I recently set up a cluster between two small PVE nodes in my homelab to provide failover on critical VMs. I was doing this in a bit of a sandbox for testing, and today I learned that if the failover node isn't turned on when you restart the primary node, none of the VMs on the primary node will start up. Is there a way to configure this failover relationship between the two nodes as "nice-to-have" instead of a "need-to-have", so that if the failover node is unavailable for whatever reason, the primary PVE VMs will still start up, and then replication will resume if and when the failover node becomes available?
Possible to designate a cluster node as "optional"?
- Thread starter serendipitist
- Start date
-
- Tags
- cluster
Ideally you have at least 3 votes in a cluster so that there are still 2 present if one of the nodes is down. In some situations, a full 3rd node is just not justifiable. But for these situations the QDevice mechanism can be used https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_corosync_external_vote_support
Everything else is working around the intended use-case, with more or less pain points.
Everything else is working around the intended use-case, with more or less pain points.
pvecm extected X is something that might work as a quick solution, but again, not the intended use-case and therefore you might see edge-cases that might cause problems.