Decommission Cluster

[Brononius]

Hey,

Today, I'm using a cluster in the wrong way.
I've got 2 nodes (a big server with all VM's, and a small desktop with only 1 VM (=backup server)). I built the cluster to have 1 management view. But after some months/years, I now came to the conclusion that I would be better off without the cluster config. And just reinstall the small one with just backupserver.

In the past, I've removed once a cluster, and this went completely wrong. I've lost the configs of all VM's and so.
That's the reason that I would like to inform myself before doing this. :$

Is it safe to follow the procedure https://pve.proxmox.com/pve-docs/pve-admin-guide.html#pvecm_separate_node_without_reinstall ?
Or should I take care of 'special' things?
The server contains everything locally (no shared storage or so).

 

[oguz]

hi,

yes the procedure should be safe to follow, but it's not the recommended one.

Or should I take care of 'special' things?

ideally you can learn from the mistake and make a backup of all your configurations in case something does go wrong.

 

[Brononius]

ideally you can learn from the mistake and make a backup of all your configurations in case something does go wrong.

When you say configurations, do you mean the VM-configuration of Proxmox?
The id.conf files from /etc/pve/nodes/MY_HOSTNAME ?
These are backupped every night through the backup server, after this script:

rsync -avh /etc/pve/qemu-server/* /backup/config_qm/.

So I'm learning.


But is there something else to take into account before removing the cluster?


ps I'm pretty sure that my data of all VM's are backupped through the backupserver (backuppc). But I looking for the fasted way to recover if my installation fails...

 

[Brononius]

I've followed this morning the procedure.
Everything is still running, I can start/stop VM's...

But when I run pvecm, it seems that corosync.conf is missing.
Normal, since I removed this in the third step?

pvecm delnode vidia
Node/IP: vidia is not a known host of the cluster.

pvecm nodes
Error: Corosync config '/etc/pve/corosync.conf' does not exist - is this node part of a cluster?

The PVE cluster is running:

systemctl status pve-cluster
● pve-cluster.service - The Proxmox VE cluster filesystem
   Loaded: loaded (/lib/systemd/system/pve-cluster.service; enabled; vendor preset: enabled)
   Active: active (running) since Tue 2020-11-24 08:30:25 CET; 11min ago
  Process: 3514 ExecStart=/usr/bin/pmxcfs (code=exited, status=0/SUCCESS)
Main PID: 3518 (pmxcfs)
    Tasks: 6 (limit: 7372)
   Memory: 45.3M
   CGroup: /system.slice/pve-cluster.service
           └─3518 /usr/bin/pmxcfs

nov 24 08:30:24 stampertj systemd[1]: Starting The Proxmox VE cluster filesystem...
nov 24 08:30:25 stampertj systemd[1]: Started The Proxmox VE cluster filesystem.

 

[oguz]

But when I run pvecm, it seems that corosync.conf is missing.
Normal, since I removed this in the third step?

yes, check also at the other node to see if it worked as expected