Over the past few months I've noticed that memory usage of my server would slowly increase over time. The culprit appears to be Plex as restarting it will lower the usage from 60-70% usage to 13%, but I'm not sure what specifically is causing the memory usage to expand. 

I only recently noticed that Fix Common Problems flagged the memory leak. 

 

Error

Out Of Memory errors detected on your server

Your server has run out of memory, and processes (potentially required) are being killed off. You should post your diagnostics and ask for assistance on the unRaid forums

 

I've attached my logs. Any help would be greatly appreciated. 

tower-diagnostics-20210405-1248.zip

The Unraid OS is in RAM, so any path that isn't actual storage is RAM. Typically storage is some path in some subfolder of /mnt (but not /mnt itself).

 

The usual reason for a docker to write into RAM is a container path that is mapped to a host path that isn't actual storage.

It's been a while since I set this up. My current container path is /mnt/user/appdata/ 

Is there a preferred location for me to remap this to? 

 

That looks OK. Plex isn't the most likely candidate for writing much data anyway except for transcoding and if you use DVR feature.

 

Is that your only docker?

Here are all my dockers. 

 

If it's not Plex, it's weird that the memory goes down once I restart it. 

Nothing obvious in the screenshot.

 

Apr  3 22:05:49 Tower kernel: Out of memory: Killed process 13624 (Plex DLNA Serve) total-vm:14714740kB, anon-rss:11700980kB, file-rss:0kB, shmem-rss:0kB, UID:99 pgtables:23460kB oom_score_adj:0
Apr  3 22:05:49 Tower kernel: oom_reaper: reaped process 13624 (Plex DLNA Serve), now anon-rss:0kB, file-rss:0kB, shmem-rss:4kB

Do you actually use the DLNA feature of Plex?

I don't. I can disable it and see if that fixes it. 

I disabled it yesterday and so far the memory usage has only gone up by 3%. It jumps randomly so I'll keep monitoring and update this thread by the end of the week. 

 

Thanks again! 

Thanks @trurl. That appears to have fixed it. 

 

Most days my memory usage sits at 18% with sometimes going up to 30%, but that may have been during a parity check or multiple transfers going on.