Immich errors and Quicksync

[Riggspsk]

I have used the community script to install Immich multiple times on 2 different nodes over the last 24 hours. Each instance has an error with Quicksync. I have Plex Server running without problems in another container, and it transcodes using Quicksync without problems. Intel Gen 12 i7.
Log:

`  built with gcc 14 (Debian 14.2.0-19)
  configuration: --prefix=/usr/lib/jellyfin-ffmpeg --target-os=linux --extra-version=Jellyfin --disable-doc --disable-ffplay --disable-static --disable-libxcb --disable-sdl2 --disable-xlib --enable-lto=auto --enable-gpl --enable-v>
  libavutil      59. 39.100 / 59. 39.100
  libavcodec     61. 19.101 / 61. 19.101
  libavformat    61.  7.100 / 61.  7.100
  libavdevice    61.  3.100 / 61.  3.100
  libavfilter    10.  4.100 / 10.  4.100
  libswscale      8.  3.100 /  8.  3.100
  libswresample   5.  3.100 /  5.  3.100
  libpostproc    58.  3.100 / 58.  3.100
[AVHWDeviceContext @ 0x6211c2ce9cc0] Failed to open /dev/dri/renderD128 as DRM device node.
[AVHWDeviceContext @ 0x6211c2ce9cc0] No VA display found for device /dev/dri/renderD128.
Device creation failed: -22.
Failed to set value 'qsv=hw,child_device=/dev/dri/renderD128' for option 'init_hw_device': Invalid argument
Error parsing global options: Invalid argument
^[[39m
^[[31m[Nest] 173  - ^[[39m10/04/2025, 12:25:10 PM ^[[31m  ERROR^[[39m ^[[33m[Microservices:MediaService]^[[39m ^[[31mError occurred during transcoding: ffmpeg exited with code 234: Device creation failed: -22.
Failed to set value 'qsv=hw,child_device=/dev/dri/renderD128' for option 'init_hw_device': Invalid argument
Error parsing global options: Invalid argument
^[[39m
^[[31m[Nest] 173  - ^[[39m10/04/2025, 12:25:10 PM ^[[31m  ERROR^[[39m ^[[33m[Microservices:MediaService]^[[39m ^[[31mRetrying with QSV acceleration disabled^[[39m
^[[32m[Nest] 173  - ^[[39m10/04/2025, 12:25:17 PM ^[[32m    LOG^[[39m ^[[33m[Microservices:MediaService]^[[39m ^[[32mSuccessfully encoded 416fdc19-d944-487b-835f-3b6ec0dc0ba9^[[39m
^[[32m[Nest] 173  - ^[[39m10/04/2025, 12:25:30 PM ^[[32m    LOG^[[39m ^[[33m[Microservices:MediaService]^[[39m ^[[32mTranscoding video 21bc8bad-79b2-4f51-bca8-b303f4899110 with QSV-accelerated encoding and decoding^[[39m
^[[31m[Nest] 173  - ^[[39m10/04/2025, 12:25:30 PM ^[[31m  ERROR^[[39m ^[[33m[Microservices:MediaRepository]^[[39m ^[[31mffmpeg version 7.1.2-Jellyfin Copyright (c) 2000-2025 the FFmpeg developers`

[vhsdream]

@Riggspsk At the moment Debian 13 LXCs that use hardware acceleration use the wrong GID for the render group when the LXC is created. Easiest way to deal with it right now is to change it manually in the ProxmoxVE UI.

Change the GID to 992:

[Riggspsk]

Yes I did. Each time I installed from scratch. First time I saw the error I thought I accidently hit 'No'. But each other time I double checked that I did.

[vhsdream]

If your LXC is privileged, then you should be able to ensure the immich user is in both the video and render groups, then make sure /dev/dri/card0 is owned by the video group, and /dev/dri/renderD128 is owned by the render group, like so:

Although it may look different in a privileged LXC.

[Riggspsk]

Looks correct to me:

I've been googling this for days, and completely out of ideas.

[vhsdream]

No it's not correct, your image proves it: The group that owns /dev/dri/renderD128 must be render.

Change it: chown root:render /dev/dri/renderD128

[Riggspsk]

Thank you, that did it. My eyes saw what they wanted to see and not what was actually on the screen.