Closed niallobr closed 2 years ago
After some searching, installing the below driver and re-initialising the device seems to have fixed it
docker exec -it sonarr apt install intel-media-va-driver-non-free -y
docker exec -it sonarr ffmpeg -init_hw_device vaapi=foo:/dev/dri/renderD128
Will leave open for now in case it's something that needs to be added to the image
Before you installed that package, was the intel-media-va-driver package installed? It should have been
Just curious if the non-free package is actually required or if the free one wasn't getting installed for some reason
Actually, looks like it's just more hardware dependent, some need non-free
updated to include this package at least on the ubuntu based images
I have sonarr-sma and radarr-sma containers set up using the same docker-compose.yaml structure but hwaccel processing has started failing in my sonarr-sma container recently. I've tried ubuntu/alpine builds and different versions of ffmpeg without luck. Any help would be great.
When I don't build the ffmpeg error in sma.log seems to be
or when I do build I get errors like 'could not allocate memory'.
Here's an example of a failed manual conversion on sonarr-sma and a successful conversion of the same file in radarr-sma
sonarr-sma manual.py
radarr-sma manual.py
autoProcess.ini
docker-compose.yaml
Not sure if relevant but I was having some issues with dependencies and log ownership so I run a script with these on both containers when they go up
Is there any obvious reason it would be failing in sonarr-sma but working in radarr-sma? I can see they are on different versions of Python but beyond that I'm not sure.