Jellyfin Forum
Docker run setup not using hardware transcode - Printable Version

+- Jellyfin Forum (https://forum.jellyfin.org)
+-- Forum: Support (https://forum.jellyfin.org/f-support)
+--- Forum: General Questions (https://forum.jellyfin.org/f-general-questions)
+--- Thread: Docker run setup not using hardware transcode (/t-docker-run-setup-not-using-hardware-transcode)



Docker run setup not using hardware transcode - sk1036 - 2025-05-05

Hello fellow fins of jelly!

I have previously had jellyfin running on ubuntu for a while with hardware transcoding working.

I have recently moved over to a docker container setup, this is mostly working except for hardware transcode.

I'm almost certainly missing something daft.

I have passed my /dev/dri path as a device.

If I docker bash into the container /dev/dri shows entries for a card1 & renderD128 device.

I've tried flip flopping between QSV/VAAPI and leaving blank the render device or explicitly naming /dev/dri/renderD128.

But my CPUs all hit the rafters when I transcode.

I'm not seeing much in the way of logs, perhaps I'm missing something but essentially the software transcode is working so I guess nothing much to report.

I'm not using compose, I run the following to create the container:

docker run -d \
  --name jellyfin \
  --user 1000:1000 \
  --device /dev/dri \
  --group-add video \
  -v /my/configConfused-faceconfig \
  -v /my/mediaConfused-facemedia \
  -p 8096:8096 \
  jellyfin/jellyfin:latest

Any pointers appreciated!


RE: Docker run setup not using hardware transcode - bitmap - 2025-05-06

Could you provide your hardware specs? A good way to ensure you're mapping the correct hardware is specificity. In general, the iGPU is /dev/dri/renderD128, which would make the device part of your run command:

Code:
--device /dev/dri/renderD128:/dev/dri/renderD128

Also, try:
Code:
docker exec -it jellyfin /usr/lib/jellyfin-ffmpeg/vainfo

Follow the instructions here and see if you missed anything: https://jellyfin.org/docs/general/post-install/transcoding/hardware-acceleration/intel#configure-with-linux-virtualization