diff --git a/heartbeat/exportfs b/heartbeat/exportfs index c10777fa9..2307a9e67 100755 --- a/heartbeat/exportfs +++ b/heartbeat/exportfs @@ -1,487 +1,490 @@ #!/bin/sh # exportfs # # Description: Manages nfs exported file system. # # (c) 2010 Ben Timby, Florian Haas, Dejan Muhamedagic, # and Linux-HA contributors # # License: GNU General Public License v2 (GPLv2) and later ####################################################################### # Initialization: : ${OCF_FUNCTIONS_DIR=${OCF_ROOT}/lib/heartbeat} . ${OCF_FUNCTIONS_DIR}/ocf-shellfuncs # Defaults OCF_RESKEY_unlock_on_stop_default=1 OCF_RESKEY_wait_for_leasetime_on_stop_default=0 OCF_RESKEY_rmtab_backup_default=".rmtab" : ${OCF_RESKEY_unlock_on_stop=${OCF_RESKEY_unlock_on_stop_default}} : ${OCF_RESKEY_wait_for_leasetime_on_stop=${OCF_RESKEY_wait_for_leasetime_on_stop_default}} : ${OCF_RESKEY_rmtab_backup=${OCF_RESKEY_rmtab_backup_default}} ####################################################################### exportfs_meta_data() { cat < 1.0 Exportfs uses the exportfs command to add/remove nfs exports. It does NOT manage the nfs server daemon. It depends on Linux specific NFS implementation details, so is considered not portable to other platforms yet. Manages NFS exports The client specification allowing remote machines to mount the directory (or directories) over NFS. Note: it follows the format defined in "man exportfs". For example, in the use case to export the directory(-ies) for multiple subnets, please do config a dedicated primitive for each subnet CIDR ip address, and do not attempt to use multiple CIDR ip addresses in a space separated list, like in /etc/exports. Client ACL. The options to pass to exportfs for the exported directory or directories. Export options. The directory or directories to be exported using NFS. Multiple directories are separated by white space. The directory or directories to export. The fsid option to pass to exportfs. This can be a unique positive integer, a UUID (assuredly sans comma characters), or the special string "root" which is functionally identical to numeric fsid of 0. If multiple directories are being exported, then they are assigned ids sequentially starting with this fsid (fsid, fsid+1, fsid+2, ...). Obviously, in that case the fsid must be an integer. 0 (root) identifies the export as the root of an NFSv4 pseudofilesystem -- avoid this setting unless you understand its special status. This value will override any fsid provided via the options parameter. Unique fsid within cluster or starting fsid for multiple exports. Relinquish NFS locks associated with this filesystem when the resource stops. Enabling this parameter is highly recommended unless the path exported by this ${__SCRIPT_NAME} resource is also exported by a different resource. Note: Unlocking is only possible on Linux systems where /proc/fs/nfsd/unlock_filesystem exists and is writable. If your system does not fulfill this requirement (on account of having an nonrecent kernel, for example), you may set this parameter to 0 to silence the associated warning. Unlock filesystem on stop? When stopping (unexporting), wait out the NFSv4 lease time. Only after all leases have expired does the NFS kernel server relinquish all server-side handles on the exported filesystem. If this ${__SCRIPT_NAME} resource manages an export that resides on a mount point designed to fail over along with the NFS export itself, then enabling this parameter will ensure such failover is working properly. Note that when this parameter is set, your stop timeout MUST accommodate for the wait period. This parameter is safe to disable if none of your NFS clients are using NFS version 4 or later. Ride out the NFSv4 lease time on resource stop? Back up those entries from the NFS rmtab that apply to the exported directory, to the specified backup file. The filename is interpreted as relative to the exported directory. This backup is required if clients are connecting to the export via NFSv3 over TCP. Note that a configured monitor operation is required for this functionality. To disable rmtab backups, set this parameter to the special string "none". Location of the rmtab backup, relative to directory. END return $OCF_SUCCESS } exportfs_methods() { cat <<-EOF start stop status monitor validate-all methods meta-data usage EOF } reset_fsid() { CURRENT_FSID=$OCF_RESKEY_fsid } bump_fsid() { CURRENT_FSID=$((CURRENT_FSID+1)) } get_fsid() { echo $CURRENT_FSID } # run a function on all directories forall() { local func=$1 shift 1 local fast_exit="" local dir rc=0 if [ "$2" = fast_exit ]; then fast_exit=1 shift 1 fi reset_fsid for dir in $OCF_RESKEY_directory; do $func $dir "$@" rc=$(($rc | $?)) [ $NUMDIRS -gt 1 ] && bump_fsid [ "$fast_exit" ] && continue [ $rc -ne 0 ] && return $rc done return $rc } backup_rmtab() { local dir=$1 local rmtab_backup rmtab_backup="$dir/${OCF_RESKEY_rmtab_backup}" if [ -r /var/lib/nfs/rmtab ]; then grep ":$dir:" /var/lib/nfs/rmtab > ${rmtab_backup} fi } restore_rmtab() { local dir=$1 local rmtab_backup rmtab_backup="$dir/${OCF_RESKEY_rmtab_backup}" if [ -r ${rmtab_backup} ]; then local tmpf=`mktemp` sort -u ${rmtab_backup} /var/lib/nfs/rmtab > $tmpf && install -o root -m 644 $tmpf /var/lib/nfs/rmtab rm -f $tmpf ocf_log debug "Restored `wc -l ${rmtab_backup}` rmtab entries from ${rmtab_backup}." else ocf_log warn "rmtab backup ${rmtab_backup} not found or not readable." fi } exportfs_usage() { cat <" instead of "*" format_exports | grep -q -x -F "$dir " rc=$? fi # log something only for monitors if [ $rc -ne 0 -a "$__OCF_ACTION" = "monitor" ]; then local sev="info" ocf_is_probe || sev="err" ocf_log $sev "$dir not exported to $spec (stopped)." fi return $rc } exportfs_monitor () { local spec if ! ha_pseudo_resource "${OCF_RESOURCE_INSTANCE}" monitor; then return $OCF_NOT_RUNNING fi # IPv6 addresses and networks are encased in brackets that need # to be removed case "$OCF_RESKEY_clientspec" in *:*:*) spec="$(echo "$OCF_RESKEY_clientspec" | tr -d '[]')" ;; *) spec="$OCF_RESKEY_clientspec" ;; esac if forall is_exported "$spec"; then if [ ${OCF_RESKEY_rmtab_backup} != "none" ]; then forall backup_rmtab fi return $OCF_SUCCESS else return $OCF_NOT_RUNNING fi } +testdir() { + if [ ! -d $1 ]; then + mkdir -p "$1" + if [ $? -ne 0 ]; then + ocf_exit_reason "Unable to create directory $1" + return 1 + fi + fi + return 0 +} export_one() { local dir=$1 local opts sep sep="" if [ -n "$OCF_RESKEY_options" ]; then opts="$OCF_RESKEY_options" sep="," fi if echo "$opts" | grep fsid >/dev/null; then #replace fsid in options list opts=`echo "$opts" | sed "s,fsid=[^,]*,fsid=$(get_fsid),g"` else #tack the fsid option onto our options list. opts="${opts}${sep}fsid=$(get_fsid)" fi opts="-o $opts" # if any of directories fails to export we can exit # immediately ocf_run exportfs -v $opts "${OCF_RESKEY_clientspec}:$dir" if [ $? -ne 0 ]; then ocf_exit_reason "exportfs failed - exportfs -v $opts ${OCF_RESKEY_clientspec}:$dir" exit $OCF_ERR_GENERIC fi ocf_log info "directory $dir exported" return $OCF_SUCCESS } exportfs_start () { + if ! forall testdir; then + return $OCF_ERR_INSTALLED + fi + if exportfs_monitor; then ocf_log debug "already exported" return $OCF_SUCCESS fi ocf_log info "Exporting file system(s) ..." ha_pseudo_resource "${OCF_RESOURCE_INSTANCE}" start forall export_one # Restore the rmtab to ensure smooth NFS-over-TCP failover if [ ${OCF_RESKEY_rmtab_backup} != "none" ]; then forall restore_rmtab fi } unlock_fs() { local dir=$1 local unlockfile unlockfile=/proc/fs/nfsd/unlock_filesystem if [ -w ${unlockfile} ]; then echo "$dir" > ${unlockfile} ocf_log info "Unlocked NFS export $dir" else ocf_log warn "Unable to unlock NFS export $dir, ${unlockfile} not found or not writable" fi } wait_for_leasetime() { local leasetimefile local sleeptime leasetimefile=/proc/fs/nfsd/nfsv4leasetime if [ -r ${leasetimefile} ]; then sleeptime=$((`cat ${leasetimefile}`+2)) ocf_log info "Sleeping ${sleeptime} seconds to accommodate for NFSv4 lease expiry" sleep ${sleeptime}s else ocf_log warn "Unable to read NFSv4 lease time from ${leasetimefile}, file not found or not readable" fi } cleanup_export_cache() { # see if the cache is blocking unexport local contentfile=/proc/net/rpc/nfsd.export/content local fsid_re local i=1 fsid_re="fsid=(echo `forall get_fsid`|sed 's/ /|/g')," while :; do grep -E -q "$fsid_re" $contentfile || break ocf_log info "Cleanup export cache ... (try $i)" ocf_run exportfs -f sleep 0.5 i=$((i + 1)) done } unexport_one() { local dir=$1 ocf_run exportfs -v -u ${OCF_RESKEY_clientspec}:$dir } exportfs_stop () { local rc exportfs_monitor if [ $? -eq $OCF_NOT_RUNNING ]; then ocf_log debug "not exported" return $OCF_SUCCESS fi ocf_log info "Un-exporting file system ..." # Backup the rmtab to ensure smooth NFS-over-TCP failover if [ ${OCF_RESKEY_rmtab_backup} != "none" ]; then forall backup_rmtab fi forall unexport_one rc=$? if ocf_is_true ${OCF_RESKEY_unlock_on_stop}; then forall unlock_fs fi if ocf_is_true ${OCF_RESKEY_wait_for_leasetime_on_stop}; then wait_for_leasetime fi if [ $rc -eq 0 ]; then cleanup_export_cache ha_pseudo_resource "${OCF_RESOURCE_INSTANCE}" stop ocf_log info "Un-exported file system(s)" return $OCF_SUCCESS else ocf_exit_reason "Failed to un-export file system(s)" return $OCF_ERR_GENERIC fi } -testdir() { - if [ ! -d $1 ]; then - ocf_is_probe || - ocf_log err "$1 does not exist or is not a directory" - return 1 - fi - return 0 -} exportfs_validate_all () { if echo "$OCF_RESKEY_fsid" | grep -q -F ','; then ocf_exit_reason "$OCF_RESKEY_fsid cannot contain a comma" return $OCF_ERR_CONFIGURED fi if [ $NUMDIRS -gt 1 ] && ! ocf_is_decimal "$OCF_RESKEY_fsid"; then ocf_exit_reason "use integer fsid when exporting multiple directories" return $OCF_ERR_CONFIGURED fi - if ! forall testdir; then - return $OCF_ERR_INSTALLED - fi } for dir in $OCF_RESKEY_directory; do # strip off trailing '/' from directory dir=$(echo $dir | sed 's/\/*$//') : ${dir:=/} if [ -e "$dir" ] ; then canonicalized_dir=$(readlink -f "$dir") if [ $? -ne 0 ]; then if [ "$__OCF_ACTION" != "stop" ]; then ocf_exit_reason "Could not canonicalize $dir because readlink failed" exit $OCF_ERR_GENERIC fi fi else case "$__OCF_ACTION" in - stop|monitor) + stop|monitor|validate-all) canonicalized_dir="$dir" ocf_log debug "$dir does not exist" ;; *) ocf_exit_reason "$dir does not exist" exit $OCF_ERR_CONFIGURED ;; esac fi directories="$directories$canonicalized_dir " done OCF_RESKEY_directory="${directories%% }" NUMDIRS=`echo "$OCF_RESKEY_directory" | wc -w` OCF_REQUIRED_PARAMS="directory fsid clientspec" OCF_REQUIRED_BINARIES="exportfs" ocf_rarun $* diff --git a/heartbeat/pgsql b/heartbeat/pgsql index aa8a13a84..532063ac5 100755 --- a/heartbeat/pgsql +++ b/heartbeat/pgsql @@ -1,2252 +1,2263 @@ #!/bin/sh # # Description: Manages a PostgreSQL Server as an OCF High-Availability # resource # # Authors: Serge Dubrouski (sergeyfd@gmail.com) -- original RA # Florian Haas (florian@linbit.com) -- makeover # Takatoshi MATSUO (matsuo.tak@gmail.com) -- support replication # David Corlette (dcorlette@netiq.com) -- add support for non-standard library locations and non-standard port # # Copyright: 2006-2012 Serge Dubrouski # and other Linux-HA contributors # License: GNU General Public License (GPL) # ############################################################################### # Initialization: : ${OCF_FUNCTIONS_DIR=${OCF_ROOT}/lib/heartbeat} . ${OCF_FUNCTIONS_DIR}/ocf-shellfuncs # Use runuser if available for SELinux. if [ -x /sbin/runuser ]; then SU=runuser else SU=su fi # # Get PostgreSQL Configuration parameter # get_pgsql_param() { local param_name param_name=$1 perl_code="if (/^\s*$param_name[\s=]+\s*(.*)$/) { \$dir=\$1; \$dir =~ s/\s*\#.*//; \$dir =~ s/^'(\S*)'/\$1/; print \$dir;}" perl -ne "$perl_code" < $OCF_RESKEY_config } # Defaults OCF_RESKEY_pgctl_default=/usr/bin/pg_ctl OCF_RESKEY_psql_default=/usr/bin/psql OCF_RESKEY_pgdata_default=/var/lib/pgsql/data OCF_RESKEY_pgdba_default=postgres OCF_RESKEY_pghost_default="" OCF_RESKEY_pgport_default=5432 OCF_RESKEY_pglibs_default=/usr/lib OCF_RESKEY_start_opt_default="" OCF_RESKEY_ctl_opt_default="" OCF_RESKEY_pgdb_default=template1 OCF_RESKEY_logfile_default=/dev/null OCF_RESKEY_socketdir_default="" OCF_RESKEY_stop_escalate_default=90 OCF_RESKEY_monitor_user_default="" OCF_RESKEY_monitor_password_default="" OCF_RESKEY_monitor_sql_default="select now();" OCF_RESKEY_check_wal_receiver_default="false" # Defaults for replication OCF_RESKEY_rep_mode_default=none OCF_RESKEY_node_list_default="" OCF_RESKEY_restore_command_default="" OCF_RESKEY_archive_cleanup_command_default="" OCF_RESKEY_recovery_end_command_default="" OCF_RESKEY_master_ip_default="" OCF_RESKEY_repuser_default="postgres" OCF_RESKEY_primary_conninfo_opt_default="" OCF_RESKEY_restart_on_promote_default="false" OCF_RESKEY_tmpdir_default="/var/lib/pgsql/tmp" OCF_RESKEY_xlog_check_count_default="3" OCF_RESKEY_crm_attr_timeout_default="5" OCF_RESKEY_stop_escalate_in_slave_default=90 OCF_RESKEY_replication_slot_name_default="" : ${OCF_RESKEY_pgctl=${OCF_RESKEY_pgctl_default}} : ${OCF_RESKEY_psql=${OCF_RESKEY_psql_default}} : ${OCF_RESKEY_pgdata=${OCF_RESKEY_pgdata_default}} : ${OCF_RESKEY_pgdba=${OCF_RESKEY_pgdba_default}} : ${OCF_RESKEY_pghost=${OCF_RESKEY_pghost_default}} : ${OCF_RESKEY_pgport=${OCF_RESKEY_pgport_default}} : ${OCF_RESKEY_pglibs=${OCF_RESKEY_pglibs_default}} : ${OCF_RESKEY_config=${OCF_RESKEY_pgdata}/postgresql.conf} : ${OCF_RESKEY_start_opt=${OCF_RESKEY_start_opt_default}} : ${OCF_RESKEY_ctl_opt=${OCF_RESKEY_ctl_opt_default}} : ${OCF_RESKEY_pgdb=${OCF_RESKEY_pgdb_default}} : ${OCF_RESKEY_logfile=${OCF_RESKEY_logfile_default}} : ${OCF_RESKEY_socketdir=${OCF_RESKEY_socketdir_default}} : ${OCF_RESKEY_stop_escalate=${OCF_RESKEY_stop_escalate_default}} : ${OCF_RESKEY_monitor_user=${OCF_RESKEY_monitor_user_default}} : ${OCF_RESKEY_monitor_password=${OCF_RESKEY_monitor_password_default}} : ${OCF_RESKEY_monitor_sql=${OCF_RESKEY_monitor_sql_default}} : ${OCF_RESKEY_check_wal_receiver=${OCF_RESKEY_check_wal_receiver_default}} # for replication : ${OCF_RESKEY_rep_mode=${OCF_RESKEY_rep_mode_default}} : ${OCF_RESKEY_node_list=${OCF_RESKEY_node_list_default}} : ${OCF_RESKEY_restore_command=${OCF_RESKEY_restore_command_default}} : ${OCF_RESKEY_archive_cleanup_command=${OCF_RESKEY_archive_cleanup_command_default}} : ${OCF_RESKEY_recovery_end_command=${OCF_RESKEY_recovery_end_command_default}} : ${OCF_RESKEY_master_ip=${OCF_RESKEY_master_ip_default}} : ${OCF_RESKEY_repuser=${OCF_RESKEY_repuser_default}} : ${OCF_RESKEY_primary_conninfo_opt=${OCF_RESKEY_primary_conninfo_opt_default}} : ${OCF_RESKEY_restart_on_promote=${OCF_RESKEY_restart_on_promote_default}} : ${OCF_RESKEY_tmpdir=${OCF_RESKEY_tmpdir_default}} : ${OCF_RESKEY_xlog_check_count=${OCF_RESKEY_xlog_check_count_default}} : ${OCF_RESKEY_crm_attr_timeout=${OCF_RESKEY_crm_attr_timeout_default}} : ${OCF_RESKEY_stop_escalate_in_slave=${OCF_RESKEY_stop_escalate_in_slave_default}} : ${OCF_RESKEY_replication_slot_name=${OCF_RESKEY_replication_slot_name_default}} usage() { cat < 1.0 Resource script for PostgreSQL. It manages a PostgreSQL as an HA resource. Manages a PostgreSQL database instance Path to pg_ctl command. pgctl Start options (-o start_opt in pg_ctl). "-i -p 5432" for example. start_opt Additional pg_ctl options (-w, -W etc..). ctl_opt Path to psql command. psql Path to PostgreSQL data directory. pgdata User that owns PostgreSQL. pgdba Hostname/IP address where PostgreSQL is listening pghost Port where PostgreSQL is listening pgport Custom location of the Postgres libraries. If not set, the standard location will be used. pglibs PostgreSQL user that pgsql RA will user for monitor operations. If it's not set pgdba user will be used. monitor_user Password for monitor user. monitor_password SQL script that will be used for monitor operations. monitor_sql Path to the PostgreSQL configuration file for the instance. Configuration file Database that will be used for monitoring. pgdb Path to PostgreSQL server log output file. logfile Unix socket directory for PostgreSQL. If you use PostgreSQL 9.3 or higher and define unix_socket_directories in the postgresql.conf, then you must set socketdir to determine which directory is used for psql command. socketdir Number of seconds to wait for stop (using -m fast) before resorting to -m immediate stop escalation Replication mode may be set to "async" or "sync" or "slave". They require PostgreSQL 9.1 or later. Once set, "async" and "sync" require node_list, master_ip, and restore_command parameters,as well as configuring PostgreSQL for replication (in postgresql.conf and pg_hba.conf). "slave" means that RA only makes recovery.conf before starting to connect to primary which is running somewhere. It doesn't need master/slave setting. It requires master_ip restore_command parameters. rep_mode All node names. Please separate each node name with a space. This is optional for replication. Defaults to all nodes in the cluster node list restore_command for recovery.conf. This is required for replication. restore_command archive_cleanup_command for recovery.conf. This is used for replication and is optional. archive_cleanup_command recovery_end_command for recovery.conf. This is used for replication and is optional. recovery_end_command Master's floating IP address to be connected from hot standby. This parameter is used for "primary_conninfo" in recovery.conf. This is required for replication. master ip User used to connect to the master server. This parameter is used for "primary_conninfo" in recovery.conf. This is required for replication. repuser primary_conninfo options of recovery.conf except host, port, user and application_name. This is optional for replication. primary_conninfo_opt If this is true, RA deletes recovery.conf and restarts PostgreSQL on promote to keep Timeline ID. It probably makes fail-over slower. It's recommended to set on-fail of promote up as fence. This is optional for replication. restart_on_promote Set this option when using replication slots. Can only use lower case letters, numbers and underscore for replication_slot_name. The replication slots would be created for each node, with the name adding the node name as postfix. For example, replication_slot_name is "sample" and 2 slaves which are "node1" and "node2" connect to their slots, the slots names are "sample_node1" and "sample_node2". If the node name contains a upper case letter, hyphen and dot, those characters will be converted to a lower case letter or an underscore. For example, Node-1.example.com to node_1_example_com. pgsql RA doesn't monitor and delete the replication slot. When the slave node has been disconnected in failure or the like, execute one of the following manually. Otherwise it may eventually cause a disk full because the master node will continue to accumulate the unsent WAL. 1. recover and reconnect the slave node to the master node as soon as possible. 2. delete the slot on the master node by following psql command. $ select pg_drop_replication_slot('replication_slot_name'); replication_slot_name Path to temporary directory. This is optional for replication. tmpdir Number of checks of xlog on monitor before promote. This is optional for replication. Note: For backward compatibility, the terms are unified with PostgreSQL 9. If you are using PostgreSQL 10 or later, replace "xlog" with "wal". Likewise, replacing "location" with "lsn". xlog check count The timeout of crm_attribute forever update command. Default value is 5 seconds. This is optional for replication. The timeout of crm_attribute forever update command. Number of seconds to wait for stop (using -m fast) before resorting to -m immediate in slave state. This is optional for replication. stop escalation_in_slave If this is true, RA checks wal_receiver process on monitor and notifies its status using "(resource name)-receiver-status" attribute. It's useful for checking whether PostgreSQL (hot standby) connects to primary. The attribute shows status as "normal" or "normal (master)" or "ERROR". Note that if you configure PostgreSQL as master/slave resource, then wal receiver is not running in the master and the attribute shows status as "normal (master)" consistently because it is normal status. check_wal_receiver EOF } # # Run the given command in the Resource owner environment... # runasowner() { local quietrun="" local loglevel="-err" local var for var in 1 2 do case "$1" in "-q") quietrun="-q" shift 1;; "info"|"warn"|"err") loglevel="-$1" shift 1;; *) ;; esac done ocf_run $quietrun $loglevel $SU $OCF_RESKEY_pgdba -c "cd $OCF_RESKEY_pgdata; $*" } # # Shell escape # escape_string() { echo "$*" | sed -e "s/'/'\\\\''/g" } # # methods: What methods/operations do we support? # pgsql_methods() { cat </dev/null 2>&1" return $? fi # No PID file false } pgsql_wal_receiver_status() { local PID local receiver_parent_pids local pgsql_real_monitor_status=$1 PID=`head -n 1 $PIDFILE` receiver_parent_pids=`ps -ef | tr -s " " | grep "[w]al\s*receiver" | cut -d " " -f 3` if echo "$receiver_parent_pids" | grep -q -w "$PID" ; then attrd_updater -n "$PGSQL_WAL_RECEIVER_STATUS_ATTR" -v "normal" return 0 fi if [ $pgsql_real_monitor_status -eq "$OCF_RUNNING_MASTER" ]; then attrd_updater -n "$PGSQL_WAL_RECEIVER_STATUS_ATTR" -v "normal (master)" return 0 fi attrd_updater -n "$PGSQL_WAL_RECEIVER_STATUS_ATTR" -v "ERROR" ocf_log warn "wal receiver process is not running" return 1 } # # pgsql_real_monitor # pgsql_real_monitor() { local loglevel local rc local output # Set the log level of the error message loglevel=${1:-err} if ! pgsql_status then ocf_log info "PostgreSQL is down" return $OCF_NOT_RUNNING fi if is_replication; then #Check replication state output=`exec_sql "${CHECK_MS_SQL}"` rc=$? if [ $rc -ne 0 ]; then report_psql_error $rc $loglevel "Can't get PostgreSQL recovery status." return $OCF_ERR_GENERIC fi case "$output" in f) ocf_log debug "PostgreSQL is running as a primary." if [ "$OCF_RESKEY_monitor_sql" = "$OCF_RESKEY_monitor_sql_default" ]; then if ocf_is_probe; then # Set initial score for primary. exec_with_retry 0 ocf_promotion_score -v $PROMOTE_ME fi return $OCF_RUNNING_MASTER fi ;; t) ocf_log debug "PostgreSQL is running as a hot standby." if ocf_is_probe; then # Set initial score for hot standby. exec_with_retry 0 ocf_promotion_score -v $CAN_NOT_PROMOTE fi return $OCF_SUCCESS;; *) ocf_exit_reason "$CHECK_MS_SQL output is $output" return $OCF_ERR_GENERIC;; esac fi OCF_RESKEY_monitor_sql=`escape_string "$OCF_RESKEY_monitor_sql"` runasowner -q $loglevel "$OCF_RESKEY_psql $psql_options \ -c '$OCF_RESKEY_monitor_sql'" rc=$? if [ $rc -ne 0 ]; then report_psql_error $rc $loglevel "PostgreSQL $OCF_RESKEY_pgdb isn't running." return $OCF_ERR_GENERIC fi if is_replication; then return $OCF_RUNNING_MASTER fi return $OCF_SUCCESS } pgsql_replication_monitor() { local rc rc=$1 if [ $rc -ne $OCF_SUCCESS -a $rc -ne "$OCF_RUNNING_MASTER" ]; then return $rc fi # If I am Master if [ $rc -eq $OCF_RUNNING_MASTER ]; then change_data_status "$NODENAME" "LATEST" change_pgsql_status "$NODENAME" "PRI" control_slave_status || return $OCF_ERR_GENERIC if [ "$RE_CONTROL_SLAVE" = "true" ]; then sleep 2 ocf_log info "re-controlling slave status." RE_CONTROL_SLAVE="none" control_slave_status || return $OCF_ERR_GENERIC fi return $rc fi # I can't get master node name from $OCF_RESKEY_CRM_meta_notify_master_uname on monitor, # so I will get master node name using crm_mon -n print_crm_mon | grep -q -i -E "/dev/null` if [ $? -eq 0 ]; then my_master_baseline=`$CRM_ATTR_REBOOT -N "$NODENAME" -n \ "$PGSQL_MASTER_BASELINE" -G -q 2>/dev/null` # get older location cmp_location=`printf "$master_baseline\n$my_master_baseline\n" |\ sort | head -1` if [ "$cmp_location" != "$my_master_baseline" ]; then # We used to set the failcount to INF for the resource here in # order to move the master to the other node. However, setting # the failcount should be done only by the CRM and so this use # got deprecated in pacemaker version 1.1.17. Now we do the # "ban resource from the node". ocf_exit_reason "My data is newer than new master's one. New master's location : $master_baseline" exec_with_retry 0 $CRM_RESOURCE -B -r $OCF_RESOURCE_INSTANCE -N $NODENAME -Q return $OCF_ERR_GENERIC fi fi fi return $OCF_SUCCESS } pgsql_notify() { local type="${OCF_RESKEY_CRM_meta_notify_type}" local op="${OCF_RESKEY_CRM_meta_notify_operation}" local rc if ! is_replication; then return $OCF_SUCCESS fi ocf_log debug "notify: ${type} for ${op}" case $type in pre) case $op in promote) pgsql_pre_promote return $? ;; esac ;; post) case $op in promote) delete_xlog_location PROMOTE_NODE=`echo $OCF_RESKEY_CRM_meta_notify_promote_uname | \ sed "s/ /\n/g" | head -1 | tr '[A-Z]' '[a-z]'` if [ "$PROMOTE_NODE" != "$NODENAME" ]; then delete_master_baseline fi return $OCF_SUCCESS ;; demote) pgsql_post_demote return $? ;; start|stop) MASTER_NODE=`echo $OCF_RESKEY_CRM_meta_notify_master_uname | \ sed "s/ /\n/g" | head -1 | tr '[A-Z]' '[a-z]'` if [ "$NODENAME" = "$MASTER_NODE" ]; then control_slave_status fi return $OCF_SUCCESS ;; esac ;; esac return $OCF_SUCCESS } control_slave_status() { local rc local data_status local target local all_data_status local tmp_data_status local number_of_nodes all_data_status=`exec_sql "${CHECK_REPLICATION_STATE_SQL}"` rc=$? if [ $rc -eq 0 ]; then if [ -n "$all_data_status" ]; then all_data_status=`echo $all_data_status | sed "s/\n/ /g"` fi else report_psql_error $rc err "Can't get PostgreSQL replication status." return 1 fi number_of_nodes=`echo $NODE_LIST | wc -w` for target in $NODE_LIST; do if [ "$target" = "$NODENAME" ]; then continue fi data_status="DISCONNECT" if [ -n "$all_data_status" ]; then for tmp_data_status in $all_data_status; do if ! echo $tmp_data_status | grep -q "^${target}|"; then continue fi data_status=`echo $tmp_data_status | cut -d "|" -f 2,3` ocf_log debug "node_name and data_status is $tmp_data_status" break done fi case "$data_status" in "STREAMING|SYNC") change_data_status "$target" "$data_status" change_master_score "$target" "$CAN_PROMOTE" change_pgsql_status "$target" "HS:sync" ;; "STREAMING|ASYNC") change_data_status "$target" "$data_status" if [ "$OCF_RESKEY_rep_mode" = "sync" ]; then change_master_score "$target" "$CAN_NOT_PROMOTE" set_sync_mode "$target" else if [ $number_of_nodes -le 2 ]; then change_master_score "$target" "$CAN_PROMOTE" else # I can't determine which slave's data is newest in async mode. change_master_score "$target" "$CAN_NOT_PROMOTE" fi fi change_pgsql_status "$target" "HS:async" ;; "STREAMING|POTENTIAL") change_data_status "$target" "$data_status" change_master_score "$target" "$CAN_NOT_PROMOTE" change_pgsql_status "$target" "HS:potential" ;; "DISCONNECT") change_data_status "$target" "$data_status" change_master_score "$target" "$CAN_NOT_PROMOTE" if [ "$OCF_RESKEY_rep_mode" = "sync" ]; then set_async_mode "$target" fi ;; *) change_data_status "$target" "$data_status" change_master_score "$target" "$CAN_NOT_PROMOTE" if [ "$OCF_RESKEY_rep_mode" = "sync" ]; then set_async_mode "$target" fi change_pgsql_status "$target" "HS:connected" ;; esac done return 0 } have_master_right() { local old local new local output local data_status local node local mylocation local count local newestXlog local oldfile local newfile ocf_log debug "Checking if I have a master right." data_status=`$CRM_ATTR_FOREVER -N "$NODENAME" -n \ "$PGSQL_DATA_STATUS_ATTR" -G -q 2>/dev/null` if [ "$OCF_RESKEY_rep_mode" = "sync" ]; then if [ -n "$data_status" -a "$data_status" != "STREAMING|SYNC" -a \ "$data_status" != "LATEST" ]; then ocf_log warn "My data is out-of-date. status=$data_status" return 1 fi else if [ -n "$data_status" -a "$data_status" != "STREAMING|SYNC" -a \ "$data_status" != "STREAMING|ASYNC" -a \ "$data_status" != "LATEST" ]; then ocf_log warn "My data is out-of-date. status=$data_status" return 1 fi fi ocf_log info "My data status=$data_status." show_xlog_location if [ $? -ne 0 ]; then ocf_exit_reason "Failed to show my xlog location." exit $OCF_ERR_GENERIC fi old=0 for count in `seq $OCF_RESKEY_xlog_check_count`; do if [ -f ${XLOG_NOTE_FILE}.$count ]; then old=$count continue fi break done new=`expr $old + 1` # get xlog locations of all nodes for node in ${NODE_LIST}; do output=`$CRM_ATTR_REBOOT -N "$node" -n \ "$PGSQL_XLOG_LOC_NAME" -G -q 2>/dev/null` if [ $? -ne 0 ]; then ocf_log warn "Can't get $node xlog location." continue else ocf_log info "$node xlog location : $output" echo "$node $output" >> ${XLOG_NOTE_FILE}.${new} if [ "$node" = "$NODENAME" ]; then mylocation=$output fi fi done oldfile=`cat ${XLOG_NOTE_FILE}.${old} 2>/dev/null` newfile=`cat ${XLOG_NOTE_FILE}.${new} 2>/dev/null` if [ "$oldfile" != "$newfile" ]; then # reset counter rm -f ${XLOG_NOTE_FILE}.* printf "$newfile\n" > ${XLOG_NOTE_FILE}.0 return 1 fi if [ "$new" -ge "$OCF_RESKEY_xlog_check_count" ]; then newestXlog=`printf "$newfile\n" | sort -t " " -k 2,3 -r | \ head -1 | cut -d " " -f 2` if [ "$newestXlog" = "$mylocation" ]; then ocf_log info "I have a master right." exec_with_retry 5 ocf_promotion_score -v $PROMOTE_ME return 0 fi change_data_status "$NODENAME" "DISCONNECT" ocf_log info "I don't have correct master data." # reset counter rm -f ${XLOG_NOTE_FILE}.* printf "$newfile\n" > ${XLOG_NOTE_FILE}.0 fi return 1 } is_replication() { if [ "$OCF_RESKEY_rep_mode" != "none" -a "$OCF_RESKEY_rep_mode" != "slave" ]; then return 0 fi return 1 } use_replication_slot() { if [ -n "$OCF_RESKEY_replication_slot_name" ]; then return 0 fi return 1 } create_replication_slot_name() { local number_of_nodes=0 local target local replication_slot_name local replication_slot_name_list_tmp local replication_slot_name_list if [ -n "$NODE_LIST" ]; then number_of_nodes=`echo $NODE_LIST | wc -w` fi if [ $number_of_nodes -le 0 ]; then replication_slot_name_list="" # The Master node should have some slots equal to the number of Slaves, and # the Slave nodes connect to their dedicated slot on the Master. # To ensuring that the slots name are each unique, add postfix to $OCF_RESKEY_replication_slot. # The postfix is "_$target". else for target in $NODE_LIST do if [ "$target" != "$NODENAME" ]; then # The Uppercase, "-" and "." don't allow to use in slot_name. # If the NODENAME contains them, convert upper case to lower case and "_" and "." to "_". target=`echo "$target" | tr 'A-Z.-' 'a-z__'` replication_slot_name="$OCF_RESKEY_replication_slot_name"_"$target" replication_slot_name_list_tmp="$replication_slot_name_list" replication_slot_name_list="$replication_slot_name_list_tmp $replication_slot_name" fi done fi echo $replication_slot_name_list } delete_replication_slot(){ DELETE_REPLICATION_SLOT_sql="SELECT pg_drop_replication_slot('$1');" output=`exec_sql "$DELETE_REPLICATION_SLOT_sql"` return $? } delete_replication_slots() { local replication_slot_name_list local replication_slot_name replication_slot_name_list=`create_replication_slot_name` ocf_log debug "replication slot names are $replication_slot_name_list." for replication_slot_name in $replication_slot_name_list do if [ `check_replication_slot $replication_slot_name` = "1" ]; then delete_replication_slot $replication_slot_name if [ $? -eq 0 ]; then ocf_log info "PostgreSQL delete the replication slot($replication_slot_name)." else ocf_exit_reason "$output" return $OCF_ERR_GENERIC fi fi done } create_replication_slots() { local replication_slot_name local replication_slot_name_list local output local rc local CREATE_REPLICATION_SLOT_sql local DELETE_REPLICATION_SLOT_sql replication_slot_name_list=`create_replication_slot_name` ocf_log debug "replication slot names are $replication_slot_name_list." for replication_slot_name in $replication_slot_name_list do # If the same name slot is already exists, initialize(delete and create) the slot. if [ `check_replication_slot $replication_slot_name` = "1" ]; then delete_replication_slot $replication_slot_name if [ $? -eq 0 ]; then ocf_log info "PostgreSQL delete the replication slot($replication_slot_name)." else ocf_exit_reason "$output" return $OCF_ERR_GENERIC fi fi CREATE_REPLICATION_SLOT_sql="SELECT pg_create_physical_replication_slot('$replication_slot_name');" output=`exec_sql "$CREATE_REPLICATION_SLOT_sql"` rc=$? if [ $rc -eq 0 ]; then ocf_log info "PostgreSQL creates the replication slot($replication_slot_name)." else ocf_exit_reason "$output" return $OCF_ERR_GENERIC fi done return 0 } # This function check the replication slot does exists. check_replication_slot(){ local replication_slot_name=$1 local output local CHECK_REPLICATION_SLOT_sql="SELECT count(*) FROM pg_replication_slots WHERE slot_name = '$replication_slot_name'" output=`exec_sql "$CHECK_REPLICATION_SLOT_sql"` echo "$output" } # On postgreSQL 10 or later, "location" means "lsn". get_my_location() { local rc local output local replay_loc local receive_loc local output1 local output2 local log1 local log2 local newer_location output=`exec_sql "$CHECK_XLOG_LOC_SQL"` rc=$? if [ $rc -ne 0 ]; then report_psql_error $rc err "Can't get my xlog location." return 1 fi replay_loc=`echo $output | cut -d "|" -f 1` receive_loc=`echo $output | cut -d "|" -f 2` output1=`echo "$replay_loc" | cut -d "/" -f 1` output2=`echo "$replay_loc" | cut -d "/" -f 2` log1=`printf "%08s\n" $output1 | sed "s/ /0/g"` log2=`printf "%08s\n" $output2 | sed "s/ /0/g"` replay_loc="${log1}${log2}" output1=`echo "$receive_loc" | cut -d "/" -f 1` output2=`echo "$receive_loc" | cut -d "/" -f 2` log1=`printf "%08s\n" $output1 | sed "s/ /0/g"` log2=`printf "%08s\n" $output2 | sed "s/ /0/g"` receive_loc="${log1}${log2}" newer_location=`printf "$replay_loc\n$receive_loc" | sort -r | head -1` echo "$newer_location" return 0 } # On postgreSQL 10 or later, "xlog_location" means "wal_lsn". show_xlog_location() { local location location=`get_my_location` || return 1 exec_with_retry 0 $CRM_ATTR_REBOOT -N "$NODENAME" -n "$PGSQL_XLOG_LOC_NAME" -v "$location" } # On postgreSQL 10 or later, "xlog_location" means "wal_lsn". delete_xlog_location() { exec_with_retry 5 $CRM_ATTR_REBOOT -N "$NODENAME" -n "$PGSQL_XLOG_LOC_NAME" -D } show_master_baseline() { local rc local location location=`get_my_location` ocf_log info "My master baseline : $location." exec_with_retry 0 $CRM_ATTR_REBOOT -N "$NODENAME" -n "$PGSQL_MASTER_BASELINE" -v "$location" } delete_master_baseline() { exec_with_retry 5 $CRM_ATTR_REBOOT -N "$NODENAME" -n "$PGSQL_MASTER_BASELINE" -D } set_async_mode_all() { [ "$OCF_RESKEY_rep_mode" = "sync" ] || return 0 ocf_log info "Set all nodes into async mode." runasowner -q err "echo \"synchronous_standby_names = ''\" > \"$REP_MODE_CONF\"" if [ $? -ne 0 ]; then ocf_exit_reason "Can't set all nodes into async mode." return 1 fi return 0 } set_async_mode() { cat $REP_MODE_CONF | grep -q -E "(\"$1\")|([,' ]$1[,' ])" if [ $? -eq 0 ]; then ocf_log info "Setup $1 into async mode." runasowner -q err "echo \"synchronous_standby_names = ''\" > \"$REP_MODE_CONF\"" else ocf_log debug "$1 is already in async mode." return 0 fi exec_with_retry 0 reload_conf } set_sync_mode() { local sync_node_in_conf sync_node_in_conf=`cat $REP_MODE_CONF | cut -d "'" -f 2` if [ -n "$sync_node_in_conf" ]; then ocf_log debug "$sync_node_in_conf is already sync mode." else ocf_log info "Setup $1 into sync mode." runasowner -q err "echo \"synchronous_standby_names = '\\\"$1\\\"'\" > \"$REP_MODE_CONF\"" [ "$RE_CONTROL_SLAVE" = "false" ] && RE_CONTROL_SLAVE="true" exec_with_retry 0 reload_conf fi } reload_conf() { # Invoke pg_ctl runasowner "$OCF_RESKEY_pgctl -D $OCF_RESKEY_pgdata reload" if [ $? -eq 0 ]; then ocf_log info "Reload configuration file." else ocf_exit_reason "Can't reload configuration file." return 1 fi return 0 } user_recovery_conf() { local nodename_tmp # put archive_cleanup_command and recovery_end_command only when defined by user if [ -n "$OCF_RESKEY_archive_cleanup_command" ]; then echo "archive_cleanup_command = '${OCF_RESKEY_archive_cleanup_command}'" fi if [ -n "$OCF_RESKEY_recovery_end_command" ]; then echo "recovery_end_command = '${OCF_RESKEY_recovery_end_command}'" fi if use_replication_slot; then nodename_tmp=`echo "$NODENAME" | tr 'A-Z.-' 'a-z__'` echo "primary_slot_name = '${OCF_RESKEY_replication_slot_name}_$nodename_tmp'" fi } make_recovery_conf() { runasowner "touch $RECOVERY_CONF" if [ $? -ne 0 ]; then ocf_exit_reason "Can't create recovery.conf." return 1 fi cat > $RECOVERY_CONF <> $RECOVERY_CONF <> $RECOVERY_CONF ocf_log debug "Created recovery.conf. host=${OCF_RESKEY_master_ip}, user=${OCF_RESKEY_repuser}" return 0 } # change pgsql-status. # arg1:node, arg2: value change_pgsql_status() { local output if ! is_node_online $1; then return 0 fi output=`$CRM_ATTR_REBOOT -N "$1" -n "$PGSQL_STATUS_ATTR" -G -q 2>/dev/null` if [ "$output" != "$2" ]; then # If slave's disk is broken, RA cannot read PID file # and misjudges the PostgreSQL as down while it is running. # It causes overwriting of pgsql-status by Master because replication is still connected. if [ "$output" = "STOP" -o "$output" = "UNKNOWN" ]; then if [ "$1" != "$NODENAME" ]; then ocf_log warn "Changing $PGSQL_STATUS_ATTR on $1 : $output->$2 by $NODENAME is prohibited." return 0 fi fi ocf_log info "Changing $PGSQL_STATUS_ATTR on $1 : $output->$2." exec_with_retry 0 $CRM_ATTR_REBOOT -N "$1" -n "$PGSQL_STATUS_ATTR" -v "$2" fi return 0 } # change pgsql-data-status. # arg1:node, arg2: value change_data_status() { local output if ! node_exist $1; then return 0 fi while : do output=`$CRM_ATTR_FOREVER -N "$1" -n "$PGSQL_DATA_STATUS_ATTR" -G -q 2>/dev/null` if [ "$output" != "$2" ]; then ocf_log info "Changing $PGSQL_DATA_STATUS_ATTR on $1 : $output->$2." exec_with_retry 0 exec_with_timeout 0 "$CRM_ATTR_FOREVER" -N $1 -n $PGSQL_DATA_STATUS_ATTR -v "$2" else break fi done return 0 } # set master-score # arg1:node, arg2: score, arg3: resoure set_master_score() { local current_score current_score=`$CRM_ATTR_REBOOT -N "$1" -n "master-$3" -G -q 2>/dev/null` if [ -n "$current_score" -a "$current_score" != "$2" ]; then ocf_log info "Changing $3 master score on $1 : $current_score->$2." exec_with_retry 0 $CRM_ATTR_REBOOT -N "$1" -n "master-$3" -v "$2" fi return 0 } # change master-score # arg1:node, arg2: score change_master_score() { local instance if ! is_node_online $1; then return 0 fi if echo $OCF_RESOURCE_INSTANCE | grep -q ":"; then # If Pacemaker version is 1.0.x instance=0 while : do if [ "$instance" -ge "$OCF_RESKEY_CRM_meta_clone_max" ]; then break fi if [ "${RESOURCE_NAME}:${instance}" = "$OCF_RESOURCE_INSTANCE" ]; then instance=`expr $instance + 1` continue fi set_master_score $1 $2 "${RESOURCE_NAME}:${instance}" || return 1 instance=`expr $instance + 1` done else # If globally-unique=false and Pacemaker version is 1.1.8 or higher # Master/Slave resource has no instance number set_master_score $1 $2 ${RESOURCE_NAME} || return 1 fi return 0 } report_psql_error() { local rc local loglevel local message rc=$1 loglevel=${2:-err} message="$3" ocf_log $loglevel "$message rc=$rc" if [ $rc -eq 1 ]; then ocf_exit_reason "Fatal error (out of memory, file not found, etc.) occurred while executing the psql command." elif [ $rc -eq 2 ]; then ocf_log $loglevel "Connection error (connection to the server went bad and the session was not interactive) occurred while executing the psql command." elif [ $rc -eq 3 ]; then ocf_exit_reason "Script error (the variable ON_ERROR_STOP was set) occurred while executing the psql command." fi } # # timeout management function # arg1 timeout >= 0 (if arg1 is 0, OCF_RESKEY_crm_attr_timeout is used.) # arg2 : command # arg3 : command's args exec_with_timeout() { local func_pid local count=$OCF_RESKEY_crm_attr_timeout local rc if [ "$1" -ne 0 ]; then count=$1 fi shift $* & func_pid=$! sleep .1 while kill -s 0 $func_pid >/dev/null 2>&1; do sleep 1 count=`expr $count - 1` if [ $count -le 0 ]; then ocf_exit_reason "\"$*\" (pid=$func_pid) timed out." kill -s 9 $func_pid >/dev/null 2>&1 return 1 fi ocf_log info "Waiting($count). \"$*\" (pid=$func_pid)." done wait $func_pid } # retry command when command doesn't return 0 # arg1 : count >= 0 (if arg1 is 0, it retries command in infinitum(1day)) # arg2..argN : command and args exec_with_retry() { local count="86400" local output local rc if [ "$1" -ne 0 ]; then count=$1 fi shift while [ $count -gt 0 ]; do output=`$*` rc=$? if [ $rc -ne 0 ]; then ocf_log warn "Retrying(remain $count). \"$*\" failed. rc=$rc. stdout=\"$output\"." count=`expr $count - 1` sleep 1 else printf "${output}" return 0 fi done ocf_exit_reason "giving up executing \"$*\"" return $rc } is_node_online() { print_crm_mon | grep -q -i "/dev/null 2>&1 - if [ ! $? -eq 0 ]; then - ocf_exit_reason "User $OCF_RESKEY_pgdba doesn't exist"; - return $OCF_ERR_INSTALLED; - fi - if ocf_is_probe; then ocf_log info "Don't check $OCF_RESKEY_pgdata during probe" else if ! runasowner "test -w $OCF_RESKEY_pgdata"; then ocf_exit_reason "Directory $OCF_RESKEY_pgdata is not writable by $OCF_RESKEY_pgdba" return $OCF_ERR_PERM; fi fi - if [ -n "$OCF_RESKEY_monitor_user" -a ! -n "$OCF_RESKEY_monitor_password" ] - then - ocf_exit_reason "monitor password can't be empty" - return $OCF_ERR_CONFIGURED - fi - - if [ ! -n "$OCF_RESKEY_monitor_user" -a -n "$OCF_RESKEY_monitor_password" ] - then - ocf_exit_reason "monitor_user has to be set if monitor_password is set" - return $OCF_ERR_CONFIGURED - fi - if is_replication || [ "$OCF_RESKEY_rep_mode" = "slave" ]; then if [ `printf "$version\n9.1" | sort -n | head -1` != "9.1" ]; then ocf_exit_reason "Replication mode needs PostgreSQL 9.1 or higher." return $OCF_ERR_INSTALLED fi ocf_version_cmp "$version" "12" rc=$? if [ $rc -eq 1 ]||[ $rc -eq 2 ]; then # change the standby method for PosrgreSQL 12 or later. USE_STANDBY_SIGNAL=true # change the path to recovery.conf because it cause PostgreSQL start error. RECOVERY_CONF=${OCF_RESKEY_tmpdir}/recovery.conf if [ $check_config_rc -eq 0 ]; then # adding recovery parameters to postgresql.conf. recovery_conf_string="include '$RECOVERY_CONF' # added by pgsql RA" if ! grep -q "^[[:space:]]*$recovery_conf_string" $OCF_RESKEY_config; then ocf_log info "adding include directive $recovery_conf_string into $OCF_RESKEY_config" echo "$recovery_conf_string" >> $OCF_RESKEY_config fi fi fi if [ ! -n "$OCF_RESKEY_master_ip" ]; then ocf_exit_reason "master_ip can't be empty." return $OCF_ERR_CONFIGURED fi fi if is_replication; then REP_MODE_CONF=${OCF_RESKEY_tmpdir}/rep_mode.conf PGSQL_LOCK=${OCF_RESKEY_tmpdir}/PGSQL.lock XLOG_NOTE_FILE=${OCF_RESKEY_tmpdir}/xlog_note CRM_ATTR_REBOOT="${HA_SBIN_DIR}/crm_attribute -l reboot" CRM_ATTR_FOREVER="${HA_SBIN_DIR}/crm_attribute -l forever" CRM_RESOURCE="${HA_SBIN_DIR}/crm_resource" CAN_NOT_PROMOTE="-INFINITY" CAN_PROMOTE="100" PROMOTE_ME="1000" CHECK_MS_SQL="select pg_is_in_recovery()" CHECK_SYNCHRONOUS_STANDBY_NAMES_SQL="show synchronous_standby_names" ocf_version_cmp "$version" "10" rc=$? if [ $rc -eq 1 ]||[ $rc -eq 2 ]; then CHECK_XLOG_LOC_SQL="select pg_last_wal_replay_lsn(),pg_last_wal_receive_lsn()" else CHECK_XLOG_LOC_SQL="select pg_last_xlog_replay_location(),pg_last_xlog_receive_location()" fi CHECK_REPLICATION_STATE_SQL="select application_name,upper(state),upper(sync_state) from pg_stat_replication" PGSQL_STATUS_ATTR="${RESOURCE_NAME}-status" PGSQL_DATA_STATUS_ATTR="${RESOURCE_NAME}-data-status" PGSQL_XLOG_LOC_NAME="${RESOURCE_NAME}-xlog-loc" PGSQL_MASTER_BASELINE="${RESOURCE_NAME}-master-baseline" NODE_LIST=`echo $OCF_RESKEY_node_list | tr '[A-Z]' '[a-z]'` RE_CONTROL_SLAVE="false" if ! ocf_is_ms; then ocf_exit_reason "Replication(rep_mode=async or sync) requires Master/Slave configuration." return $OCF_ERR_CONFIGURED fi if [ ! "$OCF_RESKEY_rep_mode" = "sync" -a ! "$OCF_RESKEY_rep_mode" = "async" ]; then ocf_exit_reason "Invalid rep_mode : $OCF_RESKEY_rep_mode" return $OCF_ERR_CONFIGURED fi if [ ! -n "$NODE_LIST" ]; then ocf_exit_reason "node_list can't be empty." return $OCF_ERR_CONFIGURED fi if [ $check_config_rc -eq 0 ]; then rep_mode_string="include '$REP_MODE_CONF' # added by pgsql RA" if [ "$OCF_RESKEY_rep_mode" = "sync" ]; then if ! grep -q "^[[:space:]]*$rep_mode_string" $OCF_RESKEY_config; then ocf_log info "adding include directive into $OCF_RESKEY_config" echo "$rep_mode_string" >> $OCF_RESKEY_config fi else if grep -q "$rep_mode_string" $OCF_RESKEY_config; then ocf_log info "deleting include directive from $OCF_RESKEY_config" rep_mode_string=`echo $rep_mode_string | sed -e 's|/|\\\\/|g'` sed -i "/$rep_mode_string/d" $OCF_RESKEY_config fi fi fi if ! mkdir -p $OCF_RESKEY_tmpdir || ! chown $OCF_RESKEY_pgdba $OCF_RESKEY_tmpdir || ! chmod 700 $OCF_RESKEY_tmpdir; then ocf_exit_reason "Can't create directory $OCF_RESKEY_tmpdir or it is not readable by $OCF_RESKEY_pgdba" return $OCF_ERR_PERM fi fi if [ "$OCF_RESKEY_rep_mode" = "slave" ]; then if ocf_is_ms; then ocf_exit_reason "Replication(rep_mode=slave) does not support Master/Slave configuration." return $OCF_ERR_CONFIGURED fi fi if use_replication_slot; then ocf_version_cmp "$version" "9.4" rc=$? if [ $rc -eq 0 ]||[ $rc -eq 3 ]; then ocf_exit_reason "Replication slot needs PostgreSQL 9.4 or higher." return $OCF_ERR_CONFIGURED fi echo "$OCF_RESKEY_replication_slot_name" | grep -q -e '[^a-z0-9_]' if [ $? -eq 0 ]; then ocf_exit_reason "Invalid replication_slot_name($OCF_RESKEY_replication_slot_name). only use lower case letters, numbers, and the underscore character." return $OCF_ERR_CONFIGURED fi fi return $OCF_SUCCESS } +# Validate most critical parameters +pgsql_validate_all() { + local rc + + getent passwd $OCF_RESKEY_pgdba >/dev/null 2>&1 + if [ ! $? -eq 0 ]; then + ocf_exit_reason "User $OCF_RESKEY_pgdba doesn't exist"; + return $OCF_ERR_INSTALLED; + fi + + if [ -n "$OCF_RESKEY_monitor_user" ] && [ -z "$OCF_RESKEY_monitor_password" ]; then + ocf_exit_reason "monitor password can't be empty" + return $OCF_ERR_CONFIGURED + fi + + if [ -z "$OCF_RESKEY_monitor_user" ] && [ -n "$OCF_RESKEY_monitor_password" ]; then + ocf_exit_reason "monitor_user has to be set if monitor_password is set" + return $OCF_ERR_CONFIGURED + fi + + if [ "$OCF_CHECK_LEVEL" -eq 10 ]; then + validate_ocf_check_level_10 + rc=$? + [ $rc -ne "$OCF_SUCCESS" ] && exit $rc + fi + + return $OCF_SUCCESS +} + # # Check if we need to create a log file # check_log_file() { if [ ! -e "$1" ] then touch $1 > /dev/null 2>&1 chown $OCF_RESKEY_pgdba:`getent passwd $OCF_RESKEY_pgdba | cut -d ":" -f 4` $1 fi #Check if $OCF_RESKEY_pgdba can write to the log file if ! runasowner "test -w $1" then return 1 fi return 0 } # # Check if we need to create stats temp directory in tmpfs # check_stat_temp_directory() { local stats_temp stats_temp=`get_pgsql_param stats_temp_directory` if [ -z "$stats_temp" ]; then return fi if [ "${stats_temp#/}" = "$stats_temp" ]; then stats_temp="$OCF_RESKEY_pgdata/$stats_temp" fi if [ -d "$stats_temp" ]; then return fi if ! mkdir -p "$stats_temp"; then ocf_exit_reason "Can't create directory $stats_temp" exit $OCF_ERR_PERM fi if ! chown $OCF_RESKEY_pgdba: "$stats_temp"; then ocf_exit_reason "Can't change ownership for $stats_temp" exit $OCF_ERR_PERM fi if ! chmod 700 "$stats_temp"; then ocf_exit_reason "Can't change permissions for $stats_temp" exit $OCF_ERR_PERM fi } # # Check socket directory # check_socket_dir() { if [ ! -d "$OCF_RESKEY_socketdir" ]; then if ! mkdir "$OCF_RESKEY_socketdir"; then ocf_exit_reason "Can't create directory $OCF_RESKEY_socketdir" exit $OCF_ERR_PERM fi if ! chown $OCF_RESKEY_pgdba:`getent passwd \ $OCF_RESKEY_pgdba | cut -d ":" -f 4` "$OCF_RESKEY_socketdir" then ocf_exit_reason "Can't change ownership for $OCF_RESKEY_socketdir" exit $OCF_ERR_PERM fi if ! chmod 2775 "$OCF_RESKEY_socketdir"; then ocf_exit_reason "Can't change permissions for $OCF_RESKEY_socketdir" exit $OCF_ERR_PERM fi else if ! runasowner "touch $OCF_RESKEY_socketdir/test.$$"; then ocf_exit_reason "$OCF_RESKEY_pgdba can't create files in $OCF_RESKEY_socketdir" exit $OCF_ERR_PERM fi rm $OCF_RESKEY_socketdir/test.$$ fi } print_crm_mon() { if [ -z "$CRM_MON_OUTPUT" ]; then ocf_version_cmp "$OCF_RESKEY_crm_feature_set" "3.1.0" res=$? if [ -z "$OCF_RESKEY_crm_feature_set" ] || [ $res -eq 2 ]; then XMLOPT="--output-as=xml" ocf_version_cmp "$OCF_RESKEY_crm_feature_set" "3.2.0" if [ $? -eq 1 ]; then crm_mon -1 $XMLOPT >/dev/null 2>&1 if [ $? -ne 0 ]; then XMLOPT="--as-xml" fi fi else XMLOPT="--as-xml" fi CRM_MON_OUTPUT=`exec_with_retry 0 crm_mon -1 $XMLOPT` fi printf "${CRM_MON_OUTPUT}\n" } # # 'main' starts here... # if [ $# -ne 1 ] then usage exit $OCF_ERR_GENERIC fi PIDFILE=${OCF_RESKEY_pgdata}/postmaster.pid BACKUPLABEL=${OCF_RESKEY_pgdata}/backup_label RESOURCE_NAME=`echo $OCF_RESOURCE_INSTANCE | cut -d ":" -f 1` PGSQL_WAL_RECEIVER_STATUS_ATTR="${RESOURCE_NAME}-receiver-status" RECOVERY_CONF=${OCF_RESKEY_pgdata}/recovery.conf NODENAME=$(ocf_local_nodename | tr '[A-Z]' '[a-z]') USE_STANDBY_SIGNAL=false case "$1" in methods) pgsql_methods exit $?;; meta-data) meta_data exit $OCF_SUCCESS;; esac +[ "$__OCF_ACTION" != "validate-all" ] && OCF_CHECK_LEVEL=10 pgsql_validate_all rc=$? [ "$1" = "validate-all" ] && exit $rc if [ $rc -ne 0 ] then case "$1" in stop) if is_replication; then change_pgsql_status "$NODENAME" "UNKNOWN" fi exit $OCF_SUCCESS;; monitor) exit $OCF_NOT_RUNNING;; status) exit $OCF_NOT_RUNNING;; *) exit $rc;; esac fi US=`id -u -n` if [ $US != root -a $US != $OCF_RESKEY_pgdba ] then ocf_exit_reason "$0 must be run as root or $OCF_RESKEY_pgdba" exit $OCF_ERR_GENERIC fi # make psql command options if [ -n "$OCF_RESKEY_monitor_user" ]; then PGUSER=$OCF_RESKEY_monitor_user; export PGUSER PGPASSWORD=$OCF_RESKEY_monitor_password; export PGPASSWORD psql_options="-p $OCF_RESKEY_pgport $OCF_RESKEY_pgdb" else psql_options="-p $OCF_RESKEY_pgport -U $OCF_RESKEY_pgdba $OCF_RESKEY_pgdb" fi if [ -n "$OCF_RESKEY_pghost" ]; then psql_options="$psql_options -h $OCF_RESKEY_pghost" else if [ -n "$OCF_RESKEY_socketdir" ]; then psql_options="$psql_options -h $OCF_RESKEY_socketdir" fi fi if [ -n "$OCF_RESKEY_pgport" ]; then export PGPORT=$OCF_RESKEY_pgport fi if [ -n "$OCF_RESKEY_pglibs" ]; then if [ -n "$LD_LIBRARY_PATH" ]; then export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$OCF_RESKEY_pglibs else export LD_LIBRARY_PATH=$OCF_RESKEY_pglibs fi fi # What kind of method was invoked? case "$1" in status) if pgsql_status then ocf_log info "PostgreSQL is up" exit $OCF_SUCCESS else ocf_log info "PostgreSQL is down" exit $OCF_NOT_RUNNING fi;; monitor) pgsql_monitor exit $?;; start) pgsql_start exit $?;; promote) pgsql_promote exit $?;; demote) pgsql_demote exit $?;; notify) pgsql_notify exit $?;; stop) pgsql_stop exit $?;; *) exit $OCF_ERR_UNIMPLEMENTED;; esac