Can't add node to existing PVE4.2 cluster

[Odiobill]

Hi All,

I used to have an HA enabled three nodes PVE cluster. On of the servers died and all the VMs were correctly migrated to the other nodes.

Since fixing the other server will require some time, I've deployed a new one, installing PVE and configuring all the iSCSI volumes as in the other members.
Unfortunately, I can't add this new node to the cluster. The broken server is named glo-pve1, the current members are glo-pve2 and glo-pve3, while the new one should be glo-pve4. Here's what it happens:

--- ON GLO-PVE2 ---
root@glo-pve2.ams-ix:~# pvecm status
Quorum information
------------------
Date: Fri Jul 29 14:09:28 2016
Quorum provider: corosync_votequorum
Nodes: 2
Node ID: 0x00000001
Ring ID: 132
Quorate: Yes

Votequorum information
----------------------
Expected votes: 3
Highest expected: 3
Total votes: 2
Quorum: 2
Flags: Quorate

Membership information
----------------------
Nodeid Votes Name
0x00000001 1 91.200.17.135 (local)
0x00000002 1 91.200.17.137
root@glo-pve2.ams-ix:~# pvecm nodes

Membership information
----------------------
Nodeid Votes Name
1 1 glo-pve2 (local)
2 1 glo-pve3

--- ON GLO-PVE4 ---
root@glo-pve4.ams-ix:~# pvecm add pve-glo2
unable to copy ssh ID

Since ssh connections are enabled and working between the nodes, I used ssh-copy-id without any problem. I can login without any problems from glo2 to glo4 and the other way around, but the pvecm add command always keeps failing.

Is there anything else that I can check?

Thanks,

--
Davide

 

[dcsapak]

--- ON GLO-PVE4 ---
root@glo-pve4.ams-ix:~# pvecm add pve-glo2
unable to copy ssh ID

is it correct that you wrote pve-glo2 instead of glo-pve2 ?

 

[Odiobill]

Hi Dominik, thanks for your quick reply. Yes, that was a typo because I changed the names in the cut and paste.

Somehow, I managed to solve the issue. Is it possible that the problem was due to some non-running service? It did work after a reboot. Weird because I didn't touch glo-pve2 at all.

Anyway, it looks working fine at the moment. Thanks again!

 

[Shotgun Teddy]

I am actually encountering this issue right now. I have the head node, which I have installed the quorom/cluster on. I have built a second server, with the exact same version of proxmox (4.2-48); running kernel: 4.4.6-1-pve). At first it was failing because I had a different port and publickey access only setup in SSH. But I have changed all of that. Its now accepting password authentication, root login and its on port 22. Restarted the pve-cluster services on both boxes. Both can ping each other, they're on the same subnet. But when I try to add the node to the cluster I get this error:

root@pvemox:~/.ssh# pvecm add xxx.xxx.xxx.xxx
Password:
Permission denied (password,keyboard-interactive).
unable to add node: command failed (ssh 216.176.200.54 -o BatchMode=yes pvecm addnode chugmox --force 1)

The password is fine, because I can ssh to and from both nodes, on each node.