Hi Guys,

 

I keep getting an "Out Of Memory errors detected on your server" error on my server every single day for the past week.

 

The weird thing is that it happens 4:46 AM every time, only thing I could correlate this with is the following:

 

- Automatic appdata backup (starts at 4AM)

 

I did have problems with a kernel panic crash but that hasn't happened in a month or two. That was also the reason I upgraded to 6.10.0-rc2 but that didn't fix it.

 

How can I check where the problem lies?

 

zeus-diagnostics-20220306-1852.zip

syslog-10.50.0.254.log

It's not related to the backup.  It's happening daily at ~11pm What else goes on at that time?  (I'd guess something with tdarr node?)

I haven't run tdarr nodes for a few months now, at 12 am there is a ssd trim, mover works every 3 hours. Everything major should start after 2 am.

 

But if it happens at 11pm maybe Plex is the problem?

I think I am using Ram as storage for plex transcoding:

 

--runtime=nvidia --device=/dev/dri --mount type=tmpfs,destination=/tmp,tmpfs-size=20000000000 --no-healthcheck --restart unless-stopped --log-opt max-size=50m

 

But the error is everyday and my watch time and error time doesn't match up

I woke up today the moment I got the warning and checked dashboard but the ram was still fine (80% unused), I checked my server health event logs via IPMI but there is nothing there (about ram and recent) so probably not a hardware fault either right?

I've got grafana installed with prometheus but I can't see anything out of the ordinary:

 

 

Does anyone have an explanation for this?