Jellyfin Instability and Crashes

Jellyfin Instability and Crashes - Printable Version

+- Jellyfin Forum (https://forum.jellyfin.org)
+-- Forum: Support (https://forum.jellyfin.org/f-support)
+--- Forum: Troubleshooting (https://forum.jellyfin.org/f-troubleshooting)
+--- Thread: Jellyfin Instability and Crashes (/t-jellyfin-instability-and-crashes)

Pages: 1 2 3

Jellyfin Instability and Crashes - Jam - 2024-07-22

Hi,

I hope someone can help me here as i've been pulling my hair out these last few days. Since last week, i've noticed my Jellyfin server halt up and crash, more and more. It started around last week on 10.9.7 and I am now on 10.9.8.

It looks to me, that ffmpeg is using up loads of CPU power which then leads to it ultimately crashing. I use my Intel iGPU and I have done for the best part of 6 months (since I built this new server). Everything has been fine, nothing has been changed.

Now? My server can barely run for more than 5 hours before halting up, front-end seeming to be very very very slow then for it to inevitably crashing. I have to restart it on a container level around 5 times a day and it's progressively getting worse and worse.

I am using the official docker container image on my Unraid server. I have attached the logs to allow some more light on this. If you need any other logs, let me know cause this is honestly getting me to the end of my tether now. I can't wrap my head around why it's been so stable and now all of sudden, it's just progressively getting worse and worse.

https://gist.githubusercontent.com/JamsRepos/3de805879e01d3bc24e6357f7398a163/raw/aa11f23b9ce1b99c8e5bf882916b19a4027fd852/gistfile1.txt

Everywhere it restarts, is me restarting it manually to fix things. I have barely any plugins anymore as mainly they're not needed and also to try and provide more stability to the server. There's no reason to me why it can't run properly as it has done for half a year perfectly, since moving up to 10.9 there's been a good few issues which is also understandable.

Any help is massively appreciated.

RE: Jellyfin Instability and Crashes - Efficient_Good_5784 - 2024-07-22

I'm thinking that the issue might be the simkl plugin you have installed.

It's trying to scrobble over and over again, but it keeps failing. Did you disable that?

Also, how are your episodes named? I see a lot of this in the logs as errors:

Code:
[ERR] [73] MediaBrowser.Providers.TV.EpisodeMetadataService: Error in "TheMovieDb"

System.FormatException: The input string '252284-the-red-king' was not in a correct format.

RE: Jellyfin Instability and Crashes - Jam - 2024-07-22

(2024-07-22, 09:34 PM)Efficient_Good_5784 Wrote: I'm thinking that the issue might be the simkl plugin you have installed.

It's trying to scrobble over and over again, but it keeps failing. Did you disable that?

Also, how are your episodes named? I see a lot of this in the logs as errors:

Code:
[ERR] [73] MediaBrowser.Providers.TV.EpisodeMetadataService: Error in "TheMovieDb" System.FormatException: The input string '252284-the-red-king' was not in a correct format.

Thanks for the response! The SIMKL Plugin seems to report an error when it's actually a warning, it's merely stating that a user is not logged in as only a small handful use SIMKL. I can however give disabling it a go though, to rule it out. I also have been speaking to SIMKL developers to see if they can see a cause with their plugin too.

My episodes are names following the trash guides naming scheme.

Code:
{Series TitleYear} - S{season:00}E{episode:00} - {Episode CleanTitle} [{Custom Formats }{Quality Full}]{[MediaInfo VideoDynamicRangeType]}{[Mediainfo AudioCodec}{ Mediainfo AudioChannels]}{[MediaInfo VideoCodec]}{-Release Group}

I have only got TMDB as my old metadata provider, should I perhaps provide another backup? Even then, would this cause such a hold up for a select few TMDB metadata clashes? Surely that's just result in the show just not being scanned in.

RE: Jellyfin Instability and Crashes - Efficient_Good_5784 - 2024-07-22

(2024-07-22, 09:48 PM)Jam Wrote: Thanks for the response! The SIMKL Plugin seems to report an error when it's actually a warning, it's merely stating that a user is not logged in as only a small handful use SIMKL. I can however give disabling it a go though, to rule it out. I also have been speaking to SIMKL developers to see if they can see a cause with their plugin too.

This is a huge chunk of the warning in your log file:

Code:
[ERR] [90] Jellyfin.Plugin.Simkl.Services.PlaybackScrobbler: Can't scrobble: User "REDACTED_USER" not logged in (True)

Maybe it's a warning, but I bet it's causing slowdowns on your system due to the huge frequency it's spammed in the logs.
Doing a find-all on the logs, that exact line is shown 1668 times on the logs.

What server version where you on before upgrading to v10.9.7?

As for the episode naming, I see two types from the logs:
1. Home Fires (2015) - S01E05 - Episode 5 [redacted].mkv
2. DARLING in the FRANXX (2018 ) - S01E07 - 007 - Shooting Star Moratorium [redacted].mkv

Jellyfin should ignore everything in the square brackets.
I would suggest to try removing the year tag and the extra episode numbers. As in, only keep the "S##E##" format:
1. Home Fires S01E05 [redacted].mkv
2. DARLING in the FRANXX - S01E07 - Shooting Star Moratorium [redacted].mkv

I think the TMBD provider may not be able to pick up the metadata properly based on the naming of the episodes.

RE: Jellyfin Instability and Crashes - Jam - 2024-07-22

I was on 10.9.6 prior to 10.9.7, every version I have updated to every version through the lifespan of 10.9.

It seems the reason it's not picking it up is due to image content is not present on TMDB though? It doesn't seem to be failing to find it, more so the images just aren't there that it's wanting.

I've also ran the SIMKL plugin for quite a long time now, since the life of my Jellyfin server so i'm not quite sure that'd be halting it up, if at all.

Like I say, my servers been running fine for a very long time, it's just as of recent which seems odd. Like right now, I have 10 friends watching content on my server and it's handling perfectly, no issues. It'll just gradually degrade throughout the uptime then seize up.

I am still happy to give it a go disabling it but I am not confident it'll change much, if at all. And the scan causing a couple errors early hours of the morning doesn't add up to when the issues are occuring.

RE: Jellyfin Instability and Crashes - Jam - 2024-07-23

Got another log capture. You can see at 20:56:38 I had to restart the server due to it seizing it up. This was without the SIMKL plugin and having TVDB and Fanart as backup metadata/image providers.

https://gist.githubusercontent.com/JamsRepos/c3215285a8dcf4bc9e1c59d1793668d0/raw/11787f2e06fd9b7e147ec8a9ce13cd2755c53446/gistfile1.txt

[Image: VLFNb78.png]

Here's also the containers usage on the server at the time.

RE: Jellyfin Instability and Crashes - TheDreadPirate - 2024-07-23

I see the logs, but its not clear what is causing it.

I saw your comment on a github thread where you said you were on Unraid. Can you go to the Unraid console to see if there are any logs that could point to something Unraid is doing in the background that could be the root cause?

I have Unraid in a separate test setup and I'm still familiarizing myself with it.

RE: Jellyfin Instability and Crashes - Jam - 2024-07-23

(2024-07-23, 08:25 PM)TheDreadPirate Wrote: I see the logs, but its not clear what is causing it.

I saw your comment on a github thread where you said you were on Unraid. Can you go to the Unraid console to see if there are any logs that could point to something Unraid is doing in the background that could be the root cause?

I have Unraid in a separate test setup and I'm still familiarizing myself with it.

Nothing was running at the time of it happening. System on glances also looked fine, nothing out the ordinary. IO was also fine.

https://gist.githubusercontent.com/JamsRepos/e73fb01134dc43322779e10966d4d949/raw/2623834679e17c6fb7200ebdf38a4f572a400b30/gistfile1.txt

That's the Unraid logs, the 100.100.1.10 is my Tailscale to my VPS which stops my server IP being exposed.

It's also worth noting this occurs to the local IP address of the server too, so that's not associated to the reverse proxy connection.

RE: Jellyfin Instability and Crashes - TheDreadPirate - 2024-07-23

Quote:Jul 23 20:56:35 Home kernel: br-094fb560e9be: port 2(vethf8ad51c) entered disabled state
Jul 23 20:56:35 Home kernel: veth087d5af: renamed from eth0
Jul 23 20:56:35 Home kernel: br-094fb560e9be: port 2(vethf8ad51c) entered disabled state
Jul 23 20:56:35 Home kernel: device vethf8ad51c left promiscuous mode
Jul 23 20:56:35 Home kernel: br-094fb560e9be: port 2(vethf8ad51c) entered disabled state
Jul 23 20:56:35 Home kernel: br-094fb560e9be: port 2(vethdfe6ef4) entered blocking state
Jul 23 20:56:35 Home kernel: br-094fb560e9be: port 2(vethdfe6ef4) entered disabled state
Jul 23 20:56:35 Home kernel: device vethdfe6ef4 entered promiscuous mode
Jul 23 20:56:35 Home kernel: br-094fb560e9be: port 2(vethdfe6ef4) entered blocking state
Jul 23 20:56:35 Home kernel: br-094fb560e9be: port 2(vethdfe6ef4) entered forwarding state
Jul 23 20:56:36 Home kernel: br-094fb560e9be: port 2(vethdfe6ef4) entered disabled state
Jul 23 20:56:37 Home kernel: eth0: renamed from veth9dfe8aa
Jul 23 20:56:37 Home kernel: IPv6: ADDRCONF(NETDEV_CHANGE): vethdfe6ef4: link becomes ready
Jul 23 20:56:37 Home kernel: br-094fb560e9be: port 2(vethdfe6ef4) entered blocking state
Jul 23 20:56:37 Home kernel: br-094fb560e9be: port 2(vethdfe6ef4) entered forwarding state

The time stamps for this message lines up with the last crash/restart. And nearly every time these messages appear correlates to the Jellyfin container restarting seconds later earlier in the day.

Jellyfin didn't restart every time these messages appeared. But every time Jellyfin restarted these messages happened seconds earlier.

RE: Jellyfin Instability and Crashes - Jam - 2024-07-23

Every restart was triggered by myself manually. So whether that's just unraid taking it's time to trigger the restart as when I click restart on Unraid, it's never instantaneous. It often takes a while for the container to actually receive any type of shutdown message. Which I would assume is that?

I've looked up the logs and they do seem to be related to when you restart a container, so I would assume that's the actual timestamp of when I clicked the restart button on Unraid.

It's also worthwhile noting I have also been restarting other containers today, so that could be also another reason why there's more even when Jellyfin does not restart.