Good to know you got over those drivers.
With Nvidia releasing a new driver so quickly after the previous one – AS PRODUCTION – I think you have your answer. That’s why I hold back and avoid anything released before a holiday
Speaking from the Plex perspective. 510.85.02 , which I have, is more than enough for things coming.
You should not need to touch your driver for quite some time unless you buy a new card which requires it
@ChuckPa All of my problems are back after using r525 for a few weeks. I’m still getting a XID 31 error, which according to Nvidia is a XID 31: MMU Error. Their recommended troubleshooting causes are Driver Error or User App Error
Secondly, the moderators on the Unraid forums are helping me troubleshoot. Is there supposed to be a ‘HardwareDevicePath=’ reference in the Preferences.xml? If so, I don’t have one.
Also, it appears Plex doesn’t have access to /dev/dri when I rundocker exec Plex-Media-Server ls -la /dev/dri. It returns ls: cannot access '/dev/dri': No such file or directory. Are there other checks I should be making to verify access?
This is the container definitions failing.
It’s not passing the host into the container.
Is there something in the docker settings or is this a corrupted Plex file or permission problem?
I have --runtime=nvidia set and all of the nvidia driver variables defined.
It’s not a corrupted Plex image file.
If using docker run, it looks like this
sudo docker run \
-d \
--name plex \
--network=host \
-e PLEX_CLAIM="claim-J3YbuBpaBhE9oJS1jsma" \
-e TZ="EST" \
-e LANG="en_US.UTF-8" \
-e PLEX_UID=1000 \
-e PLEX_GID=1000 \
-e PUID=1000 \
-e PGID=1000 \
-h dockerplex \
-v /sata/dockerplex:/config \
-v /sata/dockerplex/tmp:/tmp \
-v /sata/dockerplex/transcode:/transcode \
-v /vie:/data \
-v /glock:/glock \
--device=/dev/dri:/dev/dri \
plexinc/pms-docker:plexpass
Are you expressly declaring the --device clause ??
I now have --runtime=nvidia --device=/dev/dri:/dev/dri in extra parameters and nothing happens.
That fixed the definitions access to /dev/dri but still getting the XID31 errors. I never had to add --device=/dev/dri:/dev/dri when the gpu WAS working.
Still getting CPU transcoding
XID 31 – very easily could be the Unraid kernel driver for the Nvidia.
It could be, but I’m using the same NVENC driver to transcode video files using Handbrake and it works flawlessly.
My next step is to try out Jellyfin to see if I get the same problem.
Finishing this thread with the final solution.
The 1070 video card ended up being the problem, not the drivers, docker or Unraid. I purchased a used GTX 1660. Before I swapped the card, I ran a dockerized version of GPU-BURN and the card would not complete the test before throwing a XID: 31 READ_ERROR. My syslog was throwing a XID:31 WRITE_ERROR when Plex would try to transcode before.
I swapped the card to the 1600, ran the GPU BURN again, and successfully completed the test. HW transcoding is back to normal after swapping to a new card.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
