Jellyfin (docker) froze with 100% cpu and 100% ram usage.

[Istria]

Yesterday (27-Aug), I noticed around 1830 that my server was running at much higher cpu load than usual. After investigation, I noticed that the proxmox (LXC) container running jellyfin (in docker) had been at 100% cpu (of the 2 assigned cores) and 100% RAM usage since the day before (26Aug). Also proxmox graph says it's been reading from disk at around 500MB/s the whole time, which is impossible because my SSD isn't that fast. So maybe it is reading from RAM at that speed? It was also unresponsive. At that time, Jellyfin was not being used. The server was simply running idle. 

Rebooting the container solved the problem. But I'm wondering how I could figure out why it happened and if it could happen again. 

I attached the logs of the past 3 days and some screenshots of Proxmox.

Has anyone experienced this? Or has any hints of where to look for the problem? 

Thanks in advance!

[Venson]

Hey @Istria

The logs you uploaded do not show the server running on the 27.08 around 1830. Only a restart at 18:44 where the Scheduled tasks where triggered.

Did you recently Added a bunch of new Media?

[Istria]

Yes, that is the weird thing. It was running though. Just in that frozen 100% cpu and RAM mode for a long time.

If 2-3 weeks ago counts as "recently" then yes. I set the Jellyfin server up 2 or 3 weeks ago and added all my media then (around 3TB).

[Venson]

Well without the logs from that particular moment, i really cannot say for sure what it was. However 1gb Ram does definitely not sound sufficient and 2 cores seem a bit lacking aswell. Did you setup Hardware acceleration? My best guess would be that the server started to page RAM onto disk like hell and then more or less bricked your instance.

[Istria]

I don't transcode. Only direct play or direct stream.
I have 1GB (not 2) ram allocated to the container. But it rarely uses it all. Now, for example, it uses 450MB idle. Playing a 1080p movie makes it go to 640MB.
Could that be why it froze? That for some reason it started doing something while idle, and ran out of RAM?

EDIT: I just noticed I had no swap space configured for it. I just added 1GB of swap space just in case.