I removed one of two nodes in my cluster, reformatted for other purposes, and then tried to remove it from the web interface etc ... learned about quorum issues ... I had previously been able to get things running by bringing that down to 1 but in trying to remove the cluster entirely I'm pretty jammed up.
I think I started with instructions from https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_remove_a_cluster_node ...
and now some AI help has not gotten me any better.
I've
And When I get there, It says -bash: /etc/pve/.members: Input/output error
Something with the FUSE filesystem isn't restoring correctly, the existing .members file seems a little less than expected:
{
"nodename": "simon",
"version": 0
}
and the web interface now won't come up and containers are not starting. Any ideas?
I think I started with instructions from https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_remove_a_cluster_node ...
and now some AI help has not gotten me any better.
I've
| Step | Action | Command/Example | Notes |
|---|---|---|---|
| 1 | Stop all Proxmox services | systemctl stop pve-cluster pveproxy pvedaemon | Ensure no processes interfere. |
| 2 | Remove pmxcfs database | rm -f /var/lib/pve-cluster/config.db | Forces a fresh start, preserving backups if any. |
| 3 | Check for cluster remnants | ls /etc/pve; ls /etc/corosync (previously removed these) | Ensure no corosync.conf or cluster files remain. |
| 4 | Test pmxcfs manually | pmxcfs -d -f then echo '{"nodename":"simon","id":1,"online":1,"ip":"192.168.1.4","version":0}' > /etc/pve/.members |
And When I get there, It says -bash: /etc/pve/.members: Input/output error
Something with the FUSE filesystem isn't restoring correctly, the existing .members file seems a little less than expected:
{
"nodename": "simon",
"version": 0
}
and the web interface now won't come up and containers are not starting. Any ideas?
Last edited: