diff --git a/heartbeat/mysql b/heartbeat/mysql index b59bb2688..0d85d6045 100755 --- a/heartbeat/mysql +++ b/heartbeat/mysql @@ -1,1039 +1,1069 @@ #!/bin/sh # # # MySQL # # Description: Manages a MySQL database as Linux-HA resource # # Authors: Alan Robertson: DB2 Script # Jakub Janczak: rewrite as MySQL # Andrew Beekhof: cleanup and import # Sebastian Reitenbach: add OpenBSD defaults, more cleanup # Narayan Newton: add Gentoo/Debian defaults # Marian Marinov, Florian Haas: add replication capability # Yves Trudeau, Baron Schwartz: add VIP support and improve replication # # Support: users@clusterlabs.org # License: GNU General Public License (GPL) # # (c) 2002-2005 International Business Machines, Inc. # 2005-2010 Linux-HA contributors # # An example usage in /etc/ha.d/haresources: # node1 10.0.0.170 mysql # # See usage() function below for more details... # # OCF instance parameters: # OCF_RESKEY_binary # OCF_RESKEY_client_binary # OCF_RESKEY_config # OCF_RESKEY_datadir # OCF_RESKEY_user # OCF_RESKEY_group # OCF_RESKEY_test_table # OCF_RESKEY_test_user # OCF_RESKEY_test_passwd # OCF_RESKEY_enable_creation # OCF_RESKEY_additional_parameters # OCF_RESKEY_log # OCF_RESKEY_pid # OCF_RESKEY_socket # OCF_RESKEY_replication_user # OCF_RESKEY_replication_passwd # OCF_RESKEY_replication_port # OCF_RESKEY_max_slave_lag # OCF_RESKEY_evict_outdated_slaves # OCF_RESKEY_reader_attribute ####################################################################### # Initialization: : ${OCF_FUNCTIONS_DIR=${OCF_ROOT}/lib/heartbeat} . ${OCF_FUNCTIONS_DIR}/ocf-shellfuncs . ${OCF_FUNCTIONS_DIR}/mysql-common.sh ####################################################################### usage() { cat < 1.0 Resource script for MySQL. May manage a standalone MySQL database, a clone set with externally managed replication, or a complete master/slave replication setup. Note, when master/slave replication is in use, the resource must be setup to use notifications. Set 'notify=true' in the metadata attributes when defining a MySQL master/slave instance. -While managing replication, the default behavior is to use uname -n -values in the change master to command. Other IPs can be specified +While managing replication, the default behavior is to use uname -n +values in the change master to command. Other IPs can be specified manually by adding a node attribute \${INSTANCE_ATTR_NAME}_mysql_master_IP -giving the IP to use for replication. For example, if the mysql primitive -you are using is p_mysql, the attribute to set will be +giving the IP to use for replication. For example, if the mysql primitive +you are using is p_mysql, the attribute to set will be p_mysql_mysql_master_IP. Manages a MySQL database instance Location of the MySQL server binary MySQL server binary Location of the MySQL client binary MySQL client binary Configuration file MySQL config Directory containing databases MySQL datadir User running MySQL daemon MySQL user Group running MySQL daemon (for logfile and directory permissions) MySQL group The logfile to be used for mysqld. MySQL log file The pidfile to be used for mysqld. MySQL pid file The socket to be used for mysqld. MySQL socket Table to be tested in monitor statement (in database.table notation) MySQL test table MySQL test user, must have select privilege on test_table MySQL test user MySQL test user password MySQL test user password If the MySQL database does not exist, it will be created Create the database if it does not exist Additional parameters which are passed to the mysqld on startup. (e.g. --skip-external-locking or --skip-grant-tables) Additional parameters to pass to mysqld MySQL replication user. This user is used for starting and stopping MySQL replication, for setting and resetting the master host, and for setting and unsetting read-only mode. Because of that, this user must have SUPER, REPLICATION SLAVE, REPLICATION CLIENT, PROCESS and RELOAD privileges on all nodes within the cluster. Mandatory if you define a master-slave resource. MySQL replication user MySQL replication password. Used for replication client and slave. Mandatory if you define a master-slave resource. MySQL replication user password The port on which the Master MySQL instance is listening. MySQL replication port + + +The SSL CA certificate to be used for replication over SSL. + +MySQL replication SSL CA certificate + + + + + +The SSL CA certificate to be used for replication over SSL. + +MySQL replication SSL certificate + + + + + +The SSL certificate key to be used for replication over SSL. + +MySQL replication SSL certificate key + + + The maximum number of seconds a replication slave is allowed to lag behind its master. Do not set this to zero. What the cluster manager does in case a slave exceeds this maximum lag is determined by the evict_outdated_slaves parameter. Maximum time (seconds) a MySQL slave is allowed to lag behind a master If set to true, any slave which is more than max_slave_lag seconds behind the master has its MySQL instance shut down. If this parameter is set to false in a primitive or clone resource, it is simply ignored. If set to false in a master/slave resource, then exceeding the maximum slave lag will merely push down the master preference so the lagging slave is never promoted to the new master. Determines whether to shut down badly lagging slaves An attribute that the RA can manage to specify whether a node can be read from. This node attribute will be 1 if it's fine to read from the node, and 0 otherwise (for example, when a slave has lagged too far behind the master). A typical example for the use of this attribute would be to tie a set of IP addresses to MySQL slaves that can be read from. This parameter is only meaningful in master/slave set configurations. Sets the node attribute that determines whether a node is usable for clients to read from. END } # Convenience functions set_read_only() { # Sets or unsets read-only mode. Accepts one boolean as its # optional argument. If invoked without any arguments, defaults to # enabling read only mode. Should only be set in master/slave # setups. # Returns $OCF_SUCCESS if the operation succeeds, or # $OCF_ERR_GENERIC if it fails. local ro_val if ocf_is_true $1; then ro_val="on" else ro_val="off" fi ocf_run $MYSQL $MYSQL_OPTIONS_REPL \ -e "SET GLOBAL read_only=${ro_val}" } get_read_only() { # Check if read-only is set local read_only_state read_only_state=`$MYSQL $MYSQL_OPTIONS_REPL \ --skip-column-names -e "SHOW VARIABLES LIKE 'read_only'" | awk '{print $2}'` if [ "$read_only_state" = "ON" ]; then return 0 else return 1 fi } is_slave() { # Determine whether the machine is currently running as a MySQL # slave, as determined per SHOW SLAVE STATUS. Returns 1 if SHOW # SLAVE STATUS creates an empty result set, 0 otherwise. local rc local tmpfile # Check whether this machine should be slave if ! ocf_is_ms || ! get_read_only; then return 1 fi - + get_slave_info rc=$? rm -f $tmpfile if [ $rc -eq 0 ]; then # show slave status is not empty - # Is there a master_log_file defined? (master_log_file is deleted + # Is there a master_log_file defined? (master_log_file is deleted # by reset slave if [ "$master_log_file" ]; then return 0 else return 1 fi else # "SHOW SLAVE STATUS" returns an empty set if instance is not a # replication slave return 1 fi - + } parse_slave_info() { # Extracts field $1 from result of "SHOW SLAVE STATUS\G" from file $2 sed -ne "s/^.* $1: \(.*\)$/\1/p" < $2 } get_slave_info() { # Warning: this sets $tmpfile and LEAVE this file! You must delete it after use! local mysql_options - + if [ "$master_log_file" -a "$master_host" ]; then # variables are already defined, get_slave_info has been run before return $OCF_SUCCESS else tmpfile=`mktemp ${HA_RSCTMP}/check_slave.${OCF_RESOURCE_INSTANCE}.XXXXXX` $MYSQL $MYSQL_OPTIONS_REPL \ -e 'SHOW SLAVE STATUS\G' > $tmpfile if [ -s $tmpfile ]; then master_host=`parse_slave_info Master_Host $tmpfile` master_user=`parse_slave_info Master_User $tmpfile` master_port=`parse_slave_info Master_Port $tmpfile` master_log_file=`parse_slave_info Master_Log_File $tmpfile` master_log_pos=`parse_slave_info Read_Master_Log_Pos $tmpfile` slave_sql=`parse_slave_info Slave_SQL_Running $tmpfile` slave_io=`parse_slave_info Slave_IO_Running $tmpfile` last_errno=`parse_slave_info Last_Errno $tmpfile` secs_behind=`parse_slave_info Seconds_Behind_Master $tmpfile` ocf_log debug "MySQL instance running as a replication slave" else # Instance produced an empty "SHOW SLAVE STATUS" output -- # instance is not a slave ocf_exit_reason "check_slave invoked on an instance that is not a replication slave." return $OCF_ERR_GENERIC fi return $OCF_SUCCESS fi } check_slave() { # Checks slave status local rc new_master get_slave_info rc=$? if [ $rc -eq 0 ]; then # Did we receive an error other than max_connections? if [ $last_errno -ne 0 -a $last_errno -ne "$MYSQL_TOO_MANY_CONN_ERR" ]; then # Whoa. Replication ran into an error. This slave has # diverged from its master. Make sure this resource # doesn't restart in place. ocf_exit_reason "MySQL instance configured for replication, but replication has failed." ocf_log err "See $tmpfile for details" # Just pull the reader VIP away, killing MySQL here would be pretty evil # on a loaded server set_reader_attr 0 exit $OCF_SUCCESS fi # If we got max_connections, let's remove the vip if [ $last_errno -eq "$MYSQL_TOO_MANY_CONN_ERR" ]; then set_reader_attr 0 exit $OCF_SUCCESS fi if [ "$slave_io" != 'Yes' ]; then # Not necessarily a bad thing. The master may have # temporarily shut down, and the slave may just be # reconnecting. A warning can't hurt, though. ocf_log warn "MySQL Slave IO threads currently not running." # Sanity check, are we at least on the right master new_master=`$CRM_ATTR_REPL_INFO --query -q | cut -d'|' -f1` if [ "$master_host" != "$new_master" ]; then # Not pointing to the right master, not good, removing the VIPs set_reader_attr 0 exit $OCF_SUCCESS fi fi if [ "$slave_sql" != 'Yes' ]; then # We don't have a replication SQL thread running. Not a # good thing. Try to recoved by restarting the SQL thread # and remove reader vip. Prevent MySQL restart. ocf_exit_reason "MySQL Slave SQL threads currently not running." ocf_log err "See $tmpfile for details" # Remove reader vip set_reader_attr 0 # try to restart slave ocf_run $MYSQL $MYSQL_OPTIONS_REPL \ -e "START SLAVE" # Return success to prevent a restart exit $OCF_SUCCESS fi if ocf_is_true $OCF_RESKEY_evict_outdated_slaves; then # We're supposed to bail out if we lag too far # behind. Let's check our lag. if [ "$secs_behind" = "NULL" ] || [ $secs_behind -gt $OCF_RESKEY_max_slave_lag ]; then ocf_exit_reason "MySQL Slave is $secs_behind seconds behind master (allowed maximum: $OCF_RESKEY_max_slave_lag)." ocf_log err "See $tmpfile for details" # Remove reader vip set_reader_attr 0 exit $OCF_ERR_INSTALLED fi elif ocf_is_ms; then # Even if we're not set to evict lagging slaves, we can # still use the seconds behind master value to set our # master preference. local master_pref master_pref=$((${OCF_RESKEY_max_slave_lag}-${secs_behind})) if [ $master_pref -lt 0 ]; then # Sanitize a below-zero preference to just zero master_pref=0 fi $CRM_MASTER -v $master_pref fi # is the slave ok to have a VIP on it if [ "$secs_behind" = "NULL" ] || [ $secs_behind -gt $OCF_RESKEY_max_slave_lag ]; then set_reader_attr 0 else set_reader_attr 1 fi ocf_log debug "MySQL instance running as a replication slave" rm -f $tmpfile else # Instance produced an empty "SHOW SLAVE STATUS" output -- # instance is not a slave # TODO: Needs to handle when get_slave_info will return too many connections error rm -f $tmpfile ocf_exit_reason "check_slave invoked on an instance that is not a replication slave." exit $OCF_ERR_GENERIC fi } set_master() { local new_master master_log_file master_log_pos - local master_params + local master_params master_ssl_params new_master=`$CRM_ATTR_REPL_INFO --query -q | cut -d'|' -f1` # Keep replication position get_slave_info if [ "$master_log_file" -a "$new_master" = "$master_host" ]; then # master_params=", MASTER_LOG_FILE='$master_log_file', \ # MASTER_LOG_POS=$master_log_pos" ocf_log info "Kept master pos for $master_host : $master_log_file:$master_log_pos" rm -f $tmpfile return else master_log_file=`$CRM_ATTR_REPL_INFO --query -q | cut -d'|' -f2` master_log_pos=`$CRM_ATTR_REPL_INFO --query -q | cut -d'|' -f3` if [ -n "$master_log_file" -a -n "$master_log_pos" ]; then master_params=", MASTER_LOG_FILE='$master_log_file', \ MASTER_LOG_POS=$master_log_pos" ocf_log info "Restored master pos for $new_master : $master_log_file:$master_log_pos" fi fi # Informs the MySQL server of the master to replicate # from. Accepts one mandatory argument which must contain the host # name of the new master host. The master must either be unchanged - # from the laste master the slave replicated from, or freshly + # from the last master the slave replicated from, or freshly # reset with RESET MASTER. + if [ -n "$OCF_RESKEY_replication_master_ssl_ca" ] && [ -n "$OCF_RESKEY_replication_master_ssl_cert" ] && [ -n "$OCF_RESKEY_replication_master_ssl_key" ]; then + master_ssl_params=", MASTER_SSL=1, \ + MASTER_SSL_CA='$OCF_RESKEY_replication_master_ssl_ca', \ + MASTER_SSL_CERT='$OCF_RESKEY_replication_master_ssl_cert', \ + MASTER_SSL_KEY='$OCF_RESKEY_replication_master_ssl_key'" + fi ocf_run $MYSQL $MYSQL_OPTIONS_REPL \ -e "CHANGE MASTER TO MASTER_HOST='$new_master', \ MASTER_PORT=$OCF_RESKEY_replication_port, \ MASTER_USER='$OCF_RESKEY_replication_user', \ - MASTER_PASSWORD='$OCF_RESKEY_replication_passwd' $master_params" + MASTER_PASSWORD='$OCF_RESKEY_replication_passwd' $master_params $master_ssl_params" rm -f $tmpfile } unset_master(){ # Instructs the MySQL server to stop replicating from a master # host. # If we're currently not configured to be replicating from any # host, then there's nothing to do. But we do log a warning as # no-one but the CRM should be touching the MySQL master/slave # configuration. if ! is_slave; then ocf_log warn "Attempted to unset the replication master on an instance that is not configured as a replication slave" return $OCF_SUCCESS fi local tmpfile tmpfile=`mktemp ${HA_RSCTMP}/unset_master.${OCF_RESOURCE_INSTANCE}.XXXXXX` # At this point, the master is read only so there should not be much binlogs to transfer # Let's wait for the last bits while true; do $MYSQL $MYSQL_OPTIONS_REPL \ -e 'SHOW PROCESSLIST\G' > $tmpfile if grep -i 'Waiting for master to send event' $tmpfile >/dev/null; then ocf_log info "MySQL slave has finished reading master binary log" break fi if grep -i 'Reconnecting after a failed master event read' $tmpfile >/dev/null; then ocf_log info "Master is down, no more binary logs to come" break fi if grep -i 'Connecting to master' $tmpfile >/dev/null; then ocf_log info "Master is down, no more binary logs to come" break fi if ! grep 'system user' $tmpfile >/dev/null; then ocf_log info "Slave is not running - not waiting to finish" break fi sleep 1 done # Now, stop the slave I/O thread and wait for relay log # processing to complete ocf_run $MYSQL $MYSQL_OPTIONS_REPL \ -e "STOP SLAVE IO_THREAD" if [ $? -gt 0 ]; then ocf_exit_reason "Error stopping slave IO thread" exit $OCF_ERR_GENERIC fi while true; do $MYSQL $MYSQL_OPTIONS_REPL \ -e 'SHOW PROCESSLIST\G' > $tmpfile if grep -i 'Has read all relay log' $tmpfile >/dev/null; then ocf_log info "MySQL slave has finished processing relay log" break fi if ! grep -q 'system user' $tmpfile; then ocf_log info "Slave not runnig - not waiting to finish" break fi ocf_log info "Waiting for MySQL slave to finish processing relay log" sleep 1 done rm -f $tmpfile # Now, stop all slave activity and unset the master host ocf_run $MYSQL $MYSQL_OPTIONS_REPL \ -e "STOP SLAVE" if [ $? -gt 0 ]; then ocf_exit_reason "Error stopping rest slave threads" exit $OCF_ERR_GENERIC fi ocf_run $MYSQL $MYSQL_OPTIONS_REPL \ -e "RESET SLAVE /*!50516 ALL */;" if [ $? -gt 0 ]; then ocf_exit_reason "Failed to reset slave" exit $OCF_ERR_GENERIC fi } # Start replication as slave start_slave() { ocf_run $MYSQL $MYSQL_OPTIONS_REPL \ -e "START SLAVE" } # Set the attribute controlling the readers VIP set_reader_attr() { local curr_attr_value curr_attr_value=$(get_reader_attr) if [ "$curr_attr_value" -ne "$1" ]; then $CRM_ATTR -l reboot --name ${OCF_RESKEY_reader_attribute} -v $1 fi } # get the attribute controlling the readers VIP get_reader_attr() { local attr_value local rc attr_value=`$CRM_ATTR -l reboot --name ${OCF_RESKEY_reader_attribute} --query -q` rc=$? if [ "$rc" -eq "0" ]; then echo $attr_value else echo -1 fi } # Stores data for MASTER STATUS from MySQL update_data_master_status() { master_status_file="${HA_RSCTMP}/master_status.${OCF_RESOURCE_INSTANCE}" $MYSQL $MYSQL_OPTIONS_REPL -e "SHOW MASTER STATUS\G" > $master_status_file } # Returns the specified value from the stored copy of SHOW MASTER STATUS. # should be call after update_data_master_status for tmpfile # Arguments: # $1 The value to get. get_master_status() { awk -v var="$1" '$1 == var ":" {print substr($0, index($0, ":") + 2)}' "$master_status_file" } # Determines what IP address is attached to the current host. The output of the # crm_attribute command looks like this: # scope=nodes name=IP value=10.2.2.161 # If the ${INSTANCE_ATTR_NAME}_MYSQL_MASTER_IP node attribute is not defined, fallback is to uname -n -# The ${INSTANCE_ATTR_NAME}_MYSQL_MASTER_IP is the IP address that will be used for the +# The ${INSTANCE_ATTR_NAME}_MYSQL_MASTER_IP is the IP address that will be used for the # change master to command. get_local_ip() { local IP IP=`$CRM_ATTR -l forever -n ${INSTANCE_ATTR_NAME}_mysql_master_IP -q -G` if [ ! $? -eq 0 ]; then uname -n else echo $IP fi } ####################################################################### # Functions invoked by resource manager actions mysql_monitor() { local rc local status_loglevel="err" # Set loglevel to info during probe if ocf_is_probe; then status_loglevel="info" fi - + mysql_common_status $status_loglevel rc=$? # TODO: check max connections error # If status returned an error, return that immediately if [ $rc -ne $OCF_SUCCESS ]; then if ocf_is_ms ; then # This is a master slave setup but monitored host returned some errors. # Immediately remove it from the pool of possible masters by erasing its master-mysql key # When new mysql master election is started and node got no or negative master-mysql attribute the following is logged # nodename.com pengine: debug: master_color: mysql:0 master score: -1 # If there are NO nodes with positive vaule election of mysql master will fail with # nodename.com pengine: info: master_color: ms_mysql: Promoted 0 instances of a possible 1 to master $CRM_MASTER -D fi return $rc fi if [ $OCF_CHECK_LEVEL -gt 0 -a -n "$OCF_RESKEY_test_table" ]; then # Check if this instance is configured as a slave, and if so # check slave status if is_slave; then check_slave fi # Check for test table ocf_run -q $MYSQL $MYSQL_OPTIONS_TEST \ -e "SELECT COUNT(*) FROM $OCF_RESKEY_test_table" rc=$? if [ $rc -ne 0 ]; then # We are master/slave and test failed. Delete master score for this node as it is considered unhealthy because of this particular failed check. ocf_is_ms && $CRM_MASTER -D ocf_exit_reason "Failed to select from $test_table"; return $OCF_ERR_GENERIC; fi else # In case no exnteded tests are enabled and we are in master/slave mode _always_ set the master score to 1 if we reached this point ocf_is_ms && $CRM_MASTER -v 1 fi if ocf_is_ms && ! get_read_only; then ocf_log debug "MySQL monitor succeeded (master)"; # Always set master score for the master $CRM_MASTER -v 2 return $OCF_RUNNING_MASTER else ocf_log debug "MySQL monitor succeeded"; return $OCF_SUCCESS fi } mysql_start() { local rc if ocf_is_ms; then # Initialize the ReaderVIP attribute, monitor will enable it set_reader_attr 0 fi mysql_common_status info if [ $? = $OCF_SUCCESS ]; then ocf_log info "MySQL already running" return $OCF_SUCCESS fi mysql_common_prepare_dirs # Uncomment to perform permission clensing # - not convinced this should be enabled by default # #chmod 0755 $OCF_RESKEY_datadir #chown -R $OCF_RESKEY_user $OCF_RESKEY_datadir #chgrp -R $OCF_RESKEY_group $OCF_RESKEY_datadir mysql_extra_params= if ocf_is_ms; then mysql_extra_params="--skip-slave-start" fi mysql_common_start $mysql_extra_params rc=$? if [ $rc != $OCF_SUCCESS ]; then return $rc fi if ocf_is_ms; then # We're configured as a stateful resource. We must start as # slave by default. At this point we don't know if the CRM has # already promoted a master. So, we simply start in read only # mode. set_read_only on # Now, let's see whether there is a master. We might be a new # node that is just joining the cluster, and the CRM may have # promoted a master before. master_host=`echo $OCF_RESKEY_CRM_meta_notify_master_uname|tr -d " "` if [ "$master_host" -a "$master_host" != ${NODENAME} ]; then ocf_log info "Changing MySQL configuration to replicate from $master_host." set_master start_slave if [ $? -ne 0 ]; then ocf_exit_reason "Failed to start slave" return $OCF_ERR_GENERIC fi else ocf_log info "No MySQL master present - clearing replication state" unset_master fi # We also need to set a master preference, otherwise Pacemaker # won't ever promote us in the absence of any explicit # preference set by the administrator. We choose a low # greater-than-zero preference. $CRM_MASTER -v 1 fi # Initial monitor action if [ -n "$OCF_RESKEY_test_table" -a -n "$OCF_RESKEY_test_user" -a -n "$OCF_RESKEY_test_passwd" ]; then OCF_CHECK_LEVEL=10 fi mysql_monitor rc=$? if [ $rc != $OCF_SUCCESS -a $rc != $OCF_RUNNING_MASTER ]; then ocf_exit_reason "Failed initial monitor action" return $rc fi ocf_log info "MySQL started" return $OCF_SUCCESS } mysql_stop() { if ocf_is_ms; then # clear preference for becoming master $CRM_MASTER -D # Remove VIP capability set_reader_attr 0 fi mysql_common_stop } mysql_promote() { local master_info if ( ! mysql_common_status err ); then return $OCF_NOT_RUNNING fi ocf_run $MYSQL $MYSQL_OPTIONS_REPL \ -e "STOP SLAVE" # Set Master Info in CIB, cluster level attribute update_data_master_status master_info="$(get_local_ip)|$(get_master_status File)|$(get_master_status Position)" ${CRM_ATTR_REPL_INFO} -v "$master_info" rm -f $tmpfile set_read_only off || return $OCF_ERR_GENERIC # Existing master gets a higher-than-default master preference, so # the cluster manager does not shuffle the master role around # unnecessarily $CRM_MASTER -v $((${OCF_RESKEY_max_slave_lag}+1)) # A master can accept reads set_reader_attr 1 return $OCF_SUCCESS } mysql_demote() { if ! mysql_common_status err; then return $OCF_NOT_RUNNING fi # Return master preference to default, so the cluster manager gets # a chance to select a new master $CRM_MASTER -v 1 } mysql_notify() { # If not configured as a Stateful resource, we make no sense of # notifications. if ! ocf_is_ms; then ocf_log info "This agent makes no use of notifications unless running in master/slave mode." return $OCF_SUCCESS fi local type_op type_op="${OCF_RESKEY_CRM_meta_notify_type}-${OCF_RESKEY_CRM_meta_notify_operation}" ocf_log debug "Received $type_op notification." case "$type_op" in 'pre-promote') # Nothing to do now here, new replication info not yet published ;; 'post-promote') # The master has completed its promotion. Now is a good # time to check whether our replication slave is working # correctly. master_host=`echo $OCF_RESKEY_CRM_meta_notify_promote_uname|tr -d " "` if [ "$master_host" = ${NODENAME} ]; then ocf_log info "This will be the new master, ignoring post-promote notification." else ocf_log info "Resetting replication" unset_master if [ $? -ne 0 ]; then return $OCF_ERR_GENERIC fi ocf_log info "Changing MySQL configuration to replicate from $master_host" set_master if [ $? -ne 0 ]; then return $OCF_ERR_GENERIC fi start_slave if [ $? -ne 0 ]; then ocf_exit_reason "Failed to start slave" return $OCF_ERR_GENERIC fi fi return $OCF_SUCCESS ;; 'pre-demote') demote_host=`echo $OCF_RESKEY_CRM_meta_notify_demote_uname|tr -d " "` if [ $demote_host = ${NODENAME} ]; then ocf_log info "post-demote notification for $demote_host" set_read_only on if [ $? -ne 0 ]; then ocf_exit_reason "Failed to set read-only"; return $OCF_ERR_GENERIC; fi # Must kill all existing user threads because they are still Read/write # in order for the slaves to complete the read of binlogs local tmpfile tmpfile=`mktemp ${HA_RSCTMP}/threads.${OCF_RESOURCE_INSTANCE}.XXXXXX` $MYSQL $MYSQL_OPTIONS_REPL \ -e "SHOW PROCESSLIST" > $tmpfile for thread in `awk '$0 !~ /Binlog Dump|system user|event_scheduler|SHOW PROCESSLIST/ && $0 ~ /^[0-9]/ {print $1}' $tmpfile` do ocf_run $MYSQL $MYSQL_OPTIONS_REPL \ -e "KILL ${thread}" done else ocf_log info "Ignoring post-demote notification execpt for my own demotion." fi return $OCF_SUCCESS ;; 'post-demote') demote_host=`echo $OCF_RESKEY_CRM_meta_notify_demote_uname|tr -d " "` if [ $demote_host = ${NODENAME} ]; then ocf_log info "Ignoring post-demote notification for my own demotion." return $OCF_SUCCESS fi ocf_log info "post-demote notification for $demote_host." # The former master has just been gracefully demoted. unset_master ;; *) return $OCF_SUCCESS ;; esac } ####################################################################### case "$1" in meta-data) meta_data exit $OCF_SUCCESS;; usage|help) usage exit $OCF_SUCCESS;; esac mysql_common_validate rc=$? LSB_STATUS_STOPPED=3 if [ $rc -ne 0 ]; then case "$1" in stop) ;; monitor) mysql_common_status "info" if [ $? -eq $OCF_SUCCESS ]; then # if validatation fails and pid is active, always treat this as an error ocf_exit_reason "environment validation failed, active pid is in unknown state." exit $OCF_ERR_GENERIC fi # validation failed and pid is not active, it's safe to say this instance is inactive. exit $OCF_NOT_RUNNING;; status) exit $LSB_STATUS_STOPPED;; *) exit $rc;; esac fi # What kind of method was invoked? case "$1" in start) mysql_start;; stop) mysql_stop;; status) mysql_common_status err;; monitor) mysql_monitor;; promote) mysql_promote;; demote) mysql_demote;; notify) mysql_notify;; validate-all) exit $OCF_SUCCESS;; *) usage exit $OCF_ERR_UNIMPLEMENTED;; esac # vi:sw=4:ts=4:et: diff --git a/heartbeat/mysql-common.sh b/heartbeat/mysql-common.sh index 65db9bf85..d1b1ddb96 100755 --- a/heartbeat/mysql-common.sh +++ b/heartbeat/mysql-common.sh @@ -1,316 +1,322 @@ #!/bin/sh ####################################################################### # Use runuser if available for SELinux. if [ -x /sbin/runuser ]; then SU=runuser else SU=su fi # Attempt to detect a default binary OCF_RESKEY_binary_default=$(which mysqld_safe 2> /dev/null) if [ "$OCF_RESKEY_binary_default" = "" ]; then OCF_RESKEY_binary_default=$(which safe_mysqld 2> /dev/null) fi # Fill in some defaults if no values are specified HOSTOS=`uname` if [ "X${HOSTOS}" = "XOpenBSD" ];then if [ "$OCF_RESKEY_binary_default" = "" ]; then OCF_RESKEY_binary_default="/usr/local/bin/mysqld_safe" fi OCF_RESKEY_config_default="/etc/my.cnf" OCF_RESKEY_datadir_default="/var/mysql" OCF_RESKEY_user_default="_mysql" OCF_RESKEY_group_default="_mysql" OCF_RESKEY_log_default="/var/log/mysqld.log" OCF_RESKEY_pid_default="/var/mysql/mysqld.pid" OCF_RESKEY_socket_default="/var/run/mysql/mysql.sock" else if [ "$OCF_RESKEY_binary_default" = "" ]; then OCF_RESKEY_binary_default="/usr/bin/safe_mysqld" fi OCF_RESKEY_config_default="/etc/my.cnf" OCF_RESKEY_datadir_default="/var/lib/mysql" OCF_RESKEY_user_default="mysql" OCF_RESKEY_group_default="mysql" OCF_RESKEY_log_default="/var/log/mysqld.log" OCF_RESKEY_pid_default="/var/run/mysql/mysqld.pid" OCF_RESKEY_socket_default="/var/lib/mysql/mysql.sock" fi OCF_RESKEY_client_binary_default="mysql" OCF_RESKEY_test_user_default="root" OCF_RESKEY_test_table_default="mysql.user" OCF_RESKEY_test_passwd_default="" OCF_RESKEY_enable_creation_default=0 OCF_RESKEY_additional_parameters_default="" OCF_RESKEY_replication_user_default="root" OCF_RESKEY_replication_passwd_default="" OCF_RESKEY_replication_port_default="3306" +OCF_RESKEY_replication_master_ssl_ca_default="" +OCF_RESKEY_replication_master_ssl_cert_default="" +OCF_RESKEY_replication_master_ssl_key_default="" OCF_RESKEY_max_slave_lag_default="3600" OCF_RESKEY_evict_outdated_slaves_default="false" OCF_RESKEY_reader_attribute_default="readable" : ${OCF_RESKEY_binary=${OCF_RESKEY_binary_default}} MYSQL_BINDIR=`dirname ${OCF_RESKEY_binary}` : ${OCF_RESKEY_client_binary=${OCF_RESKEY_client_binary_default}} : ${OCF_RESKEY_config=${OCF_RESKEY_config_default}} : ${OCF_RESKEY_datadir=${OCF_RESKEY_datadir_default}} : ${OCF_RESKEY_user=${OCF_RESKEY_user_default}} : ${OCF_RESKEY_group=${OCF_RESKEY_group_default}} : ${OCF_RESKEY_log=${OCF_RESKEY_log_default}} : ${OCF_RESKEY_pid=${OCF_RESKEY_pid_default}} : ${OCF_RESKEY_socket=${OCF_RESKEY_socket_default}} : ${OCF_RESKEY_test_user=${OCF_RESKEY_test_user_default}} : ${OCF_RESKEY_test_table=${OCF_RESKEY_test_table_default}} : ${OCF_RESKEY_test_passwd=${OCF_RESKEY_test_passwd_default}} : ${OCF_RESKEY_enable_creation=${OCF_RESKEY_enable_creation_default}} : ${OCF_RESKEY_additional_parameters=${OCF_RESKEY_additional_parameters_default}} : ${OCF_RESKEY_replication_user=${OCF_RESKEY_replication_user_default}} : ${OCF_RESKEY_replication_passwd=${OCF_RESKEY_replication_passwd_default}} : ${OCF_RESKEY_replication_port=${OCF_RESKEY_replication_port_default}} +: ${OCF_RESKEY_replication_master_ssl_ca=${OCF_RESKEY_replication_master_ssl_ca_default}} +: ${OCF_RESKEY_replication_master_ssl_cert=${OCF_RESKEY_replication_master_ssl_cert_default}} +: ${OCF_RESKEY_replication_master_ssl_key=${OCF_RESKEY_replication_master_ssl_key_default}} : ${OCF_RESKEY_max_slave_lag=${OCF_RESKEY_max_slave_lag_default}} : ${OCF_RESKEY_evict_outdated_slaves=${OCF_RESKEY_evict_outdated_slaves_default}} : ${OCF_RESKEY_reader_attribute=${OCF_RESKEY_reader_attribute_default}} ####################################################################### # Convenience variables MYSQL=$OCF_RESKEY_client_binary MYSQL_OPTIONS_LOCAL="-S $OCF_RESKEY_socket" MYSQL_OPTIONS_REPL="$MYSQL_OPTIONS_LOCAL --user=$OCF_RESKEY_replication_user --password=$OCF_RESKEY_replication_passwd" MYSQL_OPTIONS_TEST="$MYSQL_OPTIONS_LOCAL --user=$OCF_RESKEY_test_user --password=$OCF_RESKEY_test_passwd" MYSQL_TOO_MANY_CONN_ERR=1040 CRM_MASTER="${HA_SBIN_DIR}/crm_master -l reboot " NODENAME=$(ocf_local_nodename) CRM_ATTR="${HA_SBIN_DIR}/crm_attribute -N $NODENAME " INSTANCE_ATTR_NAME=`echo ${OCF_RESOURCE_INSTANCE}| awk -F : '{print $1}'` CRM_ATTR_REPL_INFO="${HA_SBIN_DIR}/crm_attribute --type crm_config --name ${INSTANCE_ATTR_NAME}_REPL_INFO -s mysql_replication" ####################################################################### mysql_common_validate() { if ! have_binary "$OCF_RESKEY_binary"; then ocf_exit_reason "Setup problem: couldn't find command: $OCF_RESKEY_binary" return $OCF_ERR_INSTALLED; fi if ! have_binary "$OCF_RESKEY_client_binary"; then ocf_exit_reason "Setup problem: couldn't find command: $OCF_RESKEY_client_binary" return $OCF_ERR_INSTALLED; fi if [ ! -f $OCF_RESKEY_config ]; then ocf_exit_reason "Config $OCF_RESKEY_config doesn't exist"; return $OCF_ERR_INSTALLED; fi if [ ! -d $OCF_RESKEY_datadir ]; then ocf_exit_reason "Datadir $OCF_RESKEY_datadir doesn't exist"; return $OCF_ERR_INSTALLED; fi getent passwd $OCF_RESKEY_user >/dev/null 2>&1 if [ ! $? -eq 0 ]; then ocf_exit_reason "User $OCF_RESKEY_user doesn't exit"; return $OCF_ERR_INSTALLED; fi getent group $OCF_RESKEY_group >/dev/null 2>&1 if [ ! $? -eq 0 ]; then ocf_exit_reason "Group $OCF_RESKEY_group doesn't exist"; return $OCF_ERR_INSTALLED; fi return $OCF_SUCCESS } mysql_common_check_pid() { local pid=$1 if [ -d /proc -a -d /proc/1 ]; then [ "u$pid" != "u" -a -d /proc/$pid ] else kill -s 0 $pid >/dev/null 2>&1 fi return $? } mysql_common_status() { local loglevel=$1 local pid=$2 if [ -z "$pid" ]; then if [ ! -e $OCF_RESKEY_pid ]; then ocf_log $loglevel "MySQL is not running" return $OCF_NOT_RUNNING; fi pid=`cat $OCF_RESKEY_pid`; fi mysql_common_check_pid $pid if [ $? -eq 0 ]; then return $OCF_SUCCESS; else if [ -e $OCF_RESKEY_pid ]; then ocf_log $loglevel "MySQL not running: removing old PID file" rm -f $OCF_RESKEY_pid fi return $OCF_NOT_RUNNING; fi } mysql_common_prepare_dirs() { local rc touch $OCF_RESKEY_log chown $OCF_RESKEY_user:$OCF_RESKEY_group $OCF_RESKEY_log chmod 0640 $OCF_RESKEY_log [ -x /sbin/restorecon ] && /sbin/restorecon $OCF_RESKEY_log if ocf_is_true "$OCF_RESKEY_enable_creation" && [ ! -d $OCF_RESKEY_datadir/mysql ] ; then ocf_log info "Initializing MySQL database: " $MYSQL_BINDIR/mysql_install_db --datadir=$OCF_RESKEY_datadir rc=$? if [ $rc -ne 0 ] ; then ocf_exit_reason "Initialization failed: $rc"; exit $OCF_ERR_GENERIC fi chown -R $OCF_RESKEY_user:$OCF_RESKEY_group $OCF_RESKEY_datadir fi pid_dir=`dirname $OCF_RESKEY_pid` if [ ! -d $pid_dir ] ; then ocf_log info "Creating PID dir: $pid_dir" mkdir -p $pid_dir chown $OCF_RESKEY_user:$OCF_RESKEY_group $pid_dir fi socket_dir=`dirname $OCF_RESKEY_socket` if [ ! -d $socket_dir ] ; then ocf_log info "Creating socket dir: $socket_dir" mkdir -p $socket_dir chown $OCF_RESKEY_user:$OCF_RESKEY_group $socket_dir fi # Regardless of whether we just created the directory or it # already existed, check whether it is writable by the configured # user for dir in $pid_dir $socket_dir; do if ! $SU -s /bin/sh - $OCF_RESKEY_user -c "test -w $dir"; then ocf_exit_reason "Directory $dir is not writable by $OCF_RESKEY_user" exit $OCF_ERR_PERM; fi done } mysql_common_start() { local mysql_extra_params="$1" local pid $SU - $OCF_RESKEY_user -s /bin/sh -c \ "${OCF_RESKEY_binary} --defaults-file=$OCF_RESKEY_config \ --pid-file=$OCF_RESKEY_pid \ --socket=$OCF_RESKEY_socket \ --datadir=$OCF_RESKEY_datadir \ --log-error=$OCF_RESKEY_log \ $OCF_RESKEY_additional_parameters \ $mysql_extra_params >/dev/null 2>&1 & pid=$!" # Spin waiting for the server to come up. # Let the CRM/LRM time us out if required. start_wait=1 while [ $start_wait = 1 ]; do if ! ps $pid > /dev/null 2>&1; then wait $pid ocf_exit_reason "MySQL server failed to start (pid=$pid) (rc=$?), please check your installation" return $OCF_ERR_GENERIC fi mysql_common_status info rc=$? if [ $rc = $OCF_SUCCESS ]; then start_wait=0 elif [ $rc != $OCF_NOT_RUNNING ]; then ocf_log info "MySQL start failed: $rc" return $rc fi sleep 2 done return $OCF_SUCCESS } mysql_common_stop() { local pid local rc if [ ! -f $OCF_RESKEY_pid ]; then ocf_log info "MySQL is not running" return $OCF_SUCCESS fi pid=`cat $OCF_RESKEY_pid 2> /dev/null ` mysql_common_check_pid $pid if [ $? -ne 0 ]; then rm -f $OCF_RESKEY_pid ocf_log info "MySQL is already stopped" return $OCF_SUCCESS; fi /bin/kill $pid > /dev/null rc=$? if [ $rc != 0 ]; then ocf_exit_reason "MySQL couldn't be stopped" return $OCF_ERR_GENERIC fi # stop waiting shutdown_timeout=15 if [ -n "$OCF_RESKEY_CRM_meta_timeout" ]; then shutdown_timeout=$((($OCF_RESKEY_CRM_meta_timeout/1000)-5)) fi count=0 while [ $count -lt $shutdown_timeout ] do mysql_common_status info $pid rc=$? if [ $rc = $OCF_NOT_RUNNING ]; then break fi count=`expr $count + 1` sleep 1 ocf_log debug "MySQL still hasn't stopped yet. Waiting..." done mysql_common_status info $pid if [ $? != $OCF_NOT_RUNNING ]; then ocf_log info "MySQL failed to stop after ${shutdown_timeout}s using SIGTERM. Trying SIGKILL..." /bin/kill -KILL $pid > /dev/null fi ocf_log info "MySQL stopped"; rm -f /var/lock/subsys/mysqld rm -f $OCF_RESKEY_socket return $OCF_SUCCESS }