Description of problem: Currently we can only set galera max connections via heat templates [1]. So we have to manual set file limits by adding --pids-limits $ to pacemaker cib after the fact. This requires manual steps [2] . We would like to be able to set the file limit by template to be easier and safer.
[1]
[root@controller-0 heat-admin]# grep -r mysql_max /etc/puppet/hieradata/
/etc/puppet/hieradata/service_configs.json: "mysql_max_connections": 4096,
[root@undercloud-0 stack]# grep -R MysqlMaxConnections /usr/share/openstack-tripleo-heat-templates/
/usr/share/openstack-tripleo-heat-templates/deployment/database/mysql-base.yaml: MysqlMaxConnections:
/usr/share/openstack-tripleo-heat-templates/deployment/database/mysql-base.yaml: mysql_max_connections: {get_param: MysqlMaxConnections}
[2]
Commands to be run:
# Do this on any cluster node
# Unmanaged the resource
pcs resource unmanage galera-bundle
# Dump the original CIB
pcs cluster cib > cib-galera.xml
cp cib-galera.xml cib-galera.xml.org
# Edit cib-galera.xml and update on the resource so that the galera-bundle line looks like the following (choose the limit you want):
#
# - <podman image="cluster.common.tag/rhosp16-openstack-mariadb:pcmklatest" network="host" options="--user=root --log-driver=k8s-file --log-opt path=/var/log/containers/stdouts/galera-bundle.log -e KOLLA_CONFIG_STRATEGY=COPY_ALWAYS" promoted-max="3" replicas="3" run-command="/bin/bash /usr/local/bin/kolla_start"/>
# + <podman image="cluster.common.tag/rhosp16-openstack-mariadb:pcmklatest" network="host" options="--user=root --pids-limit 6666 --log-driver=k8s-file --log-opt path=/var/log/containers/stdouts/galera-bundle.log -e KOLLA_CONFIG_STRATEGY=COPY_ALWAYS" promoted-max="3" replicas="3" run-command="/bin/bash /usr/local/bin/kolla_start"/>
# Push the updated CIB. At this stage, no galera resource will restart
pcs cluster cib-push --config cib-galera.xml
# Force pacemaker to reprobe the state of the galera resource (that way it forgets that it has to restart it once it becomes managed again)
pcs resource refresh galera-bundle
# Give back control of the resource to pacemaker
pcs resource manage galera-bundle
# restart each resource individually for a rolling restart and the new value to take effect
Version-Release number of selected component (if applicable):
16.1
How reproducible:
100%
Steps to Reproduce:
1.above in [2]
2.
3.
Actual results:
manual edits required
Expected results:
THT to set value
Additional info:
Description of problem: Currently we can only set galera max connections via heat templates [1]. So we have to manual set file limits by adding --pids-limits $ to pacemaker cib after the fact. This requires manual steps [2] . We would like to be able to set the file limit by template to be easier and safer. [1] [root@controller-0 heat-admin]# grep -r mysql_max /etc/puppet/hieradata/ /etc/puppet/hieradata/service_configs.json: "mysql_max_connections": 4096, [root@undercloud-0 stack]# grep -R MysqlMaxConnections /usr/share/openstack-tripleo-heat-templates/ /usr/share/openstack-tripleo-heat-templates/deployment/database/mysql-base.yaml: MysqlMaxConnections: /usr/share/openstack-tripleo-heat-templates/deployment/database/mysql-base.yaml: mysql_max_connections: {get_param: MysqlMaxConnections} [2] Commands to be run: # Do this on any cluster node # Unmanaged the resource pcs resource unmanage galera-bundle # Dump the original CIB pcs cluster cib > cib-galera.xml cp cib-galera.xml cib-galera.xml.org # Edit cib-galera.xml and update on the resource so that the galera-bundle line looks like the following (choose the limit you want): # # - <podman image="cluster.common.tag/rhosp16-openstack-mariadb:pcmklatest" network="host" options="--user=root --log-driver=k8s-file --log-opt path=/var/log/containers/stdouts/galera-bundle.log -e KOLLA_CONFIG_STRATEGY=COPY_ALWAYS" promoted-max="3" replicas="3" run-command="/bin/bash /usr/local/bin/kolla_start"/> # + <podman image="cluster.common.tag/rhosp16-openstack-mariadb:pcmklatest" network="host" options="--user=root --pids-limit 6666 --log-driver=k8s-file --log-opt path=/var/log/containers/stdouts/galera-bundle.log -e KOLLA_CONFIG_STRATEGY=COPY_ALWAYS" promoted-max="3" replicas="3" run-command="/bin/bash /usr/local/bin/kolla_start"/> # Push the updated CIB. At this stage, no galera resource will restart pcs cluster cib-push --config cib-galera.xml # Force pacemaker to reprobe the state of the galera resource (that way it forgets that it has to restart it once it becomes managed again) pcs resource refresh galera-bundle # Give back control of the resource to pacemaker pcs resource manage galera-bundle # restart each resource individually for a rolling restart and the new value to take effect Version-Release number of selected component (if applicable): 16.1 How reproducible: 100% Steps to Reproduce: 1.above in [2] 2. 3. Actual results: manual edits required Expected results: THT to set value Additional info: