non-HA cluster

[Waschbüsch]

Hello everyone,

I have a three-node cluster running including ceph and would like to know if the following is possible:
If one of the nodes goes down, I would like to be able to start one of the dead node's VMs on a remaining node.
This is decidedly not about HA, fencing and the like, but about bringing up a VM manually(!) on a different node when needed.
Since virtual disks reside on ceph, they are already accessible even if one of the nodes goes down.
The whole thing is more like cold standby, I guess.

Is it possible to 'migrate' a VM when the node it is running on is currently down? Anything I have to watch out for?

Thx,

Martin

 

[m.ardito]

on a workingnode, in /etc/pve move the .conf file under a working node...
the (stopped) vm will be instantly moved to the working node
if the cluster has quorum, the vm should start there
Marco

 

[Waschbüsch]

Cool. Thanks for that info. This is what I had hoped for that it's just possible to manage without much fuss...