10.9.2 docker crash

[toytown]

Yesterday i updated my jellyFin server to 10.9.2 using the official docker.  About 2 hours later my friend connected and started streaming a film.   At some point during the stream (after transcoding had finished) the docker running jellyFin crashed.   All the remaining dockers on the server and the server itself continued to run perfectly, the docker build of jellyfin up to this point has been 100% stable using 10.8.13 so i don't believe its a hardware issue.   The only installed plugin is version 14 of "PlaybackReporting"

After restarting the docker i was able to playback the movie (albeit it using a different client) without it causing the crash.

I've included the logs @ https://text.is/52RY3

[TheDreadPirate]

I see that it was transcoding that video. Double check that you didn't run out of space in the volume that transcodes are written to.

If that is not the case, definitely submit a github issue.

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

[toytown]

The video had finished transcoding almost an hour before the crash. Transcode directory had over 500GB free.

[toytown]

Had a crash today using 10.9.7, from the log it looks like the same error as i posted with 10.9.2. The transcode had finished about an hour before and it occurred just as the client had finished watching the video.

Log @ https://text.is/3Z9Q1

[TheDreadPirate]

That looks like the NIC lost connection or the NIC itself was unavailable.

Code:

[2024-07-28 13:02:20.994 +00:00] [ERR] [232] Microsoft.AspNetCore.Server.Kestrel: Unexpected exception in TimingPipeFlusher.FlushAsync.
System.Net.Sockets.SocketException (113): No route to host

[toytown]

That looks like the NIC lost connection or the NIC itself was unavailable.

Yep, but the NIC is fine.   
The router does not show a dropped connection in its logs.
Dmesg on the docker host doesn't show anything and all the other dockers were running fine at the time and uploading data.
I've set the docker to auto restart from now on, i would prefer it not to, so i know when it crashes.

I still think its a bug, even if it loses a network connection or believes it has, it shouldn't pull down the entire docker.

[34626]

Please share your docker-compose.yml files content - what docker image do you use? The official or a third party?

[TheDreadPirate]

That looks like the NIC lost connection or the NIC itself was unavailable.

Yep, but the NIC is fine.   
The router does not show a dropped connection in its logs.
Dmesg on the docker host doesn't show anything and all the other dockers were running fine at the time and uploading data.
I've set the docker to auto restart from now on, i would prefer it not to, so i know when it crashes.

I still think its a bug, even if it loses a network connection or believes it has, it shouldn't pull down the entire docker.

Can you get your docker daemon logs and share those?

Code:

sudo journalctl -u docker.service --no-tail > docker_logs.txt

[toytown]

Can you get your docker daemon logs and share those?

I checked them at the same time, there was no entries from any of my dockers during the timeframe, it doesn't seem to log a lot though.

[TheDreadPirate]

Is your container using bridge or host networking?