charm-hacluster seems to not be updating records for redeployed pacemaker-remote peers
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack HA Cluster Charm |
New
|
Undecided
|
Unassigned |
Bug Description
This was discovered in a Masakari environment, where a hacluster-masakari app is related to a pacemaker-remote app via the pacemaker-remote interface.
Charm versions in use:
* hacluster-masakari: cs:~openstack-
* pacemaker-remote: cs:~openstack-
I see after looking this up that both charms in question were prerelease ones, so while I can't seem to find any existing bugs for this, it's possible that upgrading may resolve this. Nonetheless, please let me report this, just in case it is still an issue.
A couple of Nova compute nodes in this OpenStack environment were redeployed. However, "crm config show" on the Masakari units still shows those two nodes with out-of-date IP addresses. The nodes are up, but "crm status" shows errors since those IPs obviously don't have pacemaker instances listening - the IPs have changed.
Relation data clearly shows that hacluster has the up-to-date IP addresses available via relation data. It just seems that for some reason, hacluster is not updating the records, requiring manual intervention.
Best Regards,
Paul Goins