Audio Normalization Task

[Jamie Chapman]

It looks like DBG messages are in there. Pretty sure it was something I was doing wrong last evening. I tried deciphering the logging, but I don't see anything that stands out which might be the root of the problem. It is highly consistent though, which is good, kinda.

[TheDreadPirate]

I'm going to say go ahead and submit a github issue. This feels like a race condition, as pcm suggested. ffmpeg hasn't fully stopped before Jellyfin attempts to delete the concat file, which is generated by ffmpeg while analyzing your music to determine the normalization value.

https://github.com/jellyfin/jellyfin/issues

[Jamie Chapman]

Roger that, sounds about right. it's something under the hood.

[Jamie Chapman]

https://github.com/jellyfin/jellyfin/issues/12145

[Jamie Chapman]

I am guessing here. But, based on the log message.. it seems like the audio-normalization job creates a temp file in the cache and executes ffmpeg to normalize the audio of your album and then it tries to delete the temp file after it's done with the normalization calculations.
The schedule task spawns the ffmpeg process as a child process asynchronously (based on the github code).

I'm guessing that ffmpeg is not done doing it's normalization thingy and the scheduled task process moves on and attempts to delete the file which ffmpeg is still working on...

One thing you could do is turn on debug-logging temporarily and see what the log says when you kick off the audio normalization job and right before it fails... It might help understand the issue better.

But, I think, you might want to open an issue on github for this...

Thanks pcm. That all sounds spot on. Appreciate the help!

[Jamie Chapman]

Quick update: I moved jellyfin over to a docker last night, and the audio normalization task is now completing without errors.

Running the standard jellyfin/jellyfin:latest in Docker Desktop

[Leophir]

Quick update: I moved jellyfin over to a docker last night, and the audio normalization task is now completing without errors.

Running the standard jellyfin/jellyfin:latest in Docker Desktop

I have the opposite issue, I moved JellyFin from the OS to a container and audio normalization cannot complete (unlike before). Transcoding is also significantly slower even though everything seems alright with the hardware in the logs.

[TheDreadPirate]

Can you share your jellyfin logs? When I switched from a direct linux install to a docker install I did not notice a difference in performance.