Proxmox GUI crash, remove server from cluster

If no nodes or servers are displayed in the Proxmox VE web interface, this may be due to a crashed server or a damaged cluster configuration. If you then restart the server, it will no longer work! These instructions describe how to remove a single node from the cluster and restart it as a standalone server. If your server is still running, start at point 4.

The server has already been restarted and is no longer running / is stuck during the boot process:

The server can no longer mount /etc/pve correctly during the boot process and stops the boot process.

1. When booting in the GRUB menu

1. Restart the server.

2. In GRUB, select the current Proxmox entry and press e (Edit).

3. In the line beginning with linux, add the following at the end:

   init=/bin/bash


Result: You end up in a root shell, system services are not running, / is initially write-protected.

2. In the shell prompt (recovery mode)

mount -o remount,rw / 
systemctl disable pve-cluster 
systemctl disable corosync 
systemctl disable pve-firewall 
rm /var/lib/pve-cluster/config.db reboot

3. Reactivate services

systemctl enable pve-cluster
systemctl enable corosync
systemctl enable pve-firewall

4. Rebuild node / If your server is still running

systemctl stop pve-cluster corosync pvedaemon pveproxy pvestatd
rm -rf /etc/pve
rm -rf /var/lib/pve-cluster
rm -rf /etc/corosync

5. Start services

pvecm expected 1
systemctl start pve-cluster
systemctl start corosync
systemctl start pve-firewall
systemctl restart pveproxy pvedaemon

6. Check the web interface

Open the GUI: https://IP:8006
Log in and check whether the node is running as a standalone.

7. Add node again

On each new node to be added (adjust IP): pvecm add 192.168.123.10 --use_ssh