Automatic cleanup after node removal

[Michael Quiniola]

Environment:
-MAAS
-Juju
-Charmed Openstack (Focal/Ussuri)

In relation to Case #375039

After removal of a nova-compute node, Masakari does no clean up pacemaker nodes that are offline. They remain as "ocf::pacemaker:remote): Stopped"

Manual cleanup methods are inconsistent, but eventually work:

sudo crm_node -R <hostname> --force
sudo cibadmin --delete --xml-text '<node uname="<hostname>">'
sudo cibadmin --delete --xml-text '<node_state uname="<hostname>">'
sudo crm_resource -C

If the above commands don't work, then:

crm node standby <hostname>
crm resource unmanage <hostname>
crm configure delete <hostname>
crm resource cleanup <hostname>

stonith_admin --cleanup --history=<hostname>

Masakari should clean up its configurations automatically once a node has been removed. I am unsure whether this should be based on the nova database or based on juju information.

[Alex Kavanagh (ajkavanagh)]

Thanks for the bug report; however, I'm not sure I understand it. Masakari is an HA for instances running on OpenStack. When removing a nova-compute node, are you talking about the masakari-monitors (pacemaker remote) issues, as these were running on the compute-node host? (i.e. managed by the masakari-monitors charm).

[Launchpad Janitor (janitor)]

[Expired for OpenStack Masakari Charm because there has been no activity for 60 days.]