diff --git a/extra/resources/ClusterMon.in b/extra/resources/ClusterMon.in index 94c4c92f9b..d3a3b100e4 100755 --- a/extra/resources/ClusterMon.in +++ b/extra/resources/ClusterMon.in @@ -1,268 +1,267 @@ #!@BASH_PATH@ # # ocf:pacemaker:ClusterMon resource agent # # Original copyright 2004 SUSE LINUX AG, Lars Marowsky-Bre # Later changes copyright 2008-2021 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. ####################################################################### # Initialization: : ${OCF_FUNCTIONS:="${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs"} . "${OCF_FUNCTIONS}" : ${__OCF_ACTION:="$1"} # Explicitly list all environment variables used, to make static analysis happy : ${OCF_RESKEY_user:=""} : ${OCF_RESKEY_pidfile:="/tmp/ClusterMon_${OCF_RESOURCE_INSTANCE}.pid"} : ${OCF_RESKEY_update:="15000"} : ${OCF_RESKEY_extra_options:=""} : ${OCF_RESKEY_htmlfile:="/tmp/ClusterMon_${OCF_RESOURCE_INSTANCE}.html"} ####################################################################### 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}" case $? 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 [ $# -gt 0 ]; then false else true fi } ClusterMon_validate() { # Existence of the user if [ -n "$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 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/HealthCPU b/extra/resources/HealthCPU index 55e81ae4b0..cb88a99bab 100755 --- a/extra/resources/HealthCPU +++ b/extra/resources/HealthCPU @@ -1,188 +1,187 @@ #!/bin/sh # # ocf:pacemaker:HealthCPU resource agent # -# Copyright 2004-2019 the Pacemaker project contributors +# Copyright 2004-2022 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 System 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 } ####################################################################### healthcpu_usage() { cat < 1.1 System 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 The path to the smartctl program, used for querying device health. The path to the smartctl program - + The time to wait (dampening) for further changes to occur Dampening interval END } ####################################################################### check_temperature() { if [ $1 -lt ${lower_red_limit} ] ; then ocf_log info "Drive ${DRIVE} ${DEVICE} too cold: ${1} C" attrd_updater -n "#health-smart" -U "red" -d "${OCF_RESKEY_dampen}" return 1 fi if [ $1 -gt ${upper_red_limit} ] ; then ocf_log info "Drive ${DRIVE} ${DEVICE} too hot: ${1} C" attrd_updater -n "#health-smart" -U "red" -d "${OCF_RESKEY_dampen}" return 1 fi if [ $1 -lt ${lower_yellow_limit} ] ; then ocf_log info "Drive ${DRIVE} ${DEVICE} quite cold: ${1} C" attrd_updater -n "#health-smart" -U "yellow" -d "${OCF_RESKEY_dampen}" return 1 fi if [ $1 -gt ${upper_yellow_limit} ] ; then ocf_log info "Drive ${DRIVE} ${DEVICE} quite hot: ${1} C" attrd_updater -n "#health-smart" -U "yellow" -d "${OCF_RESKEY_dampen}" return 1 fi } common_checks() { # Each item in $OCF_RESKEY_drives must have a corresponding item in # $OCF_RESKEY_devices with the device type. Alternately, # $OCF_RESKEY_devices can be empty. drives_len=${#DRIVES[@]} devices_len=${#DEVICES[@]} if [ "${drives_len}" -ne "${devices_len}" ] && [ "${devices_len}" -gt 0 ]; then ocf_log err "OCF_RESKEY_devices must be empty or the same length as OCF_RESKEY_drives." exit $OCF_ERR_ARGS fi # Each item in $OCF_RESKEY_drives must look like a device node. for d in "${DRIVES[@]}"; do if [[ "$d" != /dev/* ]]; then ocf_log err "Device in OCF_RESKEY_devices does not look like a device node: $d" exit $OCF_ERR_ARGS fi done } init_smart() { #Set temperature defaults if [ -z "${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})) for ndx in ${!DRIVES[*]}; do DRIVE=${DRIVES[$ndx]} if [ -n "${OCF_RESKEY_devices}" ]; then DEVICE=${DEVICES[$ndx]} "${OCF_RESKEY_smartctl}" -d "${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 $OCF_ERR_INSTALLED fi else "${OCF_RESKEY_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 < $@ endif + +# Pass correct local values to validate target manually +RNG := /dev/null +OCF_ROOT := /usr/lib/ocf + +.PHONY: validate +validate: all + @cat "$(RNG)" >/dev/null 2>/dev/null || { echo "Must specify valid RNG" && false; } + @[ -x "$(OCF_ROOT)/lib/heartbeat" ] || { echo "Must specify valid OCF_ROOT" && false; } + @for AGENT in $(dist_ocf_SCRIPTS) $(ocf_SCRIPTS); do \ + echo -e "\n\n## $$AGENT:"; \ + OCF_ROOT="$(OCF_ROOT)" "$(builddir)/$$AGENT" meta-data \ + | xmllint --noout --relaxng $(RNG) - || break; \ + done + clean-generic: rm -f $(man7_MANS) $(ocf_SCRIPTS:%=%.xml) $(dist_ocf_SCRIPTS:%=%.xml) *~ diff --git a/extra/resources/SysInfo.in b/extra/resources/SysInfo.in index 6845bd03d2..f8ce6f55d9 100755 --- a/extra/resources/SysInfo.in +++ b/extra/resources/SysInfo.in @@ -1,381 +1,380 @@ #!@BASH_PATH@ # # ocf:pacemaker:SysInfo resource agent # # Original copyright 2004 SUSE LINUX AG, Lars Marowsky-Bre # Later changes copyright 2008-2021 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"} # Explicitly list all environment variables used, to make static analysis happy : ${OCF_RESKEY_CRM_meta_globally_unique:="false"} : ${OCF_RESKEY_pidfile:="${HA_VARRUN%%/}/SysInfo-${OCF_RESOURCE_INSTANCE}"} : ${OCF_RESKEY_disk_unit:="G"} : ${OCF_RESKEY_clone:="0"} : ${OCF_RESKEY_disks:=""} : ${OCF_RESKEY_delay:=""} : ${OCF_RESKEY_mind_disk_free:=""} ####################################################################### 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() { local DISK_STATUS="green" 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 $2"k"}') if [ -n "$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 $7}') fi ;; *) esac if [ -n "$cpu_type" ]; then UpdateStat cpu_info "$cpu_type" fi if [ -n "$cpu_speed" ]; then UpdateStat cpu_speed "$cpu_speed" fi if [ -n "$cpu_cores" ]; then UpdateStat cpu_cores "$cpu_cores" fi if [ -n "$cpu_load" ]; then UpdateStat cpu_load "$cpu_load" fi if [ -n "$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 $4}') if [ -n "$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" ] && [ $disk_free -le $MIN_FREE ]; then DISK_STATUS="red" fi fi done UpdateStat "#health_disk" "$DISK_STATUS" } SysInfo_megabytes() { # Size in megabytes echo $1 | awk '{ n = $0; sub( /[0-9]+(.[0-9]+)?/, "" ); if ( $0 == "" ) { $0 = "G" }; # Do not change previous behavior `if ($0 == "G" || $0 == "") { n *= 1024 };` split( n, a, $0 ); n = a[1]; if ( /^[pP]i?[bB]?/ ) { n *= 1024 * 1024 * 1024 }; if ( /^[tT]i?[bB]?/ ) { n *= 1024 * 1024 }; if ( /^[gG]i?[bB]?/ ) { n *= 1024 }; if ( /^[mM]i?[bB]?/ ) { n *= 1 }; if ( /^[kK]i?[bB]?/ ) { n /= 1024 }; if ( /^[bB]i?/ ) { n /= 1024 * 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 [ $r -ne 0 ]; then mem=$(($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 [ -z "$clone" ]; then rm "$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 $OCF_ERR_ARGS fi if [ -n "${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 480fa79c10..ba69293892 100755 --- a/extra/resources/SystemHealth +++ b/extra/resources/SystemHealth @@ -1,239 +1,243 @@ #!/bin/sh # # ocf:pacemaker:SystemHealth resource agent # -# Copyright 2009-2019 the Pacemaker project contributors +# Copyright 2009-2022 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:="$1"} ####################################################################### meta_data() { cat < - - + 1.0 This is a SystemHealth Resource Agent. It is used to monitor the health of a system via IPMI. SystemHealth resource agent + + Path to notifyServicelogEvent command + Path to notify command + + END } ####################################################################### SystemHealth_usage() { cat < /dev/null 2>&1 RC=$? if [ $RC -ne 0 ]; then ocf_log err "servicelog_notify not found!" return $OCF_ERR_INSTALLED fi which ipmiservicelogd > /dev/null 2>&1 RC=$? if [ $RC -ne 0 ]; then ocf_log err "ipmiservicelogd not found!" return $OCF_ERR_INSTALLED fi test -x "$OCF_RESKEY_program" RC=$? if [ $RC -ne 0 ]; then ocf_log err "$OCF_RESKEY_program not found!" return $OCF_ERR_INSTALLED fi } SystemHealth_start() { SystemHealth_monitor RC=$? if [ $RC -eq $OCF_ERR_GENERIC ]; then return $OCF_ERR_GENERIC elif [ $RC -eq $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 -ne 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 -ne 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 -ne 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 -eq $OCF_ERR_GENERIC ]; then return $OCF_ERR_GENERIC elif [ $RC -eq $OCF_SUCCESS ]; then killall ipmiservicelogd RC1=$? if [ $RC1 -ne 0 ]; then ocf_log err "Could not stop ipmiservicelogd!" fi servicelog_notify --remove --command="$OCF_RESKEY_program" > /dev/null 2>&1 RC2=$? if [ $RC2 -ne 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 -eq $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 -ne 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 -eq 0 ]; then return $OCF_SUCCESS else return $OCF_NOT_RUNNING fi } SystemHealth_validate() { SystemHealth_check_tools RC=$? if [ $RC -ne 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 [ $RC -ne 0 ]; then case "$__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/controld b/extra/resources/controld index 260ba43e54..1fa97f34d9 100755 --- a/extra/resources/controld +++ b/extra/resources/controld @@ -1,298 +1,297 @@ #!/bin/sh # # ocf:pacemaker:controld resource agent # -# Copyright 2008-2021 the Pacemaker project contributors +# Copyright 2008-2022 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"} # Explicitly list all environment variables used, to make static analysis happy : ${OCF_RESKEY_CRM_meta_globally_unique:="false"} : ${OCF_RESKEY_allow_stonith_disabled:="false"} : ${OCF_RESKEY_sctp:="false"} : ${OCF_RESOURCE_INSTANCE:=""} case "$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} : ${OCF_RESKEY_daemon:=dlm_controld} esac ####################################################################### if [ -e "$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 # If no-quorum-policy not set, or not set as freeze, give a warning crm_attribute --type=crm_config --name=no-quorum-policy --query|grep value=freeze >/dev/null 2>/dev/null if [ $? -ne 0 ]; then ocf_log warn "The DLM cluster best practice suggests to set the cluster property \"no-quorum-policy=freeze\"" fi "${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 -eq $OCF_NOT_RUNNING ]; then return $OCF_SUCCESS fi killall -TERM "${OCF_RESKEY_daemon}"; rc=$? if [ $rc -ne 0 ]; then return $OCF_ERR_GENERIC fi rc=$OCF_SUCCESS while [ $rc -eq $OCF_SUCCESS ]; do controld_monitor; rc=$? sleep 1 done if [ $rc -eq $OCF_NOT_RUNNING ]; then rc=$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 "$smw" ] && dlm_tool ls | grep -q "wait fencing" && \ ! stonith_admin -H '*' --output-as xml | grep -q "extended-status=\"pending\""; then ocf_log err "DLM status is: wait fencing" CM_RC=$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 $OCF_SUCCESS } case "$__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 78b1029890..c435891952 100755 --- a/extra/resources/ifspeed.in +++ b/extra/resources/ifspeed.in @@ -1,555 +1,554 @@ #!@BASH_PATH@ # # ocf:pacemaker:ifspeed resource agent # # Copyright 2011-2021 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 # Explicitly list all environment variables used, to make static analysis happy : ${OCF_RESKEY_name:=${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}} : ${OCF_RESKEY_iface:=""} : ${OCF_RESKEY_ip:=""} : ${OCF_RESKEY_debug:="false"} meta_data() { cat < - 1.0 Every time the monitor action is run, this resource agent records (in the CIB) (relative) speed of network interface it monitors. This RA can monitor physical interfaces, bonds, bridges, vlans and (hopefully) any combination of them. Examples: *) Bridge on top of one 10Gbps interface (eth2) and 802.3ad bonding (bond0) built on two 1Gbps interfaces (eth0 and eth1). *) Active-backup bonding built on top of one physical interface and one vlan on another interface. For STP-enabled bridges this RA tries to some-how guess network topology and by default looks only on ports which are connected to upstream switch. This can be overridden by 'bridge_ports' parameter. Active interfaces in this case are those in "forwarding" state. For balancing bonds this RA summs speeds of underlying "up" slave interfaces (and applies coefficient 0.8 to result). For non-balancing bonds ('active-backup' and probably 'broadcast'), only the speed of the currently active slave is used. Network interface speed monitor The name of the attribute to set. This is the name to be used in the constraints. Attribute name Network interface to monitor. Network interface Try to detect interface name by detecting the interface that holds the IP address. The IPv4 (dotted quad notation) or IPv6 address (colon hexadecimal notation) example IPv4 "192.168.1.1". example IPv6 "2001:db8:DC28:0:0:FC57:D4C8:1FFF". IPv4 or IPv6 address If not null and OCF_RESKEY_iface is a bridge, list of bridge ports to consider. Default is all ports which have designated_bridge=root_id. Bridge ports Relative weight of 1Gbps in interface speed. Can be used to tune how big attribute value will be. Weight of 1Gbps The time to wait (dampening) for further changes to occur. Dampening interval Log what have been done more verbosely. Verbose logging END } usage() { cat </dev/null) test -n "$MOUNTOUT" return $? } # # Unload a filesystem driver. # Be careful to notice if the driver is built-in and do nothing. # # 0 is success, 1 is error, 2 is already unloaded. # unload_filesystem() { if [ $# -ne 1 ] || [ -z "$1" ] then ocf_log err "unload_filesystem(): Missing an argument" return 1 fi FSNAME="$1" driver_filesystem "$FSNAME" || return 2 MODOUT=$(awk '$1 ~ /^'$FSNAME'$/{print $1,$3;exit}' < /proc/modules 2>/dev/null) if [ -z "$MODOUT" ]; then # The driver is built in, we can't unload it. return 0 fi case "$MODOUT" in "$FSNAME 0") ;; "$FSNAME "*) # The driver is busy, leave it alone ocf_log err "Module $FSNAME is still in use" return 1 ;; *) ocf_log err "Invalid module parsing! " return 1 ;; esac modprobe -rs "$FSNAME" if [ $? -ne 0 ]; then ocf_log err "Unable to unload module: $FSNAME" return 1 fi return 0 } status_daemon() { PID=$(pidof "$DAEMON") if [ -n "$PID" ]; then return $OCF_SUCCESS fi return $OCF_NOT_RUNNING } bringup_daemon() { if [ ! -e "$DAEMON" ]; then ocf_log err "Required binary not found: $DAEMON" return $OCF_ERR_INSTALLED fi "$DAEMON"; rc=$? if [ $rc -ne 0 ]; then ocf_log err "Could not start $DAEMON" return $OCF_ERR_GENERIC fi sleep 1 COUNT=0 rc=$OCF_NOT_RUNNING while [ $rc -eq $OCF_NOT_RUNNING ]; do COUNT=$(expr $COUNT + 1) if [ $COUNT -gt $OCF_RESKEY_daemon_timeout ]; then ocf_log err "$(basename $DAEMON) did not come up" return $OCF_ERR_GENERIC fi status_daemon; rc=$? sleep 1 done return $rc } kill_daemon() { status_daemon; rc=$? if [ $rc -ne $OCF_SUCCESS ]; then return $rc fi ocf_log info "Stopping $(basename "$DAEMON")" killproc "$DAEMON" while [ $rc -eq $OCF_NOT_RUNNING ]; do sleep 1 status_daemon; rc=$? done return $OCF_SUCCESS } # # Unload a module # 0 is success, 1 is error, 2 is not loaded # unload_module() { if [ $# -lt 1 ] || [ -z "$1" ] then ocf_log err "unload_module(): Requires an argument" return 1 fi MODNAME="$1" MODOUT=$(awk '$1 ~ /^'$MODNAME'$/{print $1,$3;exit}' < /proc/modules 2>/dev/null) if [ -z "$MODOUT" ] then return 2 fi case "$MODOUT" in "$MODNAME 0") ;; "$MODNAME "*) return 2 ;; *) ocf_log err "Invalid module parsing!" return 1 ;; esac modprobe -rs "$MODNAME" if [ $? -ne 0 ]; then ocf_log err "Unable to unload module \"$MODNAME\"" return 1 fi return 0 } o2cb_start() { o2cb_monitor; rc=$? if [ $rc -ne $OCF_NOT_RUNNING ]; then return $rc fi ocf_log info "Starting $OCF_RESOURCE_INSTANCE" if [ ! -e "$CLUSTER_STACK_FILE" ]; then modprobe -s ocfs2_stackglue if [ $? -ne 0 ]; then ocf_log err "Could not load ocfs2_stackglue" return $OCF_ERR_INSTALLED fi fi SP_OUT="$(awk '/^'user'$/{print; exit}' "$LOADED_PLUGINS_FILE" 2>/dev/null)" if [ -z "$SP_OUT" ] then modprobe -s ocfs2_stack_user if [ $? -ne 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 [ $? -ne 0 ]; then ocf_log err "Userspace stack '$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 -ne 0 ]; then modprobe -s ocfs2 if [ $? -ne 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 [ $? -ne 0 ]; then ocf_log err "Unable to unload modules: the cluster is still online" return $OCF_ERR_GENERIC fi unload_filesystem ocfs2 if [ $? -eq 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 [ $? -eq 1 ]; then ocf_log err "Unable to unload ocfs2_stack_${plugin}" return $OCF_ERR_GENERIC fi done <"$LOADED_PLUGINS_FILE" unload_module "ocfs2_stackglue" if [ $? -eq 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 -ne 0 ]; then ocf_log info "configfs not loaded" return $OCF_NOT_RUNNING fi check_filesystem configfs "${OCF_RESKEY_configfs}"; rc=$? if [ $rc -ne 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 -ne 0 ]; then ocf_log err "Wrong stack $(cat $LOADED_PLUGINS_FILE)" return $OCF_ERR_INSTALLED fi driver_filesystem ocfs2; rc=$? if [ $rc -ne 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 $OCF_SUCCESS } meta_data() { cat < - 1.0 This Resource Agent controls the userspace daemon needed by OCFS2. OCFS2 daemon resource agent Location where sysfs is mounted Sysfs location Location where configfs is mounted Configfs location Which userspace stack to use. Known values: pcmk Userspace stack Number of seconds to allow the control daemon to come up Daemon Timeout END } case "$__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: