I have two nodes: pve1 and pve2. I got both working fine and tried to create a cluster with them and I messed it up [*below]. I just want to uncluster them and recluster them correctly but they won't forget about each other.
I followed this to uncluster: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#pvecm_separate_node_without_reinstall
Now when I try to add pve2 to pve1 I get:
The web interface of both shows the other (just no green checkmark). I can see that pve2 still lists pve1 in /etc/pve/.vmlist:
And this is not a normal file I can edit.
How can I clear this all up without reinstalling both? I would not mind redoing pve2 as it has no VMs, but I am using one of pve1's VMs.
*What I messed up: The IP addresses get stored in /etc/hosts. I changed the IPs to put the nodes on a different network. Independently, they worked fine. But when the web-based cluster setup runs it uses /etc/hosts to look up its own IP rather than consulting the IPs on the interfaces. I have since fixed both /etc/hosts files but the clustering setup on the second node still tries to refer to the first node by the bogus IP.
I followed this to uncluster: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#pvecm_separate_node_without_reinstall
Now when I try to add pve2 to pve1 I get:
root@pve1 3% pvecm add 192.168.11.32
Please enter superuser (root) password for '192.168.11.32': ********
detected the following error(s):
* this host already contains virtual guests
Check if node may join a cluster failed!
The web interface of both shows the other (just no green checkmark). I can see that pve2 still lists pve1 in /etc/pve/.vmlist:
root@pve2 1% cat /etc/pve/.vmlist
{
"version": 1,
"ids": {
"401": { "node": "pve1", "type": "qemu", "version": 2 },
"100": { "node": "pve1", "type": "qemu", "version": 3 },
"101": { "node": "pve1", "type": "qemu", "version": 1 },
"103": { "node": "pve1", "type": "qemu", "version": 4 }}
}
And this is not a normal file I can edit.
How can I clear this all up without reinstalling both? I would not mind redoing pve2 as it has no VMs, but I am using one of pve1's VMs.
*What I messed up: The IP addresses get stored in /etc/hosts. I changed the IPs to put the nodes on a different network. Independently, they worked fine. But when the web-based cluster setup runs it uses /etc/hosts to look up its own IP rather than consulting the IPs on the interfaces. I have since fixed both /etc/hosts files but the clustering setup on the second node still tries to refer to the first node by the bogus IP.