I'm testing a HCI setup of proxmox with ceph and a fully routed setup. I have a 3 node cluster and 1 VM I'm testing with. I was under the impression that if a host failed and the VM's running on them would stay running, and just move to another host. Is this not the case? Or do you think I'm missing something? I have HA enabled as well for the VM, and maybe that is not correct?
Hyperconvergence Setup
- Thread starter rbjohnson78
- Start date
I don't know where you got the impression but I don't think it works that way. This has been discussed before on the forum and there is a HA simulator mentioned in the manual to try stuff out: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_ha_simulator
If a host fails, it might be in any kind of states (for example a power wire failed and it turned off instantly). Once the quorum decides the node is not working properly (after a minute or so without being able to reach it), the VMs are restarted on another node. The node itself reboots when it detects that it's not part of the quorum (like not able to reach enough nodes, if it is still running, to prevent split-brain).
How could they? The host is failing. Now! The processes constituting the VM vanished in a picosecond. (Hopefully - this is the best case.)
Then @leesteken 's comment regarding restarting/rebooting follows...
My apologizes. I see where my mistake was, it was the networking that would stay up, not the VM itself. Thanks all.