android tv client blocks others

[snake]

Jellyfin version: 10.10.3
client used: Android TV 0.18.2
installation method and your platform: kubernetes cluster, official jellyfin image deployed with an nfs v4 volume mount (M.2 SSD disk) for the appdata part i.e. the folders present in /config and same for the /media hormie disk part (UNraid HDD storage)
CPU: AMD Ryzen 7 5700X 8-Core @ 3400 MHz
RAM: 16G

When one of my users connects with an android tv client to play media, (which it manages to play correctly) this considerably increases the response time of my jellyfin application when browsing the different menus & libraries and blocks media playback in the sense that the application loads ad infinitum.

This only happens when using the android tv client, I've already tested with 4 media playing at the same time via browsers + android mobile client and my infrastructure manages to deliver the videos fine.

However, when a single android tv client connects and links a video, everything is congested.

This is not a transcoding problem because I tested with media that doesn't require transcoding, so ffmpeg is out of the question.

I monitor my infrastructure and I have good quality components & networks, I've done several tests to test all the hardware, it does not come from this side, so I lean for the application part, I do not know what the android TV application does more than the others but it poses a problem for me to use.

Could you suggest a solution?

[TheDreadPirate]

Can you share your full jellyfin log via pastebin?

[snake]

Can you share your full jellyfin log via pastebin?

Shure , this is the log of yesterday , i saw the probleme at 15h30 but maybe its begening befor.

https://pastebin.com/qVLLTKgx

[TheDreadPirate]

I don't see anything in the logs that would explain what is happening. It is direct playing like everything else.

I re-read your original post and are you saying that the jellyfin appdata is on a NFS share? As in the container has to access /config over NFS?

[snake]

I don't see anything in the logs that would explain what is happening.  It is direct playing like everything else.

I re-read your original post and are you saying that the jellyfin appdata is on a NFS share?  As in the container has to access /config over NFS?

exactly , this is the mount point : 

Code:

toto.lan:/mnt/user/appdata/jellyfin on /var/lib/kubelet/pods/187d3e90-fcd9-4355-bd99-9/volumes/kubernetes.io~nfs/appdata-jellyfin-nfs type nfs4 (rw,relatime,vers=4.2,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,clientaddr=192.168.5.99,local_lock=none,addr=192.168.5.98)

mapping the appdata in my NAS to the /config into container managed by pv & pvc kubernetes 

and looking this page https://jellyfin.org/docs/general/admini...n/storage/ preconize nfsv 4 for the locking file, but if it's nfs editing with a file lock I would have had the problem during my playback test with more than 4 videos simultaneously on the web client + android mobile app.

[TheDreadPirate]

SQLite should never, ever, ever be run over a network share. This can cause many issues.

Please relocate the appdata to a local disk, preferably a SSD, and try to reproduce the problem.

[snake]

Hello, first of all thank you for your quick response, and valuable advice. I'd like to give you an update on the situation. Indeed, after relocating the data (/conf) part of my server to a local filesystem, I no longer have any latency or other problems with video playback, and my display/browsing is so much smoother!



Thanks again, I now think I'll opt for a backup solution for the data part via rclone as suggested in the doc.