Plex Hardware Acceleration NVIDIA GPU

[fabio.lanza]

Hi everyone, I installed UNRAID (coming from TrueNAS after some struggles), and I want to deploy Plex with NVIDIA GPU hardware acceleration. I followed some tutorials online, but I was not able to make it work. Can you please advise?

 

I have 32GB ECC RAM and NVIDIA QUADRO P400. I have installed the NVIDIA Driver and the GPU Statistics plugins, and they seem to work. I can see the GPU stats in the dashboard. Then, I also installed the official Plex application as offered in the Apps section.

 

I have configured the Plex template to the following:

Extra Parameters: --runtime=nvidia

Additional Container Variable NVIDIA_VISIBLE_DEVICES: all

/transcode -> /dev/shm

 

When I apply the template, I get the error "unknown or invalid runtime name: nvidia.":

docker run
  -d
  --name='Plex-Media-Server'
  --net='host'
  --pids-limit 2048
  -e TZ="America/Los_Angeles"
  -e HOST_OS="Unraid"
  -e HOST_HOSTNAME="NAS"
  -e HOST_CONTAINERNAME="Plex-Media-Server"
  -e 'PLEX_CLAIM'='claim-XXXXXXXXXXX'
  -e 'PLEX_UID'='99'
  -e 'PLEX_GID'='100'
  -e 'VERSION'='latest'
  -e 'NVIDIA_VISIBLE_DEVICES'='all'
  -l net.unraid.docker.managed=dockerman
  -l net.unraid.docker.webui='http://[IP]:[PORT:32400]/web'
  -l net.unraid.docker.icon='https://raw.githubusercontent.com/plexinc/pms-docker/master/img/plex-server.png'
  -v '/dev/shm':'/transcode':'rw'
  -v '/mnt/user/STORAGE/':'/data':'ro'
  -v '/mnt/user/appdata/Plex-Media-Server':'/config':'rw'
  --runtime=nvidia 'plexinc/pms-docker'

docker: Error response from daemon: unknown or invalid runtime name: nvidia.
See 'docker run --help'.

The command failed.

 

I appreciate any support and help.

 

Thanks,

Fabio

[deadnote]

Hi

Did you try with your GPU ID instead of 'all' for NVIDIA_VISIBLE_DEVICES variable ?

You can find the ID in the nvidia driver settings. It starts with "GPU-"

[fabio.lanza]

44 minutes ago, deadnote said:

Hi

Did you try with your GPU ID instead of 'all' for NVIDIA_VISIBLE_DEVICES variable ?

You can find the ID in the nvidia driver settings. It starts with "GPU-"

I did try with the GPU ID found in the NVIDIA driver settings. Is it strange that I get an error in the runtime parameter?

[deadnote]

which docker image are you using?
I'm using the linuxserver image