unexpected new node in cluster

[endimion]

Hello, I am new to proxmox. I installed proxmox 3.4-3 on a distant server with good capacity and created a network of VMs (about 10) & 3 containers under a single node ("n34"). I also programmed daily backup locally (on the same server). Everything seemed to be fine for several weeks, then I had a strange behaviour 2 days ago, when I connected to the pve interface, the single node "ns34" dissapeared and I got another empty node instead called "213". I never created this node (as far as I know).
My applications running on the vms seem to respond (interface available over the web). I checked several configuration files (the cluster.conf is missing) and the logs. Cannot find a track of an action that could have created this node. I then restarted my cluster. Now I have my previous node ("n34") also showing in the interface. When connecting with ssh, I can see both nodes and node 213 seem to be empty. I thought of deleting it completely, but I am not sure it won't also delete my actual content. When I connect on ssh, the prompt now lists "root@213" and in the .member file only the 213 node is listed. Please help.

 

[Mr.Holmes]

Hello endimion,

the node name is defined in /etc/hostname

In file /etc/hosts you should have a line like this (IP number and my... are examples):

192.168.99.99 mynodename.mydomain.tld mynodename pvelocalhost

Moreover, GUI shows each subdirectory which can be found by

ls /etc/pve/nodes/

as node - regardless if it really exists or not.

Adapt these files/directories accordingly!

Kind regards

Mr.Holmes

 

[endimion]

Mr. Holmes,
Thank you for your reply. Hostname is correctly registered in the files you mention. On Wednesday we launched a manual backup to a datastore and while typing some commands like "ls" in the command-line, I noticed that I have again a new node in the cluster, called "ls" this time, and my prompt now lists "root@ls". We now suspect something is wrong with the hardware? I continue searching for clues in the logs and looking for a solution to inspect hardware status.

 

[endimion]

I checked the /etc/hosts file and I noticed I had two hosts declared: localhost and the remote host on which my proxmox runs. So I deleted the localhost line and removed the 2 unwanted nodes from the cluster, than rebooted the system. Everything seems to be back on the road now. Thank you for the advice.