Unable to run HW acceleration in Jellyfin with compose

1st Official Post

    Hi all,


    I would like to use HW acceleration in a docker container through compose. The computer is an intel 11th gen and should able to do so, but it doesn't work. Here you have my compose file. I've tried lots of things with no luck. Could you help me, please?


    • Official Post

    - Make sure that the GPU is active in the Bios.

    - The user running the Jellyfin container must be in the render and video groups

    - I would change the lines that you included in devices to this: /dev/dri:dev/dri

    - In the Jellyfin GUI you must activate and configure hardware acceleration. https://jellyfin.org/docs/gene…rdware-acceleration/intel

    Quote from run000

    Could you help me, please?

    Start by posting the output of:

    uname -a

    This will show on what kernel you are since it will make the need OR NOT of the dockermods on the YML


    If <= 6.1 you will need the YML as is

    If >= 6.2 then you won't need the DOCKER_MODS line although it won't hurt nothing


    Add the PUID user to the groups mentioned by chente.

    Use the GUI Users page since it's easy to manage.


    Redeploy the container and you should be able to use Hardware Acceleration on Jellyfin:


    Does video hardware acceleration work on Debian12/OMV? This is a basic prerequisite for passing them on to the container. What is the output of vainfo under ssh in the terminal? Intel's 12th CPU generation with integrated GPU (e.g. N100) requires at least kernel 6.2 for the correct function of video hardware acceleration (see Soma for kernel version control). I don't know how it is with the 11th generation. The following packages must be installed: intel-media-va-driver (for hardware-accelerated decoding... which is usually enough) or intel-media-va-driver-nonfree (if hardware-accelerated encoding is also to be used) and vainfo to control the function of the video hardware acceleration.

    Quote from mischka

    Intel's 12th CPU generation with integrated GPU (e.g. N100) requires at least kernel 6.2 for the correct function of video hardware acceleration (see Soma for kernel version control)

    If using the linuxserver image, you can have hardware acceleration with a kernel <6.2 as long as you use the DOCKER_MOD for jellyfin.

    It will install the needed packages on docker container to take advantage of intel hardware.

    Code
    USER@HOST:~ $ uname -a
Linux HOST 6.1.0-18-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.1.76-1 (2024-02-01) x86_64 GNU/Linux


    Quote from mischka

    The following packages must be installed: intel-media-va-driver (for hardware-accelerated decoding... which is usually enough) or intel-media-va-driver-nonfree (if hardware-accelerated encoding is also to be used) and vainfo to control the function of the video hardware acceleration.


    Again, not really needed to install on HOST if the same above condition is applied (read, use the DOCKER_MOD on the YML)


    FYI, ;)


    [EDIT]

    What the DOCKER_MOD does on the container:

  • chente

    Added the Label resolved

Participate now!

Don’t have an account yet? Register yourself now and be a part of our community!

Register Yourself Login