So on my backup server, I have a Lincstation S1 that came with 8gb DDR5 Sodimm. Due to the RAM sitatuation, I dont really feel like paying for a larger stick of ram. I'm running nextcloud (+mariadb) and immich (+redis and postgresql). I keep getting OOM errors nightly, and the logs state it's immich. I have immich limited to 2gb of memory using this extra parameter "--memory=2G". Normally immich is using about 1.6G memory, and the server has about 1.2G free when I look. So I'm not sure how it's getting to be OOM.

I was looking at some other posts and it seemed to imply that the server will restart the process if it gets to the memory limit of the container? So did I actually make the problem worse by making the memory limit lower than the available left on the server?

server-ak-21-diagnostics-20260803-1848.zip

I can't answer your question specifically but I solved my intermittent OOM errors with the SwapFile plugin (in community apps.) Haven't had one since.

3 hours ago, nerbonne said:

I was looking at some other posts and it seemed to imply that the server will restart the process if it gets to the memory limit of the container?

Correct, if the container hits the limit you've set then it's the one that will be killed for being out of memory. So maybe it might not have been killed if you had set more, but then setting it low enough ensures that no more important stuff gets killed.

If you have installed the immich machine learning stuff you probably want to remove that, too tight on 8GB. Run that on another machine with more resources.

Edited 41 minutes ago41 min by Kilrah