[SOLVED] General Protection Fault

[Xyrgh]

I've been running my current proxmox server for about 6 months now and it's been fine, but recently running into a couple of crashes.

I'm running several containers and a few VM's, including a Windows 10 VM. The crash only appears to happen when I log into remote desktop on the Windows VM, the host becomes unresponsive and I need to hard reset to bring it back up.

/var/log/messages attached.

The host is back up and running now and logging into the Windows VM has been fine, but wanted to troubleshoot before it happens again.

 

[Stefan_R]

All of the crashes in the syslog occur at seemingly random places, but most have in common that they occur in various 'alloc' functions, i.e. memory management. A guest crashing the host is usually also in indication of something wrong on the lower levels of the hypervisor, not the configuration.

I'd wager bad RAM, or some other hardware error...

 

[Xyrgh]

I've been having on and off issues with this and there was one common denominator, I added additional ram around December.

Proxmox is running on an Asus B450-F motherboard, non-ecc ram, 4 x 16GB sticks.

Memtest errored within an hour of testing all four sticks, but 24 hours of testing each stick and nothing. Turns out, Ryzen chipsets are overly sensitive to having four sticks installed, and I didn't have the XMP (docp) profile activated. Further, the XMP profile bumped the ram voltage from 1.2V to 1.3V.

So after making all these changes, running memtest for 24 hours with all sticks installed and zero errors, problem resolved.

Stefan, you were 100% correct, thanks for your assistance.