HW transcoding stopped working after OMV / Debian upgrade

Server Version#: 1.40.1.8227
Player Version#: 4.87.2 (web)

After upgrading Openmediavault to v7 (Debian to v12) hw accel stopped working (QuickSync). Plex is running via docker. According to server log, it seems plex wants to use hw but it fails for some reason. I assume there have been some driver changes to Debian 12 that could be the reason. Using dmesg on the host system shows some errors.

There are some other recent reports about hw accel issues that could possibly be related:
https://forums.plex.tv/t/ubuntu-24-04-hw-transcoding/873765/

https://forums.plex.tv/t/hw-transcoding-for-windows-clients-failing/875001/

https://github.com/linuxserver/docker-plex/issues/392

See also:
https://github.com/intel/media-driver/issues/1791

Compose file:

# https://jsonformatter.org/yaml-formatter
services:
  plex:
    image: 'plexinc/pms-docker:latest'
    container_name: plex
    network_mode: host
    environment:
      - PLEX_UID=1002
      - PLEX_GID=1005
#      - PLEX_CLAIM=...
    volumes:
      - '/sym/ssdStorage/appData/plex:/config'
      - '/sym/hddStorage/media/media:/media'
      - '/sym/ssdStorage/transcode:/transcode'
      - >-
        /var/log/dockerContainers/plex:/config/Library/Application Support/Plex
        Media Server/Logs
      - '/etc/timezone:/etc/timezone:ro'
      - '/etc/localtime:/etc/localtime:ro'
    devices:
      - '/dev/dri:/dev/dri'
    restart: unless-stopped

Dmesg:

[83829.956891] i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out
[83829.956982] i915 0000:00:02.0: [drm] Plex Transcoder[961924] context reset due to GPU hang
[83829.965849] i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:e7577efe, in Plex Transcoder [961924]
[83974.853874] i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out
[83974.853959] i915 0000:00:02.0: [drm] Plex Transcoder[963402] context reset due to GPU hang
[83974.860704] i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:e757fefe, in Plex Transcoder [963402]
[87065.047395] i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out
[87065.047482] i915 0000:00:02.0: [drm] Plex Transcoder[1001266 context reset due to GPU hang
[87065.051329] i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:e757fefe, in Plex Transcoder [1001266]
...

Plex Media Server.log (1.6 MB)

Bookworm 12 isn’t very new.
Quick google shows that you might only have Kernel 6.1? Debian -- Details of package kernel-image in bookworm
There were several gpu issues that came up depending on specific intel/media/qsv that I think got resolved in 6.2.
My UbuntuLTS machines are on kernel 6.5 and have no issues.

1 Like

Nice you nailed it thanks!

Search to install a 6.5 kernel via backports. Don’t install 6.8 yet.

1 Like

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.