Proxmox Node Frequently Crashes and Becomes Unreachable - Need Help Troubleshooting

[verishare]

Recently, my Proxmox node has been crashing very frequently, usually at least once every two days. The situation is becoming increasingly frustrating as it makes it difficult to remotely access and manage the node through the Proxmox Web UI.

Whenever a crash occurs, the Web UI and SSH becomes unreachable, making remote troubleshooting particularly challenging.

The node hosts several Windows and Linux VMs. However, based on my observations, CPU and RAM usage rarely reach their maximum capacity, so resource exhaustion does not seem to be an obvious cause.

 

[npiet]

Hi @verishare and welcome to the forum.

I have seen a similar problem with the Intel E1000E driver, so it would be interesting to see if this is the same driver, or a different one.

Are you able to share any journal outputs, or kernel diagnostic messaging - can you ping both to and from the host? and is the host still responsive when you are physically at the machine?

For a journal output, please run the following
journalctl --since "2 days ago" | gzip > /tmp/$(hostname)-syslog.txt.gz

This will collect all the logs (journals) from the past two days and place it in a .gz archive.

For the kernel diagnostics, could you please run:
dmesg | gzip > /tmp/$(hostname)-dmesg.txt.gz

This will dump the kernel message buffer and output it to a .txt file in a .gz archive.

If you can't perform these remotely, you may have to do this at the machine itself.

Please let me know if you have any more questions!