I was unable to repair the cluster because I was unable to remove myself from it to recreate it or to join another cluster. What I didnt expect was accidentally destroying all of my vms in the pve menu. They are still running but i desperately need to readd them to repair my situation. Any ideas?
Accidentally deleted /etc/pve/nodes in attempt to repair cluster, now I have no VMs but VMs running.
- Thread starter croso
- Start date
The config is only used when something changes/starts, meaning that because they are running right now, the settings are already loaded into memory, and as long as you don't do anything to interrupt/change that (which might include trying to make backups and the like, so if you have an automatic backup task, you might want to pause that for now), you should be fine for now.
As you're already reading in the other post [1], first thing to probably do would be to copy the /var/lib/pve-cluster/config.db file on EVERY server in the cluster to somewhere safe, so for example
That way, if something can be restored (no guarantees), you have the files for it.
[1] https://forum.proxmox.com/threads/etc-pve-nodes-accidentally-deleted.55617/#post-684030
cp /var/lib/pve-cluster/config.db /home/pve-config.dbThat way, if something can be restored (no guarantees), you have the files for it.
[1] https://forum.proxmox.com/threads/etc-pve-nodes-accidentally-deleted.55617/#post-684030
I found that recreating to VMs found the existing vms running and allowed me to access them and running
'qm rescan'
Allowed me to find and reattach their drives.
It's unfortunate I have to recreate everything from memory and have to reinstall everything exactly the way it was to prevent an issue, but it's better than losing everything on my VMs and all that work it took to set them up.
'qm rescan'
Allowed me to find and reattach their drives.
It's unfortunate I have to recreate everything from memory and have to reinstall everything exactly the way it was to prevent an issue, but it's better than losing everything on my VMs and all that work it took to set them up.
Good to hear that you're making it work, but also know that because you already worked with changing VM's, you also probably lost any chance of just recovering them all-at-once with the correct config (unless you made those backups beforehand)
and you see - no offense - that a backup would have made life so much easier
even a small usb stick on a host and regularily cp /etc and config.db (or a scp, rsnyc, cp to nas etc...) is better than no backup
cant say this often enough