[SOLVED] Problems after adding Node3

Dwx

New Member
Dec 9, 2025
2
1
3
Hello everyone,

I have a problem with my test cluster and would greatly appreciate your support.

I have set up a small Proxmox cluster consisting of 2 almost identical HP servers (HP ProLiant DL380 G7, each with 4 SSDs, etc.). Since I am missing a third full-fledged server for Ceph, I have temporarily added a PC as the third node:
Hardware: CPU: Intel Xeon E5-1650 v3 3.50GHz, RAM: 32GB

The entire setup runs in a completely isolated network; all devices are connected via a switch. PING, Corosync, pveproxy, etc. all basically work without any noticeable issues.

Problem:
As soon as the PC joins the cluster as Node3, problems begin:

  • The "Summary" view on Node3 cannot be opened and produces an error message: "Connection error 401: permission denied - invalid PVE ticket"
  • If I try to access information on Node1 or Node2 from Node3, the same error occurs.
  • Conversely, if I try to access Node3 from the cluster interface on Node1 or Node2, I also get errors.
The only thing that works consistently is the shell.

I have checked and adjusted the /etc/hosts file on all nodes.
All 3 nodes are running exactly the same Proxmox version; installation was done from the same USB stick, and I have also verified the versions via the command line.

Screenshots of the relevant configurations (including /etc/hosts) are attached below.

I appreciate any hints that could help solve the issue – thank you in advance!
 

Attachments

  • Bild (8).png
    Bild (8).png
    172.1 KB · Views: 0
  • Bild (7).png
    Bild (7).png
    161.6 KB · Views: 1
  • Bild (6).png
    Bild (6).png
    18.6 KB · Views: 1
  • Bild (5).png
    Bild (5).png
    149.2 KB · Views: 1
  • Bild (4).png
    Bild (4).png
    93.7 KB · Views: 1
  • Bild (3).png
    Bild (3).png
    34.2 KB · Views: 1
Hi!

It could be that the system times of the nodes are out of sync. You can run journalctl -u chronyd and check if it shows any errors regarding times being out of sync.
 
Last edited:
Hi!

It could be that the system times of the nodes are out of sync. You can run journalctl -u chronyd and check if it shows any errors regarding times being out of sync.
The issue was caused by the time setting. Thanks you very much - I completely forgot about that.
 
  • Like
Reactions: mkoeppl