Hi,

 

I've been pulling my hair out trying to figure out why my Unraid server is crashing when hardware transcoding with Intel QSV. I've tried searching and googling, but can't seem to find anything.

 

The following is being generated in the syslog during transcoding:

 

Dec 20 15:37:02 Apollo kernel: i915 0000:00:02.0: [drm] Resetting rcs0 for CS error
Dec 20 15:37:02 Apollo kernel: i915 0000:00:02.0: [drm] Plex Transcoder[21854] context reset due to GPU hang
Dec 20 15:37:02 Apollo kernel: i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:00280000, in Plex Transcoder [21854]
Dec 20 22:41:36 Apollo kernel: i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out
Dec 20 22:41:36 Apollo kernel: i915 0000:00:02.0: [drm] Plex Transcoder[17899] context reset due to GPU hang
Dec 20 22:41:36 Apollo kernel: i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:8ed2fff2, in Plex Transcoder [17899]

 

I'm running Plex in a docker container.

 

After the crash, the system does not respond at all to icmp, webui or ssh. The only way to regain access is to power cycle the machine.

 

Any assistance would be appreciated.

apollo-diagnostics-20221222-1924.zip

Edited December 22, 20223 yr by squiddles88

I had a issue with my server where it would hard crash unexpectedly. Figured out it was that my PSU was cutting out when something put stress on the GPU. Swapped that out for another PSU and all is well. Fortunately I had a 1500WATT PSU on a shelf from another system that was retired so I had one going spare.

It's not the PSU. It's an appliance machine with a dedicated PSU. Haven't had these problems under the Terramaster OS. 

Edited December 24, 20223 yr by squiddles88

Did you ever get to the bottom of this?  I have the same issue, and it's driving me nuts.