Random crashes about pve-edge-kernel HOT 11 CLOSED

I am investigating a similar issue with an Intel-based machine showing the same symptoms since Linux 5.14, where the kernel seems to be deadlocking somehow.

I originally suspected ZFS, due to it being the only machine in my cluster that uses it. My solution for that machine is to run the Proxmox 5.13 kernel.

from pve-edge-kernel.

That's exactly how it feels, that the kernel blocks everything. Well, I'll try the Proxmox 5.13 kernel, but then it's probably a general kernel problem.

from pve-edge-kernel.

In the meantime I changed the computer, there I had random reboots and crashes like before.
I was able to find out that this is probably due to the combination RAM and C6 sleep stats are not compatible.
In the BIOS the "Power Down Enable" option must be set to Disable because the C6 state is unstable with DRAM "Power Down Enable".
I hope this helps others who have or had the same problem.

P.S.: I am now on the 5.19 kernel and haven't had a single reboot or crash.

from pve-edge-kernel.

Thanks for the update!

from pve-edge-kernel.

Are you using ZFS as well?

from pve-edge-kernel.

No

from pve-edge-kernel.

@fabianishere a bit of a long shot but it's possible that it's a hang caused by the iGPU drivers on the machine - since 5.14 people have been reporting that the i915 kernel driver isn't super stable, especially on more modern intel cpus. I've been experiencing this in both 5.15 and 5.16, but setting i915 to /bin/false or something like that has resolved the hangups (though I do lose GPU acceleration, but it's fine for the short term).

from pve-edge-kernel.

@DjSni appreciate though that's not helpful in your case, as you don't have an iGPU at all, especially not an intel one 😅

from pve-edge-kernel.

@DjSni appreciate though that's not helpful in your case, as you don't have an iGPU at all, especially not an intel one 😅

Thanks for the info, but my APU has an iGPU, it's actually a notebook processor. I have an AMD Ryzen 9 5900HX with Radeon Graphics and not to be confused with the Ryzen 9 5900X which has no iGPU. Unfortunately, turning off the iGPU is not an option for me as I currently need it for Emby.

from pve-edge-kernel.

I have updated the machine to Linux 5.18 since then and haven't seen these issues anymore.

from pve-edge-kernel.

Thanks to you too!

from pve-edge-kernel.