[SOLVED] Revive disassembled proxmox cluster

[sandsnake]

Good evening

There was a task to revive the proxmox cluster and I ask for help to make this as painless as possible.
I'm not sure of the reasons, but as far as I understand, the cluster was disassembled by hands. In the bash history on all nodes I see something like this:

cd /etc/pve/
mv corosync.conf corosync.conf_
cd ..
cd corosync/
rm -rf *
cd ..
killall pmxcfs
systemctl stop pve-cluster
rm /var/lib/pve-cluster/.pmxcfs.lockfile
systemctl start pve-cluster

As fact:

root@pve02:/home/vlad# pvecm status
Error: Corosync config '/etc/pve/corosync.conf' does not exist - is this node part of a cluster?

Now I see 5 single nodes with old "phantom" information in /etc/pve/ about virtual machines and lxc containters on other nodes. Two nodes have running up to 30 vms/lxc, one - 50 vms/lxc and 2 nodes have 7-10 vms/lxc.

I think, first create cluster on node with 50 vms. This will allow me to save up-to-date information about running virtual machines in /etc/pve for this node. Then via 'qm backup' and 'pct backup' store all vms on next node, shutdown all vms and join hypervisor to cluster on first node. Then restore all vms and lxc in cluster. And repeat this with all nodes.

Maybe there is an easier way to build a cluster? And I'm confused by the old information about other nodes in /ets/pve. Does it need to be cleared when creating a cluster and cleaning it up, or can it be used for connection? No new virtual machines were created and the numbers are correct.

PS. Sorry, my mistake with wrong thread.

 

[bbgeek17]

The only method that you can take where you can be sure there is no trap left behind is to back it up, do a clean re-install, then restore.
Yes, you can do one node at a time. Just make sure you start with a clean install.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[sandsnake]

Hmm. Unfortunately, a clean reinstallation is difficult.

1) /etc/pve - here proxmox store inforation about virtual machines. I think this information (or part) may be useful when creating a cluster
2) /etc/corosync - here config and information about cluster. I have backup corosync.conf and can use it.
3) All keys in .ssh directory is actual

In what places other traps can be found?

 

[sandsnake]

I'll answer myself: pvecm add 10.10.10.2 -force is very hellpful