PVE unreachable after some hours of activity

[legwoju]

Hi everybody,

I have had a dedicated server with OVH since November 22, 2024.
I installed proxmox ve 8 from their installation template. I don't have a subscription.
I made the updates via the GUI, changed the ssh listening port, put a let's encrypt certificate for the hypervisor and added the openvswitch-switch package.
For the moment, there are 2 VMs running on it with standard settings (OPNSense and Debian 12), an NFS mount.

After a few hours, the hypervisor is no longer reachable. OVH maintenance informs me that there's nothing on the screen, no pinging possible, they reboot.

I can't find anything in the logs and the logctl doesn't give me any leads.

I redid the installation, still using the ovh template, and the updates, and the same problem occurred: the server stopped working.
I tested kernel 6.11 and the server lasted almost 36 hours.

I'm currently on the OVH rescue and have launched a CPU/RAM and disk test loop.

No ZFS, no HA, no CEPH, just a single server.

Do you have any clues on how to identify this problem? Which logs?

Thanks in advance for your help

Julien

Dec 06 09:06:28 hw-srv-01 pvestatd[1276]: status update time (5.208 seconds)Dec 06 09:06:48 hw-srv-01 pvestatd[1276]: status update time (5.010 seconds)-- Boot d692d8109d94491fbe345221498e6bcb --Dec 06 09:25:02 hw-srv-01 kernel: Linux version 6.11.0-1-pve (build@proxmox) (gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU B>

 

[dodokacsa97]

Hi!

I have the same issue. I don't have ovh however, I lose connection from 4 hours to 5 days randomly. I did not find anything in the logs. I have no idea what the issue could be. I tried almost everything now.

Dominik

 

[fabian]

that's quite the gap in the logs there.. maybe you can open an SSH connection, run "journalctl -f" and wait for the problem to occur once more? if the logs are just not persisted to disk, you might see the actual problem that way..

 

[legwoju]

In my case, it was a parameter in the hardware configuration of my OPNSENSE VM that was causing the problem.

I left the server running with the VMs switched off for several days, then switched on a VM every 36 hours.
When I started OPNSense, PVE became unavailable after a few hours and not allowing SSH access.
my server has been up for 5 days