmiservicelogd.pid0)" >/dev/null 2>&1 RC=$? if [ $RC != 0 ]; then ocf_log debug "ipmiservicelogd's pid `cat /var/run/ipmiservicelogd.pid0` is not running!" rm /var/run/ipmiservicelogd.pid0 return $OCF_ERR_GENERIC fi servicelog_notify --list --command="$OCF_RESKEY_program" > /dev/null 2>&1 RC=$? if [ $RC = 0 ]; then return $OCF_SUCCESS else return $OCF_NOT_RUNNING fi } SystemHealth_validate() { SystemHealth_check_tools RC=$? if [ $RC != 0 ]; then return $RC fi return $OCF_SUCCESS } : ${OCF_RESKEY_program=/usr/sbin/notifyServicelogEvent} case $__OCF_ACTION in meta-data) meta_data exit $OCF_SUCCESS ;; usage|help) SystemHealth_usage exit $OCF_SUCCESS ;; esac SystemHealth_check_tools RC=$? if [ $RC != 0 ]; then case $__OCF_ACTION in stop) exit $OCF_SUCCESS;; *) exit $RC;; esac fi case $__OCF_ACTION in start) SystemHealth_start;; stop) SystemHealth_stop;; monitor) SystemHealth_monitor;; reload) ocf_log info "Reloading..." SystemHealth_start ;; validate-all) ;; *) SystemHealth_usage exit $OCF_ERR_UNIMPLEMENTED ;; esac rc=$? ocf_log debug "${OCF_RESOURCE_INSTANCE} $__OCF_ACTION : $rc" exit $rc + +# vim: set filetype=sh expandtab tabstop=4 softtabstop=4 shiftwidth=4 textwidth=80: diff --git a/extra/resources/attribute b/extra/resources/attribute index 57b690f1c5..626607f6f6 100755 --- a/extra/resources/attribute +++ b/extra/resources/attribute @@ -1,235 +1,235 @@ #!/bin/sh # # ocf:pacemaker:attribute resource agent # -# Copyright 2016-2018 the Pacemaker project contributors +# Copyright 2016-2019 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. # USAGE="Usage: $0 {start|stop|monitor|migrate_to|migrate_from|validate-all|meta-data} Expects to have a fully populated OCF RA-compliant environment set." # Load OCF helper functions : ${OCF_FUNCTIONS=${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs} . ${OCF_FUNCTIONS} : ${__OCF_ACTION=$1} # Ensure certain variables are set and not empty : ${HA_VARRUN:="/var/run"} : ${OCF_RESKEY_CRM_meta_globally_unique:="false"} : ${OCF_RESOURCE_INSTANCE:="undef"} DEFAULT_STATE_FILE="${HA_VARRUN%%/}/opa-${OCF_RESOURCE_INSTANCE}.state" if [ ${OCF_RESKEY_CRM_meta_globally_unique} = "false" ]; then # Strip off any trailing clone marker (note + is not portable in sed) DEFAULT_STATE_FILE=$(echo "$DEFAULT_STATE_FILE" | sed s/:[0-9][0-9]*\.state/.state/) fi DEFAULT_ATTR_NAME="opa-${OCF_RESOURCE_INSTANCE}" DEFAULT_ACTIVE_VALUE="1" DEFAULT_INACTIVE_VALUE="0" : ${OCF_RESKEY_state:="$DEFAULT_STATE_FILE"} : ${OCF_RESKEY_name:="$DEFAULT_ATTR_NAME"} # Values may be empty string if [ -z ${OCF_RESKEY_active_value+x} ]; then OCF_RESKEY_active_value="$DEFAULT_ACTIVE_VALUE" fi if [ -z ${OCF_RESKEY_inactive_value+x} ]; then OCF_RESKEY_inactive_value="$DEFAULT_INACTIVE_VALUE" fi usage() { USAGE_RC=$1 cat < 1.0 Manages a node attribute This resource agent controls a node attribute for the node it's running on. It sets the attribute one way when started, and another way when stopped, according to the configuration parameters. Full path of a temporary file to store the resource state in State file Name of node attribute to manage Attribute name Value to use for node attribute when resource becomes active (empty string is discouraged, because monitor cannot distinguish it from a query error) Attribute value when active Value to use for node attribute when resource becomes inactive Attribute value when inactive END return $OCF_SUCCESS } validate() { if [ "$OCF_RESKEY_active_value" = "$OCF_RESKEY_inactive_value" ]; then ocf_exit_reason "active value '%s' must be different from inactive value '%s'" \ "$OCF_RESKEY_active_value" "$OCF_RESKEY_inactive_value" return $OCF_ERR_CONFIGURED fi VALIDATE_DIR=$(dirname "${OCF_RESKEY_state}") if [ ! -d "$VALIDATE_DIR" ]; then ocf_exit_reason "state file '$OCF_RESKEY_state' does not have a valid directory" return $OCF_ERR_PERM fi if [ ! -w "$VALIDATE_DIR" ] || [ ! -x "$VALIDATE_DIR" ]; then ocf_exit_reason "insufficient privileges on directory of state file '$OCF_RESKEY_state'" return $OCF_ERR_PERM fi return $OCF_SUCCESS } get_attribute() { GET_LINE=$(attrd_updater -n "$OCF_RESKEY_name" -Q 2>/dev/null) if [ $? -ne 0 ]; then echo "" else echo "$GET_LINE" | sed -e "s/.* value=\"\(.*\)\"$/\1/" fi } set_attribute() { attrd_updater -n "$OCF_RESKEY_name" -U "$1" 2>/dev/null # TODO if above call is async, loop until get_attribute returns expected value } check_attribute() { CHECK_VALUE=$(get_attribute) CHECK_REASON="" if [ ! -f "$OCF_RESKEY_state" ]; then if [ "$CHECK_VALUE" != "" ] && [ "$CHECK_VALUE" != "$OCF_RESKEY_inactive_value" ]; then CHECK_REASON="Node attribute $OCF_RESKEY_name='$CHECK_VALUE' differs from expected value '$OCF_RESKEY_inactive_value'" return $OCF_ERR_GENERIC fi return $OCF_NOT_RUNNING fi if [ "$CHECK_VALUE" != "$OCF_RESKEY_active_value" ]; then CHECK_REASON="Node attribute $OCF_RESKEY_name='$CHECK_VALUE' differs from expected value '$OCF_RESKEY_active_value'" return $OCF_ERR_GENERIC fi return $OCF_SUCCESS } monitor() { check_attribute MONITOR_RC=$? if [ $MONITOR_RC -eq $OCF_ERR_GENERIC ]; then ocf_exit_reason "$CHECK_REASON" fi return $MONITOR_RC } start() { check_attribute if [ $? -eq $OCF_SUCCESS ]; then return $OCF_SUCCESS fi touch "${OCF_RESKEY_state}" 2>/dev/null if [ $? -ne 0 ]; then ocf_exit_reason "Unable to manage state file $OCF_RESKEY_state" return $OCF_ERR_GENERIC fi set_attribute "${OCF_RESKEY_active_value}" if [ $? -ne 0 ]; then rm -f "${OCF_RESKEY_state}" ocf_exit_reason "Unable to set node attribute $OCF_RESKEY_name='$OCF_RESKEY_active_value'" return $OCF_ERR_GENERIC fi return $OCF_SUCCESS } stop() { check_attribute if [ $? -eq $OCF_NOT_RUNNING ]; then return $OCF_SUCCESS fi rm -f ${OCF_RESKEY_state} set_attribute "${OCF_RESKEY_inactive_value}" if [ $? -ne 0 ]; then ocf_exit_reason "Unable to set node attribute $OCF_RESKEY_name='$OCF_RESKEY_inactive_value'" return $OCF_ERR_GENERIC fi return $OCF_SUCCESS } case $__OCF_ACTION in meta-data) meta_data ;; start) start ;; stop) stop ;; monitor) monitor ;; # We don't do anything special for live migration, but we support it so that # other resources that live migrate can depend on this one. migrate_to) stop ;; migrate_from) start ;; reload) start ;; validate-all) validate ;; usage|help) usage $OCF_SUCCESS ;; *) usage $OCF_ERR_UNIMPLEMENTED ;; esac exit $? -# vim: expandtab:tabstop=4:softtabstop=4:shiftwidth=4:textwidth=80 +# vim: set filetype=sh expandtab tabstop=4 softtabstop=4 shiftwidth=4 textwidth=80: diff --git a/extra/resources/controld b/extra/resources/controld index 0936a24939..4dd27d42cd 100755 --- a/extra/resources/controld +++ b/extra/resources/controld @@ -1,286 +1,288 @@ #!/bin/sh # # ocf:pacemaker:controld resource agent # -# Copyright 2008-2018 the Pacemaker project contributors +# Copyright 2008-2019 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 # (GPLv2) WITHOUT ANY WARRANTY. # # Manages the DLM controld process # ####################################################################### # Initialization: : ${OCF_FUNCTIONS=${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs} . ${OCF_FUNCTIONS} : ${__OCF_ACTION=$1} ####################################################################### if [ -e "$OCF_ROOT/resource.d/heartbeat/controld" ]; then ocf_log info "Using heartbeat controld agent" $OCF_ROOT/resource.d/heartbeat/controld $1 exit $? fi meta_data() { cat < 1.0 This Resource Agent can control the dlm_controld services needed by cluster-aware file systems. It assumes that dlm_controld is in your default PATH. In most cases, it should be run as an anonymous clone. DLM Agent for cluster file systems Any additional options to start the dlm_controld service with DLM Options The daemon to start - supports gfs_controld and dlm_controld The daemon to start Allow DLM start-up even if STONITH/fencing is disabled in the cluster. Setting this option to true will cause cluster malfunction and hangs on fail-over for DLM clients that require fencing (such as GFS2, OCFS2, and cLVM2). This option is advanced use only. Allow start-up even without STONITH/fencing END } ####################################################################### CONFIGFS_DIR="/sys/kernel/config" DLM_CONFIGFS_DIR="${CONFIGFS_DIR}/dlm" DLM_SYSFS_DIR="/sys/kernel/dlm" controld_usage() { cat <&1) if [ $? -eq 0 ]; then if [ -n "$CUL_TMP" ]; then ocf_log err "Uncontrolled lockspace exists, system must reboot. Executing suicide fencing" stonith_admin --reboot="$(crm_node -n)" --tag controld exit $OCF_ERR_GENERIC fi fi } controld_start() { controld_monitor; rc=$? case $rc in $OCF_SUCCESS) return $OCF_SUCCESS;; $OCF_NOT_RUNNING) ;; *) return $OCF_ERR_GENERIC;; esac # Ensure configfs is mounted if [ ! -e "$CONFIGFS_DIR" ]; then modprobe configfs if [ ! -e "$CONFIGFS_DIR" ]; then ocf_log err "$CONFIGFS_DIR not available" return $OCF_ERR_INSTALLED fi fi mount -t configfs | grep " $CONFIGFS_DIR " >/dev/null 2>/dev/null if [ $? -ne 0 ]; then mount -t configfs none "$CONFIGFS_DIR" fi # Ensure DLM is available if [ ! -e "$DLM_CONFIGFS_DIR" ]; then modprobe dlm if [ ! -e "$DLM_CONFIGFS_DIR" ]; then ocf_log err "$DLM_CONFIGFS_DIR not available" return $OCF_ERR_INSTALLED fi fi if ! ocf_is_true "$OCF_RESKEY_allow_stonith_disabled" && \ ! ocf_is_true "`crm_attribute --type=crm_config --name=stonith-enabled --query --quiet --default=true`"; then ocf_log err "The cluster property stonith-enabled may not be deactivated to use the DLM" return $OCF_ERR_CONFIGURED fi ${OCF_RESKEY_daemon} $OCF_RESKEY_args while true do sleep 1 controld_monitor; rc=$? case $rc in $OCF_SUCCESS) CS_ADDR_LIST="$(cat "${DLM_CONFIGFS_DIR}"/cluster/comms/*/addr_list 2>/dev/null)" if [ $? -eq 0 ] && [ -n "$CS_ADDR_LIST" ]; then return $OCF_SUCCESS fi ;; $OCF_NOT_RUNNING) return $OCF_NOT_RUNNING ;; *) return $OCF_ERR_GENERIC ;; esac ocf_log debug "Waiting for ${OCF_RESKEY_daemon} to be ready" done } controld_stop() { controld_monitor; rc=$? if [ $rc = $OCF_NOT_RUNNING ]; then return $OCF_SUCCESS fi killall -TERM ${OCF_RESKEY_daemon}; rc=$? if [ $rc != 0 ]; then return $OCF_ERR_GENERIC fi rc=$OCF_SUCCESS while [ $rc = $OCF_SUCCESS ]; do controld_monitor; rc=$? sleep 1 done if [ $rc = $OCF_NOT_RUNNING ]; then rc=$OCF_SUCCESS fi return $rc } controld_monitor() { killall -0 ${OCF_RESKEY_daemon} >/dev/null 2>&1 ; CM_RC=$? case $CM_RC in 0) smw=$(dlm_tool status -v | grep "stateful_merge_wait=" | cut -d= -f2) if [ -n "$smw" ] && [ $smw -eq 1 ]; then ocf_log err "DLM status is: stateful_merge_wait" CM_RC=$OCF_ERR_GENERIC elif [ -z "$smw" ] && dlm_tool ls | grep -q "wait fencing" && \ ! stonith_admin -H '*' -V | grep -q "wishes to"; then ocf_log err "DLM status is: wait fencing" CM_RC=$OCF_ERR_GENERIC else CM_RC=$OCF_SUCCESS fi ;; 1) CM_RC=$OCF_NOT_RUNNING;; *) CM_RC=$OCF_ERR_GENERIC;; esac # if the dlm is not successfully running, but # dlm lockspace bits are left over, we self must fence. if [ $CM_RC -ne $OCF_SUCCESS ]; then check_uncontrolled_locks fi return $CM_RC } controld_validate() { check_binary killall check_binary ${OCF_RESKEY_daemon} case ${OCF_RESKEY_CRM_meta_globally_unique} in yes|Yes|true|True|1) ocf_log err "$OCF_RESOURCE_INSTANCE must be configured with the globally_unique=false meta attribute" exit $OCF_ERR_CONFIGURED ;; esac [ -d /var/run/cluster ] || mkdir /var/run/cluster return $OCF_SUCCESS } : ${OCF_RESKEY_sctp=false} : ${OCF_RESKEY_CRM_meta_globally_unique:="false"} case "$OCF_RESOURCE_INSTANCE" in *[gG][fF][sS]*) : ${OCF_RESKEY_args=-g 0} : ${OCF_RESKEY_daemon=gfs_controld} ;; *[dD][lL][mM]*) : ${OCF_RESKEY_args=-s 0} : ${OCF_RESKEY_daemon=dlm_controld} ;; *) : ${OCF_RESKEY_args=-s 0} : ${OCF_RESKEY_daemon=dlm_controld} esac case $__OCF_ACTION in meta-data) meta_data exit $OCF_SUCCESS ;; start) controld_validate; controld_start;; stop) controld_stop;; monitor) controld_validate; controld_monitor;; validate-all) controld_validate;; usage|help) controld_usage exit $OCF_SUCCESS ;; *) controld_usage exit $OCF_ERR_UNIMPLEMENTED ;; esac rc=$? exit $rc + +# vim: set filetype=sh expandtab tabstop=4 softtabstop=4 shiftwidth=4 textwidth=80: diff --git a/extra/resources/ifspeed.in b/extra/resources/ifspeed.in index 49d4c1c3df..d423442b19 100755 --- a/extra/resources/ifspeed.in +++ b/extra/resources/ifspeed.in @@ -1,548 +1,550 @@ #!@BASH_PATH@ # # ocf:pacemaker:ifspeed resource agent # -# Copyright 2011-2018 the Pacemaker project contributors +# Copyright 2011-2019 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. # # # OCF resource agent which monitors state of network interface and records it # as a node attribute in the CIB based on the sum of speeds of its active (up, # link detected, not blocked) underlying interfaces. # # Partially based on 'ping' RA by Andrew Beekhof # # Change on 2017 by Tomer Azran : # Add "ip" parameter to detect network interface name by ip address: # http://lists.clusterlabs.org/pipermail/users/2017-August/006224.html # # OCF instance parameters: # OCF_RESKEY_name: name of attribute to set in CIB # OCF_RESKEY_ip ip address to check # OCF_RESKEY_iface: network interface to monitor # OCF_RESKEY_bridge_ports: if not null and OCF_RESKEY_iface is a bridge, list of # bridge ports to consider. # Default is all ports which have designated_bridge=root_id # OCF_RESKEY_weight_base: Relative weight of 1Gbps. This can be used to tune # value of resulting CIB attribute. # # Initialization: : ${OCF_FUNCTIONS=${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs} # If these aren't available, we can still show help, # which is all that is needed to build the man pages. [ -r "${OCF_FUNCTIONS}" ] && . "${OCF_FUNCTIONS}" [ -r "${OCF_FUNCTIONS_DIR}/findif.sh" ] && . "${OCF_FUNCTIONS_DIR}/findif.sh" : ${OCF_SUCCESS=0} : ${__OCF_ACTION=$1} FINDIF=findif # Defaults OCF_RESKEY_name_default="ifspeed" OCF_RESKEY_bridge_ports_default="detect" OCF_RESKEY_weight_base_default=1000 OCF_RESKEY_dampen_default=5 : ${OCF_RESKEY_name=${OCF_RESKEY_name_default}} : ${OCF_RESKEY_bridge_ports=${OCF_RESKEY_bridge_ports_default}} : ${OCF_RESKEY_weight_base=${OCF_RESKEY_weight_base_default}} : ${OCF_RESKEY_dampen=${OCF_RESKEY_dampen_default}} meta_data() { cat < 1.0 Every time the monitor action is run, this resource agent records (in the CIB) (relative) speed of network interface it monitors. This RA can monitor physical interfaces, bonds, bridges, vlans and (hopefully) any combination of them. Examples: *) Bridge on top of one 10Gbps interface (eth2) and 802.3ad bonding (bond0) built on two 1Gbps interfaces (eth0 and eth1). *) Active-backup bonding built on top of one physical interface and one vlan on another interface. For STP-enabled bridges this RA tries to some-how guess network topology and by default looks only on ports which are connected to upstream switch. This can be overridden by 'bridge_ports' parameter. Active interfaces in this case are those in "forwarding" state. For balancing bonds this RA summs speeds of underlying "up" slave interfaces (and applies coefficient 0.8 to result). For non-balancing bonds ('active-backup' and probably 'broadcast'), only the speed of the currently active slave is used. Network interface speed monitor The name of the attribute to set. This is the name to be used in the constraints. Attribute name Network interface to monitor. Network interface Try to detect interface name by detecting the interface that holds the IP address. The IPv4 (dotted quad notation) or IPv6 address (colon hexadecimal notation) example IPv4 "192.168.1.1". example IPv6 "2001:db8:DC28:0:0:FC57:D4C8:1FFF". IPv4 or IPv6 address If not null and OCF_RESKEY_iface is a bridge, list of bridge ports to consider. Default is all ports which have designated_bridge=root_id. Bridge ports Relative weight of 1Gbps in interface speed. Can be used to tune how big attribute value will be. Weight of 1Gbps The time to wait (dampening) for further changes to occur. Dampening interval Log what have been done more verbosely. Verbose logging END } usage() { cat </dev/null)" if [ -z "$SP_OUT" ] then modprobe -s ocfs2_stack_user if [ $? != 0 ]; then ocf_log err "Could not load ocfs2_stack_user" return $OCF_ERR_INSTALLED fi fi SP_OUT="$(awk '/^'user'$/{print; exit}' "$LOADED_PLUGINS_FILE" 2>/dev/null)" if [ -z "$SP_OUT" ]; then ocf_log err "Switch to userspace stack unsuccessful" return $OCF_ERR_INSTALLED fi if [ -f "$CLUSTER_STACK_FILE" ]; then echo "$OCF_RESKEY_stack" >"$CLUSTER_STACK_FILE" if [ $? != 0 ]; then ocf_log err "Userspace stack '$OCF_RESKEY_stack' not supported" return $OCF_ERR_INSTALLED fi else ocf_log err "Switch to userspace stack not supported" return $OCF_ERR_INSTALLED fi driver_filesystem ocfs2; rc=$? if [ $rc != 0 ]; then modprobe -s ocfs2 if [ "$?" != 0 ]; then ocf_log err "Unable to load ocfs2 module" return $OCF_ERR_INSTALLED fi fi bringup_daemon return $? } o2cb_stop() { o2cb_monitor; rc=$? case $rc in $OCF_NOT_RUNNING) return $OCF_SUCCESS;; esac ocf_log info "Stopping $OCF_RESOURCE_INSTANCE" kill_daemon if [ $? != 0 ]; then ocf_log err "Unable to unload modules: the cluster is still online" return $OCF_ERR_GENERIC fi unload_filesystem ocfs2 if [ $? = 1 ]; then ocf_log err "Unable to unload ocfs2 module" return $OCF_ERR_GENERIC fi # If we can't find the stack glue, we have nothing to do. [ ! -e "$LOADED_PLUGINS_FILE" ] && return $OCF_SUCCESS while read plugin do unload_module "ocfs2_stack_${plugin}" if [ $? = 1 ]; then ocf_log err "Unable to unload ocfs2_stack_${plugin}" return $OCF_ERR_GENERIC fi done <"$LOADED_PLUGINS_FILE" unload_module "ocfs2_stackglue" if [ $? = 1 ]; then ocf_log err "Unable to unload ocfs2_stackglue" return $OCF_ERR_GENERIC fi # Don't unmount configfs - it's always in use by libdlm } o2cb_monitor() { o2cb_validate # Assume that ocfs2_controld will terminate if any of the conditions below are met driver_filesystem configfs; rc=$? if [ $rc != 0 ]; then ocf_log info "configfs not loaded" return $OCF_NOT_RUNNING fi check_filesystem configfs "${OCF_RESKEY_configfs}"; rc=$? if [ $rc != 0 ]; then ocf_log info "configfs not mounted" return $OCF_NOT_RUNNING fi if [ ! -e "$LOADED_PLUGINS_FILE" ]; then ocf_log info "Stack glue driver not loaded" return $OCF_NOT_RUNNING fi grep user "$LOADED_PLUGINS_FILE" >/dev/null 2>&1; rc=$? if [ $rc != 0 ]; then ocf_log err "Wrong stack `cat $LOADED_PLUGINS_FILE`" return $OCF_ERR_INSTALLED fi driver_filesystem ocfs2; rc=$? if [ $rc != 0 ]; then ocf_log info "ocfs2 not loaded" return $OCF_NOT_RUNNING fi status_daemon return $? } o2cb_usage() { echo "usage: $0 {start|stop|monitor|validate-all|meta-data}" echo " Expects to have a fully populated OCF RA-compliant environment set." echo " In particualr, a value for OCF_ROOT" } o2cb_validate() { check_binary ${DAEMON} case ${OCF_RESKEY_CRM_meta_globally_unique} in yes|Yes|true|True|1) ocf_log err "$OCF_RESOURCE_INSTANCE must be configured with the globally_unique=false meta attribute" exit $OCF_ERR_CONFIGURED ;; esac return $OCF_SUCCESS } meta_data() { cat < 1.0 OCFS2 daemon resource agent This Resource Agent controls the userspace daemon needed by OCFS2. Location where sysfs is mounted Sysfs location Location where configfs is mounted Configfs location Which userspace stack to use. Known values: pcmk Userspace stack Number of seconds to allow the control daemon to come up Daemon Timeout END } case $__OCF_ACTION in meta-data) meta_data exit $OCF_SUCCESS ;; start) o2cb_start ;; stop) o2cb_stop ;; monitor) o2cb_monitor ;; validate-all) o2cb_validate ;; usage|help) o2cb_usage exit $OCF_SUCCESS ;; *) o2cb_usage exit $OCF_ERR_UNIMPLEMENTED ;; esac exit $? + +# vim: set filetype=sh expandtab tabstop=4 softtabstop=4 shiftwidth=4 textwidth=80: diff --git a/extra/resources/ping b/extra/resources/ping index 10c272b7b5..ce46d3fe2c 100755 --- a/extra/resources/ping +++ b/extra/resources/ping @@ -1,416 +1,418 @@ #!/bin/sh # # ocf:pacemaker:ping resource agent # # 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. # ####################################################################### # Initialization: : ${OCF_FUNCTIONS=${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs} . ${OCF_FUNCTIONS} : ${__OCF_ACTION=$1} ####################################################################### meta_data() { cat < 1.0 Every time the monitor action is run, this resource agent records (in the CIB) the current number of nodes the host can connect to using the system fping (preferred) or ping tool. node connectivity PID file PID file The time to wait (dampening) further changes occur Dampening interval The name of the attributes to set. This is the name to be used in the constraints. Attribute name The number by which to multiply the number of connected ping nodes by Value multiplier A space separated list of ping nodes to count. Host list Number of ping attempts, per host, before declaring it dead no. of ping attempts How long, in seconds, to wait before declaring a ping lost ping timeout in seconds A catch all for any other options that need to be passed to ping. Extra Options Resource is failed if the score is less than failure_score. Default never fails. failure_score Use fping rather than ping, if found. If set to 0, fping will not be used even if present. Use fping if available Enables to use default attrd_updater verbose logging on every call. Verbose logging END } ####################################################################### ping_conditional_log() { level=$1; shift if [ ${OCF_RESKEY_debug} = "true" ]; then ocf_log $level "$*" fi } ping_usage() { cat <&1`; rc=$? active=`echo "$output" | grep "is alive" | wc -l` case $rc in 0) ;; 1) for h in `echo "$output" | grep "is unreachable" | awk '{print $1}'`; do ping_conditional_log warn "$h is inactive" done ;; *) ocf_log err "Unexpected result for '$cmd' $rc: `echo "$output" | tr '\n' ';'`" ;; esac return $active } ping_check() { active=0 for host in $OCF_RESKEY_host_list; do p_exe=ping case `uname` in Linux) p_args="-n -q -W $OCF_RESKEY_timeout -c $OCF_RESKEY_attempts";; Darwin) p_args="-n -q -t $OCF_RESKEY_timeout -c $OCF_RESKEY_attempts -o";; FreeBSD) p_args="-n -q -t $OCF_RESKEY_timeout -c $OCF_RESKEY_attempts -o";; *) ocf_log err "Unknown host type: `uname`"; exit $OCF_ERR_INSTALLED;; esac case $host in *:*) p_exe=ping6 esac p_out=`$p_exe $p_args $OCF_RESKEY_options $host 2>&1`; rc=$? case $rc in 0) active=`expr $active + 1`;; 1) ping_conditional_log warn "$host is inactive: $p_out";; *) ocf_log err "Unexpected result for '$p_exe $p_args $OCF_RESKEY_options $host' $rc: $p_out";; esac done return $active } ping_update() { if use_fping; then fping_check active=$? else ping_check active=$? fi score=`expr $active \* $OCF_RESKEY_multiplier` if [ "$__OCF_ACTION" = "start" ] ; then attrd_updater -n $OCF_RESKEY_name -B $score -d $OCF_RESKEY_dampen $attrd_options else attrd_updater -n $OCF_RESKEY_name -v $score -d $OCF_RESKEY_dampen $attrd_options fi rc=$? case $rc in 0) ping_conditional_log debug "Updated $OCF_RESKEY_name = $score" ;; *) ocf_log warn "Could not update $OCF_RESKEY_name = $score: rc=$rc";; esac if [ $rc -ne 0 ]; then return $rc fi if [ -n "$OCF_RESKEY_failure_score" ] && [ "$score" -lt "$OCF_RESKEY_failure_score" ]; then ocf_log warn "$OCF_RESKEY_name is less than failure_score($OCF_RESKEY_failure_score)" return 1 fi return 0 } use_fping() { ocf_is_true "$OCF_RESKEY_use_fping" && have_binary fping; } # return values: # 4 IPv4 # 6 IPv6 # 0 indefinite (i.e. hostname) host_family() { case $1 in *[0-9].*[0-9].*[0-9].*[0-9]) return 4 ;; *:*) return 6 ;; *) return 0 ;; esac } # return values same as host_family plus # 99 ambiguous families hosts_family() { # For fping allow only same IP versions or hostnames family=0 for host in $OCF_RESKEY_host_list; do host_family $host f=$? if [ $family -ne 0 ] && [ $f -ne 0 ] && [ $f -ne $family ] ; then family=99 break fi [ $f -ne 0 ] && family=$f done return $family } : ${OCF_RESKEY_name:="pingd"} : ${OCF_RESKEY_dampen:="5s"} : ${OCF_RESKEY_attempts:="3"} : ${OCF_RESKEY_multiplier:="1"} : ${OCF_RESKEY_debug:="false"} : ${OCF_RESKEY_failure_score:="0"} : ${OCF_RESKEY_use_fping:="1"} : ${OCF_RESKEY_CRM_meta_timeout:="20000"} : ${OCF_RESKEY_CRM_meta_globally_unique:="false"} integer=`echo ${OCF_RESKEY_timeout} | egrep -o '[0-9]*'` case ${OCF_RESKEY_timeout} in *[0-9]ms|*[0-9]msec) OCF_RESKEY_timeout=`expr $integer / 1000`;; *[0-9]m|*[0-9]min) OCF_RESKEY_timeout=`expr $integer \* 60`;; *[0-9]h|*[0-9]hr) OCF_RESKEY_timeout=`expr $integer \* 60 \* 60`;; *) OCF_RESKEY_timeout=$integer;; esac if [ -z ${OCF_RESKEY_timeout} ]; then if [ x"$OCF_RESKEY_host_list" != x ]; then host_count=`echo $OCF_RESKEY_host_list | awk '{print NF}'` OCF_RESKEY_timeout=`expr $OCF_RESKEY_CRM_meta_timeout / $host_count / $OCF_RESKEY_attempts` OCF_RESKEY_timeout=`expr $OCF_RESKEY_timeout / 1100` # Convert to seconds and finish 10% early else OCF_RESKEY_timeout=5 fi fi if [ ${OCF_RESKEY_timeout} -lt 1 ]; then OCF_RESKEY_timeout=5 elif [ ${OCF_RESKEY_timeout} -gt 1000 ]; then # ping actually complains if this value is too high, 5 minutes is plenty OCF_RESKEY_timeout=300 fi if [ ${OCF_RESKEY_CRM_meta_globally_unique} = "false" ]; then : ${OCF_RESKEY_pidfile:="${HA_VARRUN%%/}/ping-${OCF_RESKEY_name}"} else : ${OCF_RESKEY_pidfile:="${HA_VARRUN%%/}/ping-${OCF_RESOURCE_INSTANCE}"} fi # Check the debug option case "${OCF_RESKEY_debug}" in true|True|TRUE|1) OCF_RESKEY_debug=true;; false|False|FALSE|0) OCF_RESKEY_debug=false;; *) ocf_log warn "Value for 'debug' is incorrect. Please specify 'true' or 'false' not: ${OCF_RESKEY_debug}" OCF_RESKEY_debug=false ;; esac attrd_options='-q' if [ ${OCF_RESKEY_debug} = "true" ]; then attrd_options='' fi case $__OCF_ACTION in meta-data) meta_data exit $OCF_SUCCESS ;; start) ping_start;; stop) ping_stop;; monitor) ping_monitor;; validate-all) ping_validate;; usage|help) ping_usage exit $OCF_SUCCESS ;; *) ping_usage exit $OCF_ERR_UNIMPLEMENTED ;; esac exit $? + +# vim: set filetype=sh expandtab tabstop=4 softtabstop=4 shiftwidth=4 textwidth=80: diff --git a/extra/resources/pingd b/extra/resources/pingd index 2723ad03c6..f84a772734 100755 --- a/extra/resources/pingd +++ b/extra/resources/pingd @@ -1,187 +1,189 @@ #!/bin/sh # # ocf:pacemaker:pingd resource agent # # Copyright 2006-2019 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 # (GPLv2) WITHOUT ANY WARRANTY. # # # Records (in the CIB) the current number of ping nodes a cluster node can # connect to. # ####################################################################### # Initialization: : ${OCF_FUNCTIONS=${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs} . ${OCF_FUNCTIONS} : ${__OCF_ACTION=$1} : ${OCF_RESKEY_name:="pingd"} : ${OCF_RESKEY_interval:="1"} : ${OCF_RESKEY_CRM_meta_interval:=0} upgrade1="This agent (ocf:pacemaker:pingd) has been replaced by the more reliable ocf:pacemaker:ping." upgrade2="Attempting automated conversion, run 'crm ra info ocf:pacemaker:ping' for all configuration options" upgrade3="You will need to remove the existing resource and replace it with one that uses 'ocf:pacemaker:ping' directly" case $__OCF_ACTION in start|monitor) if [ "x" != "x$OCF_RESKEY_host_list" ]; then ocf_log err "$upgrade1" ocf_log err "$upgrade2" ocf_log err "Automatic conversion to ocf:pacemaker:ping failed: no hosts were configured to check for connectivity" ocf_log err "$upgrade3" exit $OCF_ERR_ARGS fi recurring=`crm configure show $OCF_RESOURCE_INSTANCE | grep "op monitor.*interval=\"[1-9]" | sed s/.*interval=// | awk -F\" '{print $2}' | sed s/.*interval=// | awk -F\" '{print $2}' | sort | head -n 1` if [ -z $recurring ]; then ocf_log err "$upgrade1" ocf_log err "$upgrade2" ocf_log err "Automatic conversion to ocf:pacemaker:ping failed: no monitor operation configured" ocf_log err "Without an explicit monitor operation for '$OCF_RESOURCE_INSTANCE', connectivity changes will not be noticed" ocf_log err "Preventing startup to ensure the issue is addressed before it matters" exit $OCF_ERR_ARGS fi if [ $OCF_RESKEY_CRM_meta_interval = 0 ]; then ocf_log warn "$upgrade1" ocf_log warn "$upgrade2" if [ $recurring != $OCF_RESKEY_interval ]; then ocf_log warn "Your monitor operation happens every $recurring, which means that the $OCF_RESKEY_name attribute will be updated with a different frequency than the previously configured ( $OCF_RESKEY_interval )" ocf_log warn "Either change the monitor interval to match or, ideally, switch to the ocf:pacemaker:ping agent and avoid all this compatibility nonsense." fi fi ;; meta-data) cat < 1.0 This agent (ocf:pacemaker:pingd) has been replaced by the more reliable ocf:pacemaker:ping. It records (in the CIB) the current number of ping nodes (specified in the 'host_list' parameter) a cluster node can connect to. pingd resource agent PID file PID file The user we want to run pingd as The user we want to run pingd as The time to wait (dampening) further changes occur Dampening interval The name of the instance_attributes set to place the value in. Rarely needs to be specified. Set name The name of the attributes to set. This is the name to be used in the constraints. Attribute name The section place the value in. Rarely needs to be specified. Section name The number by which to multiply the number of connected ping nodes by Value multiplier The list of ping nodes to count. Defaults to all configured ping nodes. Rarely needs to be specified. Host list How often, in seconds, to check for node liveliness ping interval in seconds Number of ping attempts, per host, before declaring it dead no. of ping attempts How long, in seconds, to wait before declaring a ping lost ping timeout in seconds A catch all for any other options that need to be passed to pingd. Extra Options END exit $OCF_SUCCESS ;; esac ${OCF_ROOT}/resource.d/pacemaker/ping $1 exit $? + +# vim: set filetype=sh expandtab tabstop=4 softtabstop=4 shiftwidth=4 textwidth=80: diff --git a/extra/resources/remote b/extra/resources/remote index 9112bbfcde..6da7a36f3a 100755 --- a/extra/resources/remote +++ b/extra/resources/remote @@ -1,111 +1,113 @@ #!/bin/sh # # ocf:pacemaker:remote OCF resource agent # # Copyright 2013-2019 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 # (GPLv2) WITHOUT ANY WARRANTY. # # This script provides metadata for Pacemaker's internal remote agent. # Outside of acting as a placeholder so the agent can be indexed, and # providing metadata, this script should never be invoked. The actual # functionality behind the remote connection lives within Pacemaker's # controller daemon. # ####################################################################### # Initialization: : ${OCF_FUNCTIONS=${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs} . ${OCF_FUNCTIONS} : ${__OCF_ACTION=$1} ####################################################################### meta_data() { cat < 1.0 remote resource agent Server location to connect to. This can be an ip address or hostname. Server location tcp port to connect to. tcp port Interval in seconds at which Pacemaker will attempt to reconnect to a remote node after an active connection to the remote node has been severed. When this value is nonzero, Pacemaker will retry the connection indefinitely, at the specified interval. As with any time-based actions, this is not guaranteed to be checked more frequently than the value of the cluster-recheck-interval cluster option. reconnect interval END } ####################################################################### remote_usage() { cat <