pacemaker: no failover of vm in case of resource failed on remote-node (with our configuration)

Solution Unverified - Updated 2024-08-02T05:45:54+00:00 -

Issue

On each resource running on remote-nodes, we have op mon on-failed=fence so that in case of monitoring failed the vm is fenced and restarted. When a resource on remote-node fails, the vm is always restarted locally on the same node and never migrated to a second node.
When a resource running in a remote node fails a monitor operation, and that resource has op monitor on-fail=fence, that fence of the remote-node does not reflect in the failcount of the container VM resource.

Red Hat Enterprise Linux (RHEL) 6 or 7 with the High Availability Add-On
pacemaker
One or more ocf:heartbeat:VirtualDomain resources with meta attribute remote-node set
One or more resources configured to run on a remote-node, also having op monitor on-fail=fence set.
migration-threshold > 0 on the VirtualDomain resource(s)

A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.