Excessive RAM consumption

[TheDreadPirate]

How much RAM is Jellyfin using? AFAIK, the malloc fix is in LSIO's image. For comparison, mine grows to ~2.5GB RAM and doesn't really grow beyond that.

Excessive ffmpeg RAM usage is transient and only happens during initial library scans, which doesn't sound like that is happening here.

[Petitgnoll6]

It's around 4gb RAM. And when I restart the container, it goes back to 440mb

And yes, I think I remember that the initial scan was really resource impactful in ram but it's not the issue here, it's the RAM being eat up slowly over time (but I don't see impact on the server' performance, so I guess it's ram caching ?).

[TheDreadPirate]

Jellyfin does aggressively cache.

[Breakintr]

I might have the same or at least a similar issue. I have jellyfin running in proxmox/LXC, and looking at the system stats, since March the memory usage keeps increasing until it leaves the container unresponsive and I have to restart it.

   

   

I'm trying to disable some tasks to see if it helps, but so far it keeps happening.

   

Can't see anything on logs either, do you have any tip on how I can debug this further?

Edit: Just found out this, I'm giving it a go and see if it fixes
https://github.com/jellyfin/jellyfin/iss...1774082390
https://github.com/dlemstra/Magick.NET/i...-785351620

[TheDreadPirate]

Within the LXC are you using any additional virtualization like docker? Or did you use the tteck scripts to install Jellyfin?

Regardless, 2GB for the LXC is not enough. Our documented recommended minimum is 8GB of RAM, but you could probably get away with 4GB.

https://jellyfin.org/docs/general/admini...memory-ram

During media scanning when you add new media jellyfin spawns a bunch of ffmpeg and ffprobe processes that consume memory.

[Breakintr]

I used tteck scripts. I'll increase to 4Gb, I have enough to provision it. I just find odd that it was working fine with 2Gb since I've installed in August until March this year, looking at the releases, the only one that I saw that kinda matches it was jellyfin-ffmpeg v6.0.1-4.

Anyway, I'll increase to 4Gb or even 8Gb and see if eventually memory consumption stops increases or if still creeps up until it gets out of memory. From what I could see, 2Gb would take around 1 week, so I'll get back in a few weeks.

Thanks!

[TheDreadPirate]

My instance levels out at around 2.5GB and stays there. During big library scans where I add a bunch of new media it climbs to 4-6GB from the ffmpeg processes it spawns, but Jellyfin itself stays at 2.5GB.

[34626]

My Jellyfin uses 9.5 GB of 32 GB

[SyrupSplashin]

Most of the memory jellyfin uses is cache.  Memory used for caching and reserved memory for applications and application data behave significantly differently.  If another app needs more memory, the cached memory will be immediately released to the other application.  Most  dashboards, like the screenshot you provided, do not distinguish between reserved memory and cache.

I'm running into this same situation where Jellyfin holds onto cache indefinitely; bare metal this is of little concern but I'm running it in proxmox and want that memory freed up for other VMs.

Assuming it's expected for jellyfin to hold data in cache/buffer indefinitely, how can I change this behavior? If it's not expected, what could be the issue? I'm running Jellyfin v10.8.13.0 pulled from Ubuntu's repos (i.e. not a docker image). Github says the env variables to help with memory consumption should be within this version so I don't know what I can really do here.

[Breakintr]

It looks like whatever it was causing this, on my end it seem that was fixed on v10.9.0