From dde12b075ff51d4def4f49e635dd390a7f1f2cac Mon Sep 17 00:00:00 2001 From: Michele Baldessari Date: Wed, 9 Nov 2016 09:05:08 +0100 Subject: Fix race during major-upgrade-pacemaker step Currently when we call the major-upgrade step we do the following: """ ... if [[ -n $(is_bootstrap_node) ]]; then check_clean_cluster fi ... if [[ -n $(is_bootstrap_node) ]]; then migrate_full_to_ng_ha fi ... for service in $(services_to_migrate); do manage_systemd_service stop "${service%%-clone}" ... done """ The problem with the above code is that it is open to the following race condition: 1. Code gets run first on a non-bootstrap controller node so we start stopping a bunch of services 2. Pacemaker notices will notice that services are down and will mark the service as stopped 3. Code gets run on the bootstrap node (controller-0) and the check_clean_cluster function will fail and exit 4. Eventually also the script on the non-bootstrap controller node will timeout and exit because the cluster never shut down (it never actually started the shutdown because we failed at 3) Let's make sure we first only call the HA NG migration step as a separate heat step. Only afterwards we start shutting down the systemd services on all nodes. We also need to move the STONITH_STATE variable into a file because it is being used across two different scripts (1 and 2) and we need to store that state. Co-Authored-By: Athlan-Guyot Sofer Closes-Bug: #1640407 Change-Id: Ifb9b9e633fcc77604cca2590071656f4b2275c60 --- extraconfig/tasks/major_upgrade_check.sh | 19 +- .../tasks/major_upgrade_controller_pacemaker_1.sh | 182 +---------------- .../tasks/major_upgrade_controller_pacemaker_2.sh | 221 ++++++++++++++++----- .../tasks/major_upgrade_controller_pacemaker_3.sh | 69 ++++++- .../tasks/major_upgrade_controller_pacemaker_4.sh | 17 +- .../tasks/major_upgrade_controller_pacemaker_5.sh | 13 +- .../tasks/major_upgrade_controller_pacemaker_6.sh | 15 ++ extraconfig/tasks/major_upgrade_pacemaker.yaml | 42 +++- 8 files changed, 315 insertions(+), 263 deletions(-) create mode 100755 extraconfig/tasks/major_upgrade_controller_pacemaker_6.sh diff --git a/extraconfig/tasks/major_upgrade_check.sh b/extraconfig/tasks/major_upgrade_check.sh index b65f6915..8bdff5e7 100755 --- a/extraconfig/tasks/major_upgrade_check.sh +++ b/extraconfig/tasks/major_upgrade_check.sh @@ -18,14 +18,8 @@ check_pcsd() fi } -check_disk_for_mysql_dump() +mysql_need_update() { - # Where to backup current database if mysql need to be upgraded - MYSQL_BACKUP_DIR=/var/tmp/mysql_upgrade_osp - MYSQL_TEMP_UPGRADE_BACKUP_DIR=/var/lib/mysql-temp-upgrade-backup - # Spare disk ratio for extra safety - MYSQL_BACKUP_SIZE_RATIO=1.2 - # Shall we upgrade mysql data directory during the stack upgrade? if [ "$mariadb_do_major_upgrade" = "auto" ]; then ret=$(is_mysql_upgrade_needed) @@ -40,6 +34,17 @@ check_disk_for_mysql_dump() else DO_MYSQL_UPGRADE=1 fi +} + +check_disk_for_mysql_dump() +{ + # Where to backup current database if mysql need to be upgraded + MYSQL_BACKUP_DIR=/var/tmp/mysql_upgrade_osp + MYSQL_TEMP_UPGRADE_BACKUP_DIR=/var/lib/mysql-temp-upgrade-backup + # Spare disk ratio for extra safety + MYSQL_BACKUP_SIZE_RATIO=1.2 + + mysql_need_update if [ "$(hiera -c /etc/puppet/hiera.yaml bootstrap_nodeid)" = "$(facter hostname)" ]; then if [ $DO_MYSQL_UPGRADE -eq 1 ]; then diff --git a/extraconfig/tasks/major_upgrade_controller_pacemaker_1.sh b/extraconfig/tasks/major_upgrade_controller_pacemaker_1.sh index fbdbc30b..739c01dd 100755 --- a/extraconfig/tasks/major_upgrade_controller_pacemaker_1.sh +++ b/extraconfig/tasks/major_upgrade_controller_pacemaker_1.sh @@ -2,8 +2,6 @@ set -eu -cluster_sync_timeout=1800 - check_cluster check_pcsd if [[ -n $(is_bootstrap_node) ]]; then @@ -19,6 +17,11 @@ check_disk_for_mysql_dump # at the end of this script if [[ -n $(is_bootstrap_node) ]]; then STONITH_STATE=$(pcs property show stonith-enabled | grep "stonith-enabled" | awk '{ print $2 }') + # We create this empty file if stonith was set to true so we can reenable stonith in step2 + rm -f /var/tmp/stonith-true + if [ $STONITH_STATE == "true" ]; then + touch /var/tmp/stonith-true + fi pcs property set stonith-enabled=false fi @@ -31,178 +34,3 @@ if [[ -n $(is_bootstrap_node) ]]; then rabbitmq_mitaka_newton_upgrade fi -# After migrating the cluster to HA-NG the services not under pacemaker's control -# are still up and running. We need to stop them explicitely otherwise during the yum -# upgrade the rpm %post sections will try to do a systemctl try-restart , which -# is going to take a long time because rabbit is down. By having the service stopped -# systemctl try-restart is a noop - -for service in $(services_to_migrate); do - manage_systemd_service stop "${service%%-clone}" - # So the reason for not reusing check_resource_systemd is that - # I have observed systemctl is-active returning unknown with at least - # one service that was stopped (See LP 1627254) - timeout=600 - tstart=$(date +%s) - tend=$(( $tstart + $timeout )) - check_interval=3 - while (( $(date +%s) < $tend )); do - if [[ "$(systemctl is-active ${service%%-clone})" = "active" ]]; then - echo "$service still active, sleeping $check_interval seconds." - sleep $check_interval - else - # we do not care if it is inactive, unknown or failed as long as it is - # not running - break - fi - - done -done - -# In case the mysql package is updated, the database on disk must be -# upgraded as well. This typically needs to happen during major -# version upgrades (e.g. 5.5 -> 5.6, 5.5 -> 10.1...) -# -# Because in-place upgrades are not supported across 2+ major versions -# (e.g. 5.5 -> 10.1), we rely on logical upgrades via dump/restore cycle -# https://bugzilla.redhat.com/show_bug.cgi?id=1341968 -# -# The default is to determine automatically if upgrade is needed based -# on mysql package versionning, but this can be overriden manually -# to support specific upgrade scenario - -if [[ -n $(is_bootstrap_node) ]]; then - if [ $DO_MYSQL_UPGRADE -eq 1 ]; then - mysqldump $backup_flags > "$MYSQL_BACKUP_DIR/openstack_database.sql" - cp -rdp /etc/my.cnf* "$MYSQL_BACKUP_DIR" - fi - - pcs resource disable redis - check_resource redis stopped 600 - pcs resource disable rabbitmq - check_resource rabbitmq stopped 600 - pcs resource disable galera - check_resource galera stopped 600 - pcs resource disable openstack-cinder-volume - check_resource openstack-cinder-volume stopped 600 - # Disable all VIPs before stopping the cluster, so that pcs doesn't use one as a source address: - # https://bugzilla.redhat.com/show_bug.cgi?id=1330688 - for vip in $(pcs resource show | grep ocf::heartbeat:IPaddr2 | grep Started | awk '{ print $1 }'); do - pcs resource disable $vip - check_resource $vip stopped 60 - done - pcs cluster stop --all -fi - - -# Swift isn't controlled by pacemaker -systemctl_swift stop - -tstart=$(date +%s) -while systemctl is-active pacemaker; do - sleep 5 - tnow=$(date +%s) - if (( tnow-tstart > cluster_sync_timeout )) ; then - echo_error "ERROR: cluster shutdown timed out" - exit 1 - fi -done - -# The reason we do an sql dump *and* we move the old dir out of -# the way is because it gives us an extra level of safety in case -# something goes wrong during the upgrade. Once the restore is -# successful we go ahead and remove it. If the directory exists -# we bail out as it means the upgrade process had issues in the last -# run. -if [ $DO_MYSQL_UPGRADE -eq 1 ]; then - if [ -d $MYSQL_TEMP_UPGRADE_BACKUP_DIR ]; then - echo_error "ERROR: mysql backup dir already exist" - exit 1 - fi - mv /var/lib/mysql $MYSQL_TEMP_UPGRADE_BACKUP_DIR -fi - -# Special-case OVS for https://bugs.launchpad.net/tripleo/+bug/1635205 -if [[ -n $(rpm -q --scripts openvswitch | awk '/postuninstall/,/*/' | grep "systemctl.*try-restart") ]]; then - echo "Manual upgrade of openvswitch - restart in postun detected" - mkdir OVS_UPGRADE || true - pushd OVS_UPGRADE - echo "Attempting to downloading latest openvswitch with yumdownloader" - yumdownloader --resolve openvswitch - echo "Updating openvswitch with nopostun option" - rpm -U --replacepkgs --nopostun ./*.rpm - popd -else - echo "Skipping manual upgrade of openvswitch - no restart in postun detected" -fi - -yum -y install python-zaqarclient # needed for os-collect-config -yum -y -q update - -# We need to ensure at least those two configuration settings, otherwise -# mariadb 10.1+ won't activate galera replication. -# wsrep_cluster_address must only be set though, its value does not -# matter because it's overriden by the galera resource agent. -cat >> /etc/my.cnf.d/galera.cnf < /dev/null; do sleep 1; done' - mysql -u root < "$MYSQL_BACKUP_DIR/openstack_database.sql" - mv /root/.my.cnf.temporary /root/.my.cnf - mysqladmin -u root shutdown - # The import was successful so we may remove the folder - rm -r "$MYSQL_BACKUP_DIR" - fi -fi - -# If we reached here without error we can safely blow away the origin -# mysql dir from every controller - -# TODO: What if the upgrade fails on the bootstrap node, but not on -# this controller. Data may be lost. -if [ $DO_MYSQL_UPGRADE -eq 1 ]; then - rm -r $MYSQL_TEMP_UPGRADE_BACKUP_DIR -fi - -# Let's reset the stonith back to true if it was true, before starting the cluster -if [[ -n $(is_bootstrap_node) ]]; then - if [ $STONITH_STATE == "true" ]; then - pcs -f /var/lib/pacemaker/cib/cib.xml property set stonith-enabled=true - fi -fi - -# Pin messages sent to compute nodes to kilo, these will be upgraded later -crudini --set /etc/nova/nova.conf upgrade_levels compute "$upgrade_level_nova_compute" -# https://bugzilla.redhat.com/show_bug.cgi?id=1284047 -# Change-Id: Ib3f6c12ff5471e1f017f28b16b1e6496a4a4b435 -crudini --set /etc/ceilometer/ceilometer.conf DEFAULT rpc_backend rabbit -# https://bugzilla.redhat.com/show_bug.cgi?id=1284058 -# Ifd1861e3df46fad0e44ff9b5cbd58711bbc87c97 Swift Ceilometer middleware no longer exists -crudini --set /etc/swift/proxy-server.conf pipeline:main pipeline "catch_errors healthcheck cache ratelimit tempurl formpost authtoken keystone staticweb proxy-logging proxy-server" -# LP: 1615035, required only for M/N upgrade. -crudini --set /etc/nova/nova.conf DEFAULT scheduler_host_manager host_manager -# LP: 1627450, required only for M/N upgrade -crudini --set /etc/nova/nova.conf DEFAULT scheduler_driver filter_scheduler - -crudini --set /etc/sahara/sahara.conf DEFAULT plugins ambari,cdh,mapr,vanilla,spark,storm diff --git a/extraconfig/tasks/major_upgrade_controller_pacemaker_2.sh b/extraconfig/tasks/major_upgrade_controller_pacemaker_2.sh index 37061512..7cc6735f 100755 --- a/extraconfig/tasks/major_upgrade_controller_pacemaker_2.sh +++ b/extraconfig/tasks/major_upgrade_controller_pacemaker_2.sh @@ -2,67 +2,186 @@ set -eu -cluster_form_timeout=600 -cluster_settle_timeout=1800 -galera_sync_timeout=600 +cluster_sync_timeout=1800 -if [[ -n $(is_bootstrap_node) ]]; then - pcs cluster start --all +# After migrating the cluster to HA-NG the services not under pacemaker's control +# are still up and running. We need to stop them explicitely otherwise during the yum +# upgrade the rpm %post sections will try to do a systemctl try-restart , which +# is going to take a long time because rabbit is down. By having the service stopped +# systemctl try-restart is a noop +for service in $(services_to_migrate); do + manage_systemd_service stop "${service%%-clone}" + # So the reason for not reusing check_resource_systemd is that + # I have observed systemctl is-active returning unknown with at least + # one service that was stopped (See LP 1627254) + timeout=600 tstart=$(date +%s) - while pcs status 2>&1 | grep -E '(cluster is not currently running)|(OFFLINE:)'; do - sleep 5 - tnow=$(date +%s) - if (( tnow-tstart > cluster_form_timeout )) ; then - echo_error "ERROR: timed out forming the cluster" - exit 1 - fi + tend=$(( $tstart + $timeout )) + check_interval=3 + while (( $(date +%s) < $tend )); do + if [[ "$(systemctl is-active ${service%%-clone})" = "active" ]]; then + echo "$service still active, sleeping $check_interval seconds." + sleep $check_interval + else + # we do not care if it is inactive, unknown or failed as long as it is + # not running + break + fi + done +done - if ! timeout -k 10 $cluster_settle_timeout crm_resource --wait; then - echo_error "ERROR: timed out waiting for cluster to finish transition" - exit 1 +# In case the mysql package is updated, the database on disk must be +# upgraded as well. This typically needs to happen during major +# version upgrades (e.g. 5.5 -> 5.6, 5.5 -> 10.1...) +# +# Because in-place upgrades are not supported across 2+ major versions +# (e.g. 5.5 -> 10.1), we rely on logical upgrades via dump/restore cycle +# https://bugzilla.redhat.com/show_bug.cgi?id=1341968 +# +# The default is to determine automatically if upgrade is needed based +# on mysql package versionning, but this can be overriden manually +# to support specific upgrade scenario + +# Calling this function will set the DO_MYSQL_UPGRADE variable which is used +# later +mysql_need_update + +if [[ -n $(is_bootstrap_node) ]]; then + if [ $DO_MYSQL_UPGRADE -eq 1 ]; then + mysqldump $backup_flags > "$MYSQL_BACKUP_DIR/openstack_database.sql" + cp -rdp /etc/my.cnf* "$MYSQL_BACKUP_DIR" fi - for vip in $(pcs resource show | grep ocf::heartbeat:IPaddr2 | grep Stopped | awk '{ print $1 }'); do - pcs resource enable $vip - check_resource_pacemaker $vip started 60 + pcs resource disable redis + check_resource redis stopped 600 + pcs resource disable rabbitmq + check_resource rabbitmq stopped 600 + pcs resource disable galera + check_resource galera stopped 600 + pcs resource disable openstack-cinder-volume + check_resource openstack-cinder-volume stopped 600 + # Disable all VIPs before stopping the cluster, so that pcs doesn't use one as a source address: + # https://bugzilla.redhat.com/show_bug.cgi?id=1330688 + for vip in $(pcs resource show | grep ocf::heartbeat:IPaddr2 | grep Started | awk '{ print $1 }'); do + pcs resource disable $vip + check_resource $vip stopped 60 done + pcs cluster stop --all fi -start_or_enable_service galera -check_resource galera started 600 -start_or_enable_service redis -check_resource redis started 600 -# We need mongod which is now a systemd service up and running before calling -# ceilometer-dbsync. There is still a race here: mongod might not be up on all nodes -# so ceilometer-dbsync will fail a couple of times before that. As it retries indefinitely -# we should be good. -# Due to LP Bug https://bugs.launchpad.net/tripleo/+bug/1627254 am using systemctl directly atm -systemctl start mongod -check_resource mongod started 600 -if [[ -n $(is_bootstrap_node) ]]; then - tstart=$(date +%s) - while ! clustercheck; do - sleep 5 - tnow=$(date +%s) - if (( tnow-tstart > galera_sync_timeout )) ; then - echo_error "ERROR galera sync timed out" - exit 1 - fi - done +# Swift isn't controlled by pacemaker +systemctl_swift stop - # Run all the db syncs - # TODO: check if this can be triggered in puppet and removed from here - ceilometer-dbsync --config-file=/etc/ceilometer/ceilometer.conf - cinder-manage db sync - glance-manage --config-file=/etc/glance/glance-registry.conf db_sync - heat-manage --config-file /etc/heat/heat.conf db_sync - keystone-manage db_sync - neutron-db-manage --config-file /etc/neutron/neutron.conf --config-file /etc/neutron/plugin.ini upgrade head - nova-manage db sync - nova-manage api_db sync - nova-manage db online_data_migrations - sahara-db-manage --config-file /etc/sahara/sahara.conf upgrade head +tstart=$(date +%s) +while systemctl is-active pacemaker; do + sleep 5 + tnow=$(date +%s) + if (( tnow-tstart > cluster_sync_timeout )) ; then + echo_error "ERROR: cluster shutdown timed out" + exit 1 + fi +done + +# The reason we do an sql dump *and* we move the old dir out of +# the way is because it gives us an extra level of safety in case +# something goes wrong during the upgrade. Once the restore is +# successful we go ahead and remove it. If the directory exists +# we bail out as it means the upgrade process had issues in the last +# run. +if [ $DO_MYSQL_UPGRADE -eq 1 ]; then + if [ -d $MYSQL_TEMP_UPGRADE_BACKUP_DIR ]; then + echo_error "ERROR: mysql backup dir already exist" + exit 1 + fi + mv /var/lib/mysql $MYSQL_TEMP_UPGRADE_BACKUP_DIR +fi + +# Special-case OVS for https://bugs.launchpad.net/tripleo/+bug/1635205 +if [[ -n $(rpm -q --scripts openvswitch | awk '/postuninstall/,/*/' | grep "systemctl.*try-restart") ]]; then + echo "Manual upgrade of openvswitch - restart in postun detected" + mkdir OVS_UPGRADE || true + pushd OVS_UPGRADE + echo "Attempting to downloading latest openvswitch with yumdownloader" + yumdownloader --resolve openvswitch + echo "Updating openvswitch with nopostun option" + rpm -U --replacepkgs --nopostun ./*.rpm + popd +else + echo "Skipping manual upgrade of openvswitch - no restart in postun detected" fi + +yum -y install python-zaqarclient # needed for os-collect-config +yum -y -q update + +# We need to ensure at least those two configuration settings, otherwise +# mariadb 10.1+ won't activate galera replication. +# wsrep_cluster_address must only be set though, its value does not +# matter because it's overriden by the galera resource agent. +cat >> /etc/my.cnf.d/galera.cnf < /dev/null; do sleep 1; done' + mysql -u root < "$MYSQL_BACKUP_DIR/openstack_database.sql" + mv /root/.my.cnf.temporary /root/.my.cnf + mysqladmin -u root shutdown + # The import was successful so we may remove the folder + rm -r "$MYSQL_BACKUP_DIR" + fi +fi + +# If we reached here without error we can safely blow away the origin +# mysql dir from every controller + +# TODO: What if the upgrade fails on the bootstrap node, but not on +# this controller. Data may be lost. +if [ $DO_MYSQL_UPGRADE -eq 1 ]; then + rm -r $MYSQL_TEMP_UPGRADE_BACKUP_DIR +fi + +# Let's reset the stonith back to true if it was true, before starting the cluster +if [[ -n $(is_bootstrap_node) ]]; then + if [ -f /var/tmp/stonith-true ]; then + pcs -f /var/lib/pacemaker/cib/cib.xml property set stonith-enabled=true + fi + rm -f /var/tmp/stonith-true +fi + +# Pin messages sent to compute nodes to kilo, these will be upgraded later +crudini --set /etc/nova/nova.conf upgrade_levels compute "$upgrade_level_nova_compute" +# https://bugzilla.redhat.com/show_bug.cgi?id=1284047 +# Change-Id: Ib3f6c12ff5471e1f017f28b16b1e6496a4a4b435 +crudini --set /etc/ceilometer/ceilometer.conf DEFAULT rpc_backend rabbit +# https://bugzilla.redhat.com/show_bug.cgi?id=1284058 +# Ifd1861e3df46fad0e44ff9b5cbd58711bbc87c97 Swift Ceilometer middleware no longer exists +crudini --set /etc/swift/proxy-server.conf pipeline:main pipeline "catch_errors healthcheck cache ratelimit tempurl formpost authtoken keystone staticweb proxy-logging proxy-server" +# LP: 1615035, required only for M/N upgrade. +crudini --set /etc/nova/nova.conf DEFAULT scheduler_host_manager host_manager +# LP: 1627450, required only for M/N upgrade +crudini --set /etc/nova/nova.conf DEFAULT scheduler_driver filter_scheduler + +crudini --set /etc/sahara/sahara.conf DEFAULT plugins ambari,cdh,mapr,vanilla,spark,storm + diff --git a/extraconfig/tasks/major_upgrade_controller_pacemaker_3.sh b/extraconfig/tasks/major_upgrade_controller_pacemaker_3.sh index d2cb9553..37061512 100755 --- a/extraconfig/tasks/major_upgrade_controller_pacemaker_3.sh +++ b/extraconfig/tasks/major_upgrade_controller_pacemaker_3.sh @@ -2,16 +2,67 @@ set -eu -start_or_enable_service rabbitmq -check_resource rabbitmq started 600 +cluster_form_timeout=600 +cluster_settle_timeout=1800 +galera_sync_timeout=600 + +if [[ -n $(is_bootstrap_node) ]]; then + pcs cluster start --all + + tstart=$(date +%s) + while pcs status 2>&1 | grep -E '(cluster is not currently running)|(OFFLINE:)'; do + sleep 5 + tnow=$(date +%s) + if (( tnow-tstart > cluster_form_timeout )) ; then + echo_error "ERROR: timed out forming the cluster" + exit 1 + fi + done + + if ! timeout -k 10 $cluster_settle_timeout crm_resource --wait; then + echo_error "ERROR: timed out waiting for cluster to finish transition" + exit 1 + fi + + for vip in $(pcs resource show | grep ocf::heartbeat:IPaddr2 | grep Stopped | awk '{ print $1 }'); do + pcs resource enable $vip + check_resource_pacemaker $vip started 60 + done +fi + +start_or_enable_service galera +check_resource galera started 600 start_or_enable_service redis check_resource redis started 600 -start_or_enable_service openstack-cinder-volume -check_resource openstack-cinder-volume started 600 +# We need mongod which is now a systemd service up and running before calling +# ceilometer-dbsync. There is still a race here: mongod might not be up on all nodes +# so ceilometer-dbsync will fail a couple of times before that. As it retries indefinitely +# we should be good. +# Due to LP Bug https://bugs.launchpad.net/tripleo/+bug/1627254 am using systemctl directly atm +systemctl start mongod +check_resource mongod started 600 -# start httpd so keystone is available for gnocchi -# upgrade to run. -systemctl start httpd +if [[ -n $(is_bootstrap_node) ]]; then + tstart=$(date +%s) + while ! clustercheck; do + sleep 5 + tnow=$(date +%s) + if (( tnow-tstart > galera_sync_timeout )) ; then + echo_error "ERROR galera sync timed out" + exit 1 + fi + done -# Swift isn't controled by pacemaker -systemctl_swift start + # Run all the db syncs + # TODO: check if this can be triggered in puppet and removed from here + ceilometer-dbsync --config-file=/etc/ceilometer/ceilometer.conf + cinder-manage db sync + glance-manage --config-file=/etc/glance/glance-registry.conf db_sync + heat-manage --config-file /etc/heat/heat.conf db_sync + keystone-manage db_sync + neutron-db-manage --config-file /etc/neutron/neutron.conf --config-file /etc/neutron/plugin.ini upgrade head + nova-manage db sync + nova-manage api_db sync + nova-manage db online_data_migrations + sahara-db-manage --config-file /etc/sahara/sahara.conf upgrade head +fi diff --git a/extraconfig/tasks/major_upgrade_controller_pacemaker_4.sh b/extraconfig/tasks/major_upgrade_controller_pacemaker_4.sh index fa95f1f8..d2cb9553 100755 --- a/extraconfig/tasks/major_upgrade_controller_pacemaker_4.sh +++ b/extraconfig/tasks/major_upgrade_controller_pacemaker_4.sh @@ -2,7 +2,16 @@ set -eu -if [[ -n $(is_bootstrap_node) ]]; then - # run gnocchi upgrade - gnocchi-upgrade -fi +start_or_enable_service rabbitmq +check_resource rabbitmq started 600 +start_or_enable_service redis +check_resource redis started 600 +start_or_enable_service openstack-cinder-volume +check_resource openstack-cinder-volume started 600 + +# start httpd so keystone is available for gnocchi +# upgrade to run. +systemctl start httpd + +# Swift isn't controled by pacemaker +systemctl_swift start diff --git a/extraconfig/tasks/major_upgrade_controller_pacemaker_5.sh b/extraconfig/tasks/major_upgrade_controller_pacemaker_5.sh index d569084d..fa95f1f8 100755 --- a/extraconfig/tasks/major_upgrade_controller_pacemaker_5.sh +++ b/extraconfig/tasks/major_upgrade_controller_pacemaker_5.sh @@ -2,14 +2,7 @@ set -eu -# We need to start the systemd services we explicitely stopped at step _1.sh -# FIXME: Should we let puppet during the convergence step do the service enabling or -# should we add it here? -services=$(services_to_migrate) -if [[ ${keep_sahara_services_on_upgrade} =~ [Ff]alse ]] ; then - services=${services%%openstack-sahara*} +if [[ -n $(is_bootstrap_node) ]]; then + # run gnocchi upgrade + gnocchi-upgrade fi -for service in $services; do - manage_systemd_service start "${service%%-clone}" - check_resource_systemd "${service%%-clone}" started 600 -done diff --git a/extraconfig/tasks/major_upgrade_controller_pacemaker_6.sh b/extraconfig/tasks/major_upgrade_controller_pacemaker_6.sh new file mode 100755 index 00000000..d569084d --- /dev/null +++ b/extraconfig/tasks/major_upgrade_controller_pacemaker_6.sh @@ -0,0 +1,15 @@ +#!/bin/bash + +set -eu + +# We need to start the systemd services we explicitely stopped at step _1.sh +# FIXME: Should we let puppet during the convergence step do the service enabling or +# should we add it here? +services=$(services_to_migrate) +if [[ ${keep_sahara_services_on_upgrade} =~ [Ff]alse ]] ; then + services=${services%%openstack-sahara*} +fi +for service in $services; do + manage_systemd_service start "${service%%-clone}" + check_resource_systemd "${service%%-clone}" started 600 +done diff --git a/extraconfig/tasks/major_upgrade_pacemaker.yaml b/extraconfig/tasks/major_upgrade_pacemaker.yaml index e13aada3..b0418a56 100644 --- a/extraconfig/tasks/major_upgrade_pacemaker.yaml +++ b/extraconfig/tasks/major_upgrade_pacemaker.yaml @@ -113,7 +113,20 @@ resources: config: list_join: - '' - - - get_file: pacemaker_common_functions.sh + - - str_replace: + template: | + #!/bin/bash + upgrade_level_nova_compute='UPGRADE_LEVEL_NOVA_COMPUTE' + params: + UPGRADE_LEVEL_NOVA_COMPUTE: {get_param: UpgradeLevelNovaCompute} + - str_replace: + template: | + #!/bin/bash + mariadb_do_major_upgrade='MYSQL_MAJOR_UPGRADE' + params: + MYSQL_MAJOR_UPGRADE: {get_param: MySqlMajorUpgrade} + - get_file: pacemaker_common_functions.sh + - get_file: major_upgrade_check.sh - get_file: major_upgrade_pacemaker_migrations.sh - get_file: major_upgrade_controller_pacemaker_2.sh @@ -164,6 +177,25 @@ resources: input_values: {get_param: input_values} ControllerPacemakerUpgradeConfig_Step5: + type: OS::Heat::SoftwareConfig + properties: + group: script + config: + list_join: + - '' + - - get_file: pacemaker_common_functions.sh + - get_file: major_upgrade_pacemaker_migrations.sh + - get_file: major_upgrade_controller_pacemaker_5.sh + + ControllerPacemakerUpgradeDeployment_Step5: + type: OS::Heat::SoftwareDeploymentGroup + depends_on: ControllerPacemakerUpgradeDeployment_Step4 + properties: + servers: {get_param: [servers, Controller]} + config: {get_resource: ControllerPacemakerUpgradeConfig_Step5} + input_values: {get_param: input_values} + + ControllerPacemakerUpgradeConfig_Step6: type: OS::Heat::SoftwareConfig properties: group: script @@ -178,12 +210,12 @@ resources: KEEP_SAHARA_SERVICES_ON_UPGRADE: {get_param: KeepSaharaServicesOnUpgrade} - get_file: pacemaker_common_functions.sh - get_file: major_upgrade_pacemaker_migrations.sh - - get_file: major_upgrade_controller_pacemaker_5.sh + - get_file: major_upgrade_controller_pacemaker_6.sh - ControllerPacemakerUpgradeDeployment_Step5: + ControllerPacemakerUpgradeDeployment_Step6: type: OS::Heat::SoftwareDeploymentGroup - depends_on: ControllerPacemakerUpgradeDeployment_Step4 + depends_on: ControllerPacemakerUpgradeDeployment_Step5 properties: servers: {get_param: [servers, Controller]} - config: {get_resource: ControllerPacemakerUpgradeConfig_Step5} + config: {get_resource: ControllerPacemakerUpgradeConfig_Step6} input_values: {get_param: input_values} -- cgit 1.2.3-korg