Web UI not working after joining cluster

[nerdynetworking]

Hello,

I have two Proxmox VE servers on the same version (the newest right now) Proxmox VE 7.2 they both were working independently today and for a while. But now I joined server A to server B in this example server A no longer loads the web UI anymore but SSH is working.

Server B where the cluster was made on web ui and ssh still works on it and you can see in the server B inside the UI host a is listed there but with an RED X indicating an issue.

I probably broke the server because I followed forums post to use SSH to join the host into a cluster that already has VMs on it using these steps:

On node1 (with guests)
Create a new cluster or get join information.

On node2 (with guests)
scp -r /etc/pve/nodes/* to node1:/etc/pve/nodes
rm -r /etc/pve/nodes/*
Join cluster.

I suppose I am reaching out to find out how to reverse this change so I can get host A back up and running. I dont even care if its still in the cluster at this point.

I've physically rebooted the server using iDrac 2 times already and still get nowhere.

 

[nerdynetworking]

Also tried this:
root@nerdyhost1:~# pvecm nodes

Membership information
----------------------
Nodeid Votes Name
2 1 nerdyhost1 (local)
root@nerdyhost1:~# pvecm delnode nerdyhost1
Cannot delete myself from cluster!
root@nerdyhost1:~#

 

[nerdynetworking]

Update:

using this command on the affected host A fixed the web UI: pvecm updatecerts

So maybe just a cert issue it was but now all my LXC and VMs are gone but the storage is still there. How do I persay re-register them like I would in like VMWare equivalent?

 

[fiona]

Hi,

So maybe just a cert issue it was but now all my LXC and VMs are gone but the storage is still there. How do I persay re-register them like I would in like VMWare equivalent?

not sure why the guest configuration files got lost. I don't think that should happen. What is the output of pvecm status now and ls /etc/pve/nodes/*/*/*.conf?

If you have any backups, you can use the Show configuration button and copy it back. Otherwise, I'm afraid you need to re-create the configuration files with IDs matching the IDs of the disks manually. One way would be to create dummy VMs/LXCs with the matching ID and then manually edit the disk settings in the .conf file. For VMs you can even choose to not create any new disks during creation, for LXCs that's not possible, so you'd need to take care of the unwanted rootfs volume.

 

[neilzhao200]

Add my experience here, it's the joining server's cert issue for me.
I have both server1 and server2 on tailscale, and I've been using tailscale dns to access both server UI.
After server2 joining server1 cluster, the tailscale generated ns for server2 stops working.
The appearance is
* After accessing server2 on browser, I can login, but it shows "connection error"

reading the thread, I realize I can ssh to the server correctly, and web UI works with ip address.
So that means it's the server2's cert has problem