diff options
| -rwxr-xr-x | packaging/common/rabbitmq-server-ha.ocf | 262 |
1 files changed, 143 insertions, 119 deletions
diff --git a/packaging/common/rabbitmq-server-ha.ocf b/packaging/common/rabbitmq-server-ha.ocf index c68c8c5b0d..35a5a0c701 100755 --- a/packaging/common/rabbitmq-server-ha.ocf +++ b/packaging/common/rabbitmq-server-ha.ocf @@ -1,4 +1,4 @@ -#!/bin/bash +#!/bin/sh # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at @@ -222,7 +222,7 @@ END # Invokes the given command as a rabbitmq user and wrapped in the # timeout command. su_rabbit_cmd() { - local cmd=${1:-status} + local cmd="${1:-status}" local LH="${LL} su_rabbit_cmd():" local rc=1 local user=$OCF_RESKEY_username @@ -244,7 +244,7 @@ now() { master_score() { local score=$1 - if [[ -z $score ]] ; then + if [ -z $score ] ; then score=0 fi ocf_run crm_master -l reboot -v $score || return $OCF_ERR_GENERIC @@ -255,7 +255,7 @@ master_score() { # Otherwise, return 10 my_host() { local hostlist="$1" - local hostname=$(hostname -s) + local hostname="$(hostname -s)" local hn local rc=10 local LH="${LL} my_host():" @@ -264,7 +264,7 @@ my_host() { for host in $hostlist ; do hn=$(echo "$host" | awk -F. '{print $1}') ocf_log debug "${LH} comparing '$hostname' with '$hn'" - if [[ "X${hostname}" == "X${hn}" ]] ; then + if [ "${hostname}" = "${hn}" ] ; then rc=$OCF_SUCCESS break fi @@ -277,7 +277,7 @@ srv_uptime() { local stime stime=$( crm_attribute -N $THIS_PCMK_NODE -l reboot --name 'rabbit-start-time' --query 2>/dev/null | awk '{print $3}' | awk -F "=" '{print $2}' | sed -e '/(null)/d' ) - if [ -z "${stime}" -o x"${stime}" == x"(null)" ] ; then + if [ -z "${stime}" -o "${stime}" = "(null)" ] ; then echo 0 else echo $(( $(now) - ${stime} )) @@ -310,7 +310,8 @@ rmq_setup_env() { # user for dir in ${PID_DIR} "${OCF_RESKEY_mnesia_base}" "${OCF_RESKEY_log_dir}"; do if test -e ${dir}; then - if [ ! -z $(su -s /bin/sh - $OCF_RESKEY_username -c "find ${dir} ! -writable") ]; then + local files=$(su -s /bin/sh - $OCF_RESKEY_username -c "find ${dir} ! -writable") + if [ "${files}" ]; then ocf_log warn "Directory ${dir} is not writable by ${OCF_RESKEY_username}, chowning." chown -R ${OCF_RESKEY_username}:${OCF_RESKEY_groupname} "${dir}" fi @@ -336,17 +337,17 @@ reset_mnesia() { # check status of a beam process get_status rc=$? - if [[ $rc == 0 ]] ; then + if [ $rc -eq 0 ] ; then # beam is running # check status of rabbit app and stop it, if it is running get_status rabbit rc=$? - if [[ $rc == 0 ]] ; then + if [ $rc -eq 0 ] ; then # rabbit app is running, have to stop it ocf_log info "${LH} Stopping RMQ-app prior to reset the mnesia." stop_rmq_server_app rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then ocf_log warn "${LH} RMQ-app can't be stopped." make_amnesia=true fi @@ -357,11 +358,11 @@ reset_mnesia() { ocf_log info "${LH} Execute reset with timeout: ${TIMEOUT_ARG}" su_rabbit_cmd "${OCF_RESKEY_ctl} reset" rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then ocf_log info "${LH} Execute force_reset with timeout: ${TIMEOUT_ARG}" su_rabbit_cmd "${OCF_RESKEY_ctl} force_reset" rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then ocf_log warn "${LH} Mnesia couldn't cleaned, even by force-reset command." make_amnesia=true fi @@ -389,8 +390,8 @@ block_client_access() # do not add temporary RMQ blocking rule, if it is already exist # otherwise, try to add a blocking rule with max of 5 retries local tries=5 - until $(iptables -nvL | grep -q 'temporary RMQ block') || [[ $tries -eq 0 ]]; do - ((tries--)) + until $(iptables -nvL | grep -q 'temporary RMQ block') || [ $tries -eq 0 ]; do + tries=$((tries-1)) iptables -I INPUT -p tcp -m tcp --dport ${OCF_RESKEY_node_port} -m state --state NEW,RELATED,ESTABLISHED \ -m comment --comment 'temporary RMQ block' -j REJECT --reject-with tcp-reset sleep 1 @@ -414,22 +415,23 @@ unblock_client_access() get_nodes__base(){ local infotype='' local rc=$OCF_ERR_GENERIC + local c_status - if [ "$1" == 'nodes' ] + if [ "$1" = 'nodes' ] then infotype='db_nodes' - elif [ "$1" == 'running' ] + elif [ "$1" = 'running' ] then infotype='running_db_nodes' fi - local c_status=$(${OCF_RESKEY_ctl} eval "mnesia:system_info(${infotype})." 2>/dev/null) + c_status=`${OCF_RESKEY_ctl} eval "mnesia:system_info(${infotype})." 2>/dev/null` rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then echo '' return $OCF_ERR_GENERIC fi # translate line like '{running_nodes,['rabbit@node-1','rabbit@node-2','rabbit@node-3']},' to node_list - echo $(echo "${c_status}" | grep "${cl}" | awk -F, '{ for (i=1;i<=NF;i++) { if ($i ~ /@/) { gsub(/[\[\]}{]/,"",$i); print $i; } }}' | tr -d "\'") + echo $(echo "${c_status}" | awk -F, '{ for (i=1;i<=NF;i++) { if ($i ~ /@/) { gsub(/[\[\]}{]/,"",$i); print $i; } }}' | tr -d "\'") return $OCF_SUCCESS } @@ -453,7 +455,7 @@ get_all_pacemaker_nodes() # Get alive cluster nodes in visible partition, but the specified one get_alive_pacemaker_nodes_but() { - if [ -z $1 ]; then + if [ -z "$1" ]; then echo `crm_node -l -p | sed -e '/(null)/d'` else echo `crm_node -l -p | sed -e "s/${1}//g" | sed -e '/(null)/d'` @@ -462,14 +464,16 @@ get_alive_pacemaker_nodes_but() } check_need_join_to() { - local join_to=$(rabbit_node_name $1) + local join_to local node - local running_nodes=$(get_running_nodes) + local running_nodes local rc=$OCF_ERR_GENERIC rc=0 + join_to=$(rabbit_node_name $1) + running_nodes=$(get_running_nodes) for node in $running_nodes ; do - if [[ ${join_to} == ${node} ]] ; then + if [ "${join_to}" = "${node}" ] ; then rc=1 break fi @@ -480,7 +484,7 @@ check_need_join_to() { # Update erlang cookie, if it has been specified update_cookie() { - if [[ "${OCF_RESKEY_erlang_cookie}" != false ]] ; then + if [ "${OCF_RESKEY_erlang_cookie}" != 'false' ] ; then echo "${OCF_RESKEY_erlang_cookie}" > "${OCF_RESKEY_erlang_cookie_file}" && \ chown ${OCF_RESKEY_username}:${OCF_RESKEY_groupname} "${OCF_RESKEY_erlang_cookie_file}" && \ chmod 600 "${OCF_RESKEY_erlang_cookie_file}" @@ -492,13 +496,13 @@ kill_rmq_and_remove_pid() { local pid local LH="${LL} kill_rmq_and_remove_pid():" - if [[ -f $OCF_RESKEY_pid_file ]] ; then + if [ -f "${OCF_RESKEY_pid_file}" ] ; then pid=$(cat $OCF_RESKEY_pid_file) - if [[ -z ${pid} ]] ; then + if [ -z "${pid}" ] ; then ocf_log err "${LH} pidfile is empty, cannot kill by unknown PID! Try to stop it manually!" fi # todo: check content for digital - if [[ -d /proc/${pid}/ ]] ; then + if [ -d "/proc/${pid}/" ] ; then ocf_run kill -9 $pid ocf_log warn "${LH} RMQ-runtime (beam) PID=${pid} stopped by 'kill -9', sorry..." fi @@ -537,20 +541,21 @@ action_validate() { join_to_cluster() { local node="$1" - local rmq_node=$(rabbit_node_name $node) + local rmq_node local rc=$OCF_ERR_GENERIC local LH="${LL} join_to_cluster():" ocf_log info "${LH} start." ocf_log info "${LH} Joining to cluster by node '${rmq_node}'." + rmq_node=$(rabbit_node_name $node) get_status rabbit rc=$? - if [[ $rc == $OCF_SUCCESS ]] ; then + if [ $rc -eq $OCF_SUCCESS ] ; then ocf_log info "${LH} rabbitmq app will be stopped." stop_rmq_server_app rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then ocf_log err "${LH} Can't stop rabbitmq app by stop_app command. Stopping." action_stop return $OCF_ERR_GENERIC @@ -559,7 +564,7 @@ join_to_cluster() { ocf_log info "${LH} Execute join_cluster with timeout: ${TIMEOUT_ARG}" su_rabbit_cmd "${OCF_RESKEY_ctl} join_cluster $rmq_node" rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then ocf_log err "${LH} Can't join to cluster by node '${rmq_node}'. Stopping." action_stop return $OCF_ERR_GENERIC @@ -567,7 +572,7 @@ join_to_cluster() { sleep 2 try_to_start_rmq_app rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then ocf_log err "${LH} Can't start RMQ app after join to cluster. Stopping." action_stop return $OCF_ERR_GENERIC @@ -594,7 +599,7 @@ unjoin_nodes_from_cluster() { nodes_in_cluster=$(get_nodes) rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then # no nodes in node list, nothing to do return $OCF_SUCCESS fi @@ -603,15 +608,15 @@ unjoin_nodes_from_cluster() { # before to unjoin the nodes, make sure they were disconnected from *this* node for hostname in $nodelist ; do nodename=$(rabbit_node_name $hostname) - if [[ "$nodename" == "$RABBITMQ_NODENAME" ]] ; then + if [ "${nodename}" = "${RABBITMQ_NODENAME}" ] ; then continue fi for rnode in $nodes_in_cluster ; do - if [[ "$nodename" == "$rnode" ]] ; then + if [ "${nodename}" = "${rnode}" ] ; then # disconnect node being unjoined from this node ocf_run ${OCF_RESKEY_ctl} eval "disconnect_node(list_to_atom(\"${nodename}\"))." 2>&1 rc=$? - if [[ $rc == $OCF_SUCCESS ]] ; then + if [ $rc -eq $OCF_SUCCESS ] ; then ocf_log info "${LH} node '${nodename}' disconnected succesfully." else ocf_log info "${LH} disconnecting node '${nodename}' failed." @@ -622,7 +627,7 @@ unjoin_nodes_from_cluster() { # remains 'running' for a while, so few retries are required local tries=0 until [ $tries -eq 5 ]; do - ((tries++)) + tries=$((tries+1)) if get_running_nodes | grep -q $(rabbit_node_name $nodename) then ocf_log info "${LH} the ${nodename} is alive and cannot be kicked from the cluster yet" @@ -632,7 +637,7 @@ unjoin_nodes_from_cluster() { ocf_log info "${LH} Execute forget_cluster_node with timeout: ${TIMEOUT_ARG}" su_rabbit_cmd "${OCF_RESKEY_ctl} forget_cluster_node ${nodename}" rc=$? - if [[ $rc == 0 ]] ; then + if [ $rc -eq 0 ] ; then ocf_log info "${LH} node '${nodename}' unjoined succesfully." else ocf_log warn "${LH} unjoining node '${nodename}' failed." @@ -651,11 +656,11 @@ stop_server_process() { pid=$(cat ${OCF_RESKEY_pid_file}) rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then ocf_log err "${LH} RMQ-server process PIDFILE was not found!" su_rabbit_cmd "${OCF_RESKEY_ctl} stop 2>&1 >> \"${OCF_RESKEY_log_dir}/shutdown_log\"" rc=$? - if [[ $rc == 0 ]] ; then + if [ $rc -eq 0 ] ; then ocf_log info "${LH} RMQ-server process stopped succesfully, although there was no PIDFILE found." return $OCF_SUCCESS else @@ -664,7 +669,7 @@ stop_server_process() { fi fi - if [[ -z ${pid} ]] ; then + if [ -z "${pid}" ] ; then kill_rmq_and_remove_pid return $OCF_ERR_GENERIC fi @@ -672,7 +677,7 @@ stop_server_process() { ocf_log info "${LH} Execute stop with timeout: ${TIMEOUT_ARG}" su_rabbit_cmd "${OCF_RESKEY_ctl} stop ${OCF_RESKEY_pid_file} 2>&1 >> \"${OCF_RESKEY_log_dir}/shutdown_log\"" rc=$? - if [[ $rc == 0 ]] ; then + if [ $rc -eq 0 ] ; then ocf_log info "${LH} RMQ-server process (PID=${pid}) stopped succesfully." fi @@ -688,7 +693,7 @@ stop_rmq_server_app() { # if the beam process isn't running, then rabbit app is stopped as well get_status rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then return $OCF_SUCCESS fi @@ -696,14 +701,14 @@ stop_rmq_server_app() { ocf_log info "${LH} Execute stop_app with timeout: ${TIMEOUT_ARG}" su_rabbit_cmd "${OCF_RESKEY_ctl} stop_app 2>&1 >> \"${OCF_RESKEY_log_dir}/shutdown_log\"" rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then ocf_log err "${LH} RMQ-server app cannot be stopped." return $OCF_ERR_GENERIC fi get_status rabbit rc=$? - if [[ $rc != $OCF_SUCCESS ]] ; then + if [ $rc -ne $OCF_SUCCESS ] ; then ocf_log info "${LH} RMQ-server app stopped succesfully." rc=$OCF_SUCCESS else @@ -715,6 +720,7 @@ stop_rmq_server_app() { } start_beam_process() { + local command local rc=$OCF_ERR_GENERIC local ts_end local pf_end @@ -722,13 +728,13 @@ start_beam_process() { local LH="${LL} start_beam_process():" # remove old PID-file if it exists - if [[ -f $OCF_RESKEY_pid_file ]] ; then + if [ -f "${OCF_RESKEY_pid_file}" ] ; then ocf_log warn "${LH} found old PID-file '${OCF_RESKEY_pid_file}'." pid=$(cat ${OCF_RESKEY_pid_file}) - if [[ -d /proc/${pid} && ! -z ${pid} ]] ; then + if [ "${pid}" -a -d "/proc/${pid}" ] ; then ocf_run cat /proc/${pid}/cmdline | grep -c 'bin/beam' 2>&1 > /dev/null rc=$? - if [[ $rc == $OCF_SUCCESS ]] ; then + if [ $rc -eq $OCF_SUCCESS ] ; then ocf_log warn "${LH} found beam process with PID=${pid}, killing...'." ocf_run kill -9 $pid else @@ -742,7 +748,7 @@ start_beam_process() { [ -f /etc/default/rabbitmq-server ] && . /etc/default/rabbitmq-server # run beam process - local command="${OCF_RESKEY_binary} >> \"${OCF_RESKEY_log_dir}/startup_log\" 2>/dev/null" + command="${OCF_RESKEY_binary} >> \"${OCF_RESKEY_log_dir}/startup_log\" 2>/dev/null" RABBITMQ_NODE_ONLY=1 su rabbitmq -s /bin/sh -c "${command}"& ts_end=$(( $(now) + ${OCF_RESKEY_start_time} )) rc=$OCF_ERR_GENERIC @@ -752,20 +758,20 @@ start_beam_process() { pf_end=$(( $(now) + 3 )) while [ $(now) -lt ${pf_end} ]; do # waiting for OCF_RESKEY_pid_file of beam process - if [[ -f $OCF_RESKEY_pid_file ]] ; then + if [ -f "${OCF_RESKEY_pid_file}" ] ; then pid=$(cat ${OCF_RESKEY_pid_file}) break fi sleep 1 done - if [[ $pid != 0 && -d /proc/${pid} ]] ; then + if [ "${pid}" != "0" -a -d "/proc/${pid}" ] ; then rc=$OCF_SUCCESS break fi sleep 2 done - if [[ $rc != $OCF_SUCCESS ]]; then - if [[ "${pid}" == "0" ]] ; then + if [ $rc -ne $OCF_SUCCESS ]; then + if [ "${pid}" = "0" ] ; then ocf_log warn "${LH} PID-file '${OCF_RESKEY_pid_file}' not found" fi ocf_log err "${LH} RMQ-runtime (beam) didn't start succesfully (rc=${rc})." @@ -785,7 +791,8 @@ check_plugins() { load_plugins() { check_plugins - if [[ $? == 0 ]] ; then + local rc=$? + if [ $rc -eq 0 ] ; then return 0 else ${OCF_RESKEY_ctl} eval 'ToBeLoaded = rabbit_plugins:setup(), ok = app_utils:load_applications(ToBeLoaded), StartupApps = app_utils:app_dependency_order(ToBeLoaded,false), app_utils:start_applications(StartupApps).' @@ -794,8 +801,9 @@ load_plugins() { } list_active_plugins() { - local LIST=`${OCF_RESKEY_ctl} eval 'rabbit_plugins:active().'` - echo "${LIST}" + local list + list=`${OCF_RESKEY_ctl} eval 'rabbit_plugins:active().'` + echo "${list}" } try_to_start_rmq_app() { @@ -805,18 +813,18 @@ try_to_start_rmq_app() { get_status rc=$? - if [[ $rc != $OCF_SUCCESS ]] ; then + if [ $rc -ne $OCF_SUCCESS ] ; then ocf_log info "${LH} RMQ-runtime (beam) not started, starting..." start_beam_process rc=$? - if [[ $rc != $OCF_SUCCESS ]]; then + if [ $rc -ne $OCF_SUCCESS ]; then ocf_log err "${LH} Failed to start beam - returning from the function" return $OCF_ERR_GENERIC fi fi - if [[ -z $startup_log ]] ; then + if [ -z "${startup_log}" ] ; then startup_log="${OCF_RESKEY_log_dir}/startup_log" fi @@ -824,12 +832,12 @@ try_to_start_rmq_app() { ocf_log info "${LH} Execute start_app with timeout: ${TIMEOUT_ARG}" su_rabbit_cmd "${OCF_RESKEY_ctl} start_app >>${startup_log} 2>&1" rc=$? - if [[ $rc == 0 ]] ; then + if [ $rc -eq 0 ] ; then ocf_log info "${LH} start_app was successful." ocf_log info "${LH} waiting for start to finish with timeout: ${TIMEOUT_ARG}" su_rabbit_cmd "${OCF_RESKEY_ctl} wait ${OCF_RESKEY_pid_file}" rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then ocf_log err "${LH} RMQ-server app failed to wait for start." return $OCF_ERR_GENERIC fi @@ -838,9 +846,10 @@ try_to_start_rmq_app() { ocf_log info "${LH} start plugins." load_plugins local mrc=$? - if [[ $mrc == 0 ]] ; then - local MLIST=`list_active_plugins` - ocf_log info "${LH} Starting plugins: $MLIST" + if [ $mrc -eq 0 ] ; then + local mlist + mlist=`list_active_plugins` + ocf_log info "${LH} Starting plugins: ${mlist}" else ocf_log info "${LH} Starting plugins: failed." fi @@ -869,7 +878,7 @@ start_rmq_server_app() { # Apply the blocking rule block_client_access rc=$? - if [[ $rc == $OCF_SUCCESS ]]; then + if [ $rc -eq $OCF_SUCCESS ]; then ocf_log info "${LH} blocked access to RMQ port" else ocf_log err "${LH} cannot block access to RMQ port!" @@ -877,11 +886,11 @@ start_rmq_server_app() { fi get_status rc=$? - if [[ $rc != $OCF_SUCCESS ]] ; then + if [ $rc -ne $OCF_SUCCESS ] ; then ocf_log info "${LH} RMQ-runtime (beam) not started, starting..." start_beam_process rc=$? - if [[ $rc != $OCF_SUCCESS ]]; then + if [ $rc -ne $OCF_SUCCESS ]; then unblock_client_access ocf_log info "${LH} unblocked access to RMQ port" return $OCF_ERR_GENERIC @@ -891,12 +900,12 @@ start_rmq_server_app() { ocf_log info "${LH} RMQ-server app not started, starting..." try_to_start_rmq_app "$startup_log" rc=$? - if [[ $rc == $OCF_SUCCESS ]] ; then + if [ $rc -eq $OCF_SUCCESS ] ; then # rabbitmq-server started successfuly as master of cluster master_score 1 # minimal positive master-score for this node. stop_rmq_server_app rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then ocf_log err "${LH} RMQ-server app can't be stopped. Beam will be killed." kill_rmq_and_remove_pid unblock_client_access @@ -906,15 +915,15 @@ start_rmq_server_app() { else # error at start RMQ-server ocf_log warn "${LH} RMQ-server app can't start without Mnesia cleaning." - for ((a=10; a > 0 ; a--)) ; do + for a in $(seq 1 10) ; do rc=$OCF_ERR_GENERIC reset_mnesia || break try_to_start_rmq_app "$startup_log" rc=$? - if [[ $rc == $OCF_SUCCESS ]]; then + if [ $rc -eq $OCF_SUCCESS ]; then stop_rmq_server_app rc=$? - if [[ $rc == $OCF_SUCCESS ]]; then + if [ $rc -eq $OCF_SUCCESS ]; then ocf_log info "${LH} RMQ-server app Mnesia cleaned successfully." rc=$OCF_SUCCESS master_score 1 @@ -929,7 +938,7 @@ start_rmq_server_app() { fi done fi - if [[ $rc == $OCF_ERR_GENERIC ]] ; then + if [ $rc -eq $OCF_ERR_GENERIC ] ; then ocf_log err "${LH} RMQ-server can't be started while many tries. Beam will be killed." kill_rmq_and_remove_pid fi @@ -949,11 +958,11 @@ get_status() { body=$( ${COMMAND_TIMEOUT} ${OCF_RESKEY_ctl} eval 'rabbit_misc:which_applications().' 2>&1 ) rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then return $OCF_NOT_RUNNING fi - if [[ ! -z $what ]] ; then + if [ "${what}" ] ; then rc=$OCF_NOT_RUNNING echo "$body" | grep "\{${what}," 2>&1 > /dev/null && rc=$OCF_SUCCESS fi @@ -975,7 +984,7 @@ is_master() { local result result=`crm_attribute -N "${1}" -l reboot --name 'rabbit-master' --query 2>/dev/null |\ awk '{print $3}' | awk -F "=" '{print $2}' | sed -e '/(null)/d'` - if [[ "${result}" != "true" ]] ; then + if [ "${result}" != 'true' ] ; then return 1 fi return 0 @@ -984,10 +993,11 @@ is_master() { get_monitor() { local rc=$OCF_ERR_GENERIC - local scope local LH="${LL} get_monitor():" local status_master local rabbit_running + local name + local node local nodelist local prev_rc local max @@ -998,12 +1008,12 @@ get_monitor() { ocf_log info "${LH} CHECK LEVEL IS: ${OCF_CHECK_LEVEL}" get_status rc=$? - if [[ $rc == $OCF_NOT_RUNNING ]] ; then + if [ $rc -eq $OCF_NOT_RUNNING ] ; then ocf_log info "${LH} get_status() returns ${rc}." ocf_log info "${LH} ensuring this slave does not get promoted." master_score 0 return $OCF_NOT_RUNNING - elif [[ $rc == $OCF_SUCCESS ]] ; then + elif [ $rc -eq $OCF_SUCCESS ] ; then ocf_log info "${LH} get_status() returns ${rc}." ocf_log info "${LH} also checking if we are master." get_status rabbit @@ -1020,7 +1030,7 @@ get_monitor() { rabbit_running=$? ocf_log info "${LH} checking if rabbit app is running" - if [ $rabbit_running == $OCF_SUCCESS ] + if [ $rabbit_running -eq $OCF_SUCCESS ] then ocf_log info "${LH} rabbit app is running. checking if we are the part of healthy cluster" prev_rc=$rc @@ -1067,7 +1077,7 @@ get_monitor() { fi fi - if [[ $rc == $OCF_ERR_GENERIC ]]; then + if [ $rc -eq $OCF_ERR_GENERIC ]; then ocf_log err "${LH} get_status() returns generic error ${rc}" ocf_log info "${LH} ensuring this slave does not get promoted." master_score 0 @@ -1080,7 +1090,7 @@ get_monitor() { for node in $nodelist do node_start_time=`crm_attribute -N $node -l reboot --name 'rabbit-start-time' --query 2>/dev/null | awk '{print $3}' | awk -F "=" '{print $2}' | sed -e '/(null)/d'` - if [ -z "${node_start_time}" -o x"${node_start_time}" == x"(null)" ] ; then + if [ -z "${node_start_time}" -o "${node_start_time}" = "(null)" ] ; then node_uptime=0 else node_uptime=$(( $(now) - ${node_start_time} )) @@ -1115,13 +1125,29 @@ get_monitor() { # Check if the rabbitmqctl control plane is alive. # The rabbit app may be not running and the command # will return > 0, so we only check if the command execution - # has timed out (which is a code 137) + # has timed out (which is a code 137 or 124) su_rabbit_cmd "${OCF_RESKEY_ctl} list_channels 2>&1 > /dev/null" - rc2=$? - if [ $rc2 -eq 137 -o $rc2 -eq 124 ]; then + local rc_alive=$? + if [ $rc_alive -eq 137 -o $rc_alive -eq 124 ]; then ocf_log err "${LH} rabbitmqctl is not responding. The resource is failed." return $OCF_ERR_GENERIC fi + + # Check if the list of all queues is available, + # Skip the check if rabbit app is not running yet. + su_rabbit_cmd "${OCF_RESKEY_ctl} -q list_queues" + local rc_queues=$? + + # If the rabbit app is running, + # we have to additionally check here if the channels/queues list results were ok. + if [ $rabbit_running -eq $OCF_SUCCESS ]; then + # Check if the rabbitmqctl control plane returned no errors for issued requests. + if [ $rc_alive -ne 0 -o $rc_queues -ne 0 ]; then + ocf_log err "${LH} rabbitmqctl exited with errors." + rc=$OCF_ERR_GENERIC + fi + fi + ocf_log info "${LH} get_monitor function ready to return ${rc}" return $rc } @@ -1131,7 +1157,7 @@ action_monitor() { local rc=$OCF_ERR_GENERIC local LH="${LL} monitor:" ocf_log debug "${LH} action start." - if [[ "${OCF_RESKEY_debug}" == "true" ]] ; then + if [ "${OCF_RESKEY_debug}" = 'true' ] ; then d=`date '+%Y%m%d %H:%M:%S'` echo $d >> /tmp/rmq-monitor.log env >> /tmp/rmq-monitor.log @@ -1148,11 +1174,9 @@ action_monitor() { action_start() { local rc=$OCF_ERR_GENERIC - local msg - local master_node local LH="${LL} start:" - if [[ ${OCF_RESKEY_debug} == "true" ]] ; then + if [ "${OCF_RESKEY_debug}" = 'true' ] ; then d=`date '+%Y%m%d %H:%M:%S'` echo $d >> /tmp/rmq-start.log env >> /tmp/rmq-start.log @@ -1163,7 +1187,7 @@ action_start() { get_status rc=$? - if [[ $rc == $OCF_SUCCESS ]] ; then + if [ $rc -eq $OCF_SUCCESS ] ; then ocf_log warn "${LH} RMQ-runtime (beam) already started." return $OCF_SUCCESS fi @@ -1171,7 +1195,7 @@ action_start() { ocf_log info "${LH} RMQ going to start." start_rmq_server_app rc=$? - if [[ $rc == $OCF_SUCCESS ]] ; then + if [ $rc -eq $OCF_SUCCESS ] ; then ocf_log info "${LH} RMQ prepared for start succesfully." fi @@ -1184,7 +1208,7 @@ action_stop() { local rc=$OCF_ERR_GENERIC local LH="${LL} stop:" - if [[ ${OCF_RESKEY_debug} == "true" ]] ; then + if [ "${OCF_RESKEY_debug}" = 'true' ] ; then d=$(date '+%Y%m%d %H:%M:%S') echo $d >> /tmp/rmq-stop.log env >> /tmp/rmq-stop.log @@ -1210,7 +1234,7 @@ action_stop() { ocf_log info "${LH} action end." get_status rc=$? - if [[ $rc == $OCF_NOT_RUNNING ]] ; then + if [ $rc -eq $OCF_NOT_RUNNING ] ; then ocf_log info "${LH} RMQ-runtime (beam) not running." return $OCF_SUCCESS else @@ -1234,11 +1258,11 @@ jjj_join () { # Check whether we are joining to ourselves # or master host is not given - if [[ $rc != 0 && $join_to != '' ]] ; then + if [ $rc -ne 0 -a "${join_to}" ] ; then ocf_log info "${LH} Joining to cluster by node '${join_to}'" join_to_cluster "${join_to}" rc=$? - if [[ $rc != $OCF_SUCCESS ]] ; then + if [ $rc -ne $OCF_SUCCESS ] ; then ocf_log err "${LH} Failed to join the cluster. The mnesia will be reset." reset_mnesia rc=$OCF_ERR_GENERIC @@ -1254,21 +1278,21 @@ action_notify() { local LH="${LL} notify:" local nodelist - if [[ ${OCF_RESKEY_debug} == "true" ]] ; then + if [ "${OCF_RESKEY_debug}" = 'true' ] ; then d=`date '+%Y%m%d %H:%M:%S'` echo $d >> /tmp/rmq-notify.log env >> /tmp/rmq-notify.log echo "$d [notify] ${OCF_RESKEY_CRM_meta_notify_type}-${OCF_RESKEY_CRM_meta_notify_operation} promote='${OCF_RESKEY_CRM_meta_notify_promote_uname}' demote='${OCF_RESKEY_CRM_meta_notify_demote_uname}' master='${OCF_RESKEY_CRM_meta_notify_master_uname}' slave='${OCF_RESKEY_CRM_meta_notify_slave_uname}' start='${OCF_RESKEY_CRM_meta_notify_start_uname}' stop='${OCF_RESKEY_CRM_meta_notify_stop_uname}' active='${OCF_RESKEY_CRM_meta_notify_active_uname}' inactive='${OCF_RESKEY_CRM_meta_notify_inactive_uname}'" >> /tmp/rmq-ocf.log fi - if [[ ${OCF_RESKEY_CRM_meta_notify_type} == 'pre' ]] ; then + if [ "${OCF_RESKEY_CRM_meta_notify_type}" = 'pre' ] ; then # PRE- anything notify section case "$OCF_RESKEY_CRM_meta_notify_operation" in promote) ocf_log info "${LH} pre-promote begin." my_host "$OCF_RESKEY_CRM_meta_notify_promote_uname" rc=$? - if [[ $rc == $OCF_SUCCESS ]] ; then + if [ $rc -eq $OCF_SUCCESS ] ; then nodelist=$(get_all_pacemaker_nodes) for i in $nodelist do @@ -1282,13 +1306,13 @@ action_notify() { esac fi - if [[ ${OCF_RESKEY_CRM_meta_notify_type} == 'post' ]] ; then + if [ "${OCF_RESKEY_CRM_meta_notify_type}" = 'post' ] ; then # POST- anything notify section case "$OCF_RESKEY_CRM_meta_notify_operation" in promote) ocf_log info "${LH} post-promote begin." # Report not running, if the list of nodes being promoted reported empty - if [ -z ${OCF_RESKEY_CRM_meta_notify_promote_uname} ] ; then + if [ -z "${OCF_RESKEY_CRM_meta_notify_promote_uname}" ] ; then ocf_log warn "${LH} there are no nodes to join to reported on post-promote. The resource will be restarted." ocf_log info "${LH} post-promote end." return $OCF_NOT_RUNNING @@ -1296,10 +1320,10 @@ action_notify() { # Note, this should fail when the mnesia is inconsistent. # For example, when the "old" master processing the promition of the new one. # Later this ex-master node will rejoin the cluster at post-start. - jjj_join ${OCF_RESKEY_CRM_meta_notify_promote_uname} + jjj_join "${OCF_RESKEY_CRM_meta_notify_promote_uname}" rc=$? ocf_log info "${LH} post-promote end." - if [[ $rc == $OCF_ERR_GENERIC ]] ; then + if [ $rc -eq $OCF_ERR_GENERIC ] ; then ocf_log err "${LH} Failed to join the cluster on post-promote. The resource will be restarted." return $OCF_NOT_RUNNING fi @@ -1317,24 +1341,24 @@ action_notify() { my_host "${nodes_list}" rc=$? # Report not running, if there is no master reported - if [ -z ${OCF_RESKEY_CRM_meta_notify_master_uname} ] ; then + if [ -z "${OCF_RESKEY_CRM_meta_notify_master_uname}" ] ; then ocf_log warn "${LH} there are no nodes to join to reported on post-start. The resource will be restarted." ocf_log info "${LH} post-start end." return $OCF_NOT_RUNNING fi - if [[ $rc == $OCF_SUCCESS ]] ; then - check_need_join_to ${OCF_RESKEY_CRM_meta_notify_master_uname} + if [ $rc -eq $OCF_SUCCESS ] ; then + check_need_join_to "${OCF_RESKEY_CRM_meta_notify_master_uname}" rc_join=$? - if [[ ${rc_join} == $OCF_SUCCESS ]]; then + if [ $rc_join -eq $OCF_SUCCESS ]; then ocf_log warn "${LH} Going to join node ${OCF_RESKEY_CRM_meta_notify_master_uname}" - jjj_join ${OCF_RESKEY_CRM_meta_notify_master_uname} + jjj_join "${OCF_RESKEY_CRM_meta_notify_master_uname}" rc2=$? else ocf_log warn "${LH} We are already clustered with node ${OCF_RESKEY_CRM_meta_notify_master_uname}" rc2=$OCF_SUCCESS fi ocf_log info "${LH} post-start end." - if [[ $rc2 == $OCF_ERR_GENERIC ]] ; then + if [ $rc2 -eq $OCF_ERR_GENERIC ] ; then ocf_log warn "${LH} Failed to join the cluster on post-start. The resource will be restarted." ocf_log info "${LH} post-start end." return $OCF_NOT_RUNNING @@ -1345,14 +1369,14 @@ action_notify() { # if rabbitmq-server stops on any another node, we should remove it from cluster (as ordinary operation) ocf_log info "${LH} post-stop begin." # Report not running, if there are no nodes being stopped reported - if [ -z ${OCF_RESKEY_CRM_meta_notify_stop_uname} ] ; then + if [ -z "${OCF_RESKEY_CRM_meta_notify_stop_uname}" ] ; then ocf_log warn "${LH} there are no nodes being stopped reported on post-stop. The resource will be restarted." ocf_log info "${LH} post-stop end." return $OCF_NOT_RUNNING fi my_host "${OCF_RESKEY_CRM_meta_notify_stop_uname}" rc=$? - if [[ $rc != $OCF_SUCCESS ]] ; then + if [ $rc -ne $OCF_SUCCESS ] ; then # On ohter nodes processing the post-stop, make sure the stopped node will be forgotten unjoin_nodes_from_cluster "${OCF_RESKEY_CRM_meta_notify_stop_uname}" else @@ -1367,14 +1391,14 @@ action_notify() { # if rabbitmq-server stops on any another node, we should remove it from cluster (as ordinary operation) ocf_log info "${LH} post-demote begin." # Report not running, if the list of nodes being demoted reported empty - if [ -z ${OCF_RESKEY_CRM_meta_notify_demote_uname} ] ; then + if [ -z "${OCF_RESKEY_CRM_meta_notify_demote_uname}" ] ; then ocf_log warn "${LH} there are no nodes being demoted reported on post-demote. The resource will be restarted." ocf_log info "${LH} post-demote end." return $OCF_NOT_RUNNING fi my_host "${OCF_RESKEY_CRM_meta_notify_demote_uname}" rc=$? - if [[ $rc != $OCF_SUCCESS ]] ; then + if [ $rc -ne $OCF_SUCCESS ] ; then # On ohter nodes processing the post-demote, make sure the demoted node will be forgotten unjoin_nodes_from_cluster "${OCF_RESKEY_CRM_meta_notify_demote_uname}" else @@ -1385,7 +1409,7 @@ action_notify() { stop_rmq_server_app rc2=$? crm_attribute -N $THIS_PCMK_NODE -l reboot --name 'rabbit-start-time' --delete - if [[ $rc2 != $OCF_SUCCESS ]] ; then + if [ $rc2 -ne $OCF_SUCCESS ] ; then ocf_log err "${LH} RMQ-server app can't be stopped on post-demote. Master resource is failed" ocf_log info "${LH} post-demote end." exit $OCF_FAILED_MASTER @@ -1405,7 +1429,7 @@ action_promote() { local rc=$OCF_ERR_GENERIC local LH="${LL} promote:" - if [[ ${OCF_RESKEY_debug} == "true" ]] ; then + if [ "${OCF_RESKEY_debug}" = 'true' ] ; then d=$(date '+%Y%m%d %H:%M:%S') echo $d >> /tmp/rmq-promote.log env >> /tmp/rmq-promote.log @@ -1426,14 +1450,14 @@ action_promote() { rc=$? ocf_log info "${LH} Updating cluster master attribute" ocf_run crm_attribute -N $THIS_PCMK_NODE -l reboot --name 'rabbit-master' --update 'true' - if [[ $rc != $OCF_SUCCESS ]] ; then + if [ $rc -ne $OCF_SUCCESS ] ; then ocf_log info "${LH} RMQ app is not started. Starting..." start_rmq_server_app rc=$? - if [[ $rc == 0 ]] ; then + if [ $rc -eq 0 ] ; then try_to_start_rmq_app rc=$? - if [[ $rc != 0 ]] ; then + if [ $rc -ne 0 ] ; then ocf_log err "${LH} Can't start RMQ app. Master resource is failed." ocf_log info "${LH} action end." exit $OCF_FAILED_MASTER @@ -1450,7 +1474,7 @@ action_promote() { get_monitor rc=$? ocf_log info "${LH} Master status is $rc" - if [ $rc == $OCF_RUNNING_MASTER ] + if [ $rc = $OCF_RUNNING_MASTER ] then rc=$OCF_SUCCESS else @@ -1502,7 +1526,7 @@ action_demote() { local rc=$OCF_ERR_GENERIC local LH="${LL} demote:" - if [[ ${OCF_RESKEY_debug} == "true" ]] ; then + if [ "${OCF_RESKEY_debug}" = 'true' ] ; then d=`date '+%Y%m%d %H:%M:%S'` echo $d >> /tmp/rmq-demote.log env >> /tmp/rmq-demote.log |