Michael_P

FWIW- I had a case with a flaky power button, my windows media server in a closet would randomly shutdown. Took me months to figure out what was causing it.

Looks more like you have a tab open to your dashboard for too long?

Shutdowns under load sounds like a flaky power supply

The OOM event happened on NOV 05 and looks like it may have been HA/Homematic related, and it likely has nothing to do with your other system instability. You'll need to reboot to clear the log or FCP will continue to alert on it. For random shutdowns, the first check for me is always the power supply.

And it was back on the 30th - once you figure out what it was, you can restart the server to clear the log so FCP stops warning you about it.

Something is spawning a LOT of Chrome processes, so if you have a proxy container or the like I'd check there

You have to take into account the age of the drive, too - and if another drive fails you will be relying this dodgy drive to rebuild it, a process which puts the drive under the most stress. The drive is telling you ahead of time, I'd listen to it.

It's lived a good life, time to replace.

It's killing python3 in your other logs and I didn't see anything but what I listed as running a python3 process, you can try running without HA for a bit and see what happens, then keep moving down the list until you find it - but I'd start with frigate as that seems to be the culprit in a lot of cases I've seen

The reaper is consistently killing off python3 because it's got the most memory usage, and there's a lot of nginx processes, too - so I'm looking at whatever's spawning those (and frigate is always a suspect anyway) The containers I see spawning those are busybox, whatever this is 6ba7fd76579f9d0638ab122feba7c103e977ac04b07fe58fcfaeab655fe0b13f (using fail2ban?), netalertx, whatever this is 1d1c308d88054708609009b1a4ab5c02180e54ad530e239c4f418fff5e8138fd, and redis (are you using immich?)

Next guess would be frigate and busybox

You've got a lot of stuff running on there, but it looks like mosquitto or openeats is spawning a bunch of threads and python3 instances. Start by disabling all of your containers then re-enabling them one by one until you find the culprit Also, your VPN credentials are exposed in your diagnostics, you should remove it from your post @JorgeB

And it happened only the once back on the 23rd , so follow Jorge's advice if it happens again - in the meantime, you'll need to reboot to clear the log or FCP will continue to warn you.

Looks like Plex's scheduled maintenance choked on something back on the 12th - you'll need to reboot to clear the log so FCP stops warning you. If it happens again, limit the memory the Plex container is allowed to use or figure out which file it doesn't like.

Looks like any container that passes it as a variable - in this case their deluge container

Then its not being respected - you'll need to fix the containers config so that any temp space for encoding and such isn't being done in RAM or you can force docker to limit the RAM instead of the kernel with the --oom-kill-disable flag, but use it cautiously as it's highly likely to bring down the entire system if docker doesn't do it right. https://docs.docker.com/engine/containers/resource_constraints/ When you've fixed it, you'll need to reboot so FCP doesn't see it in the log anymore and continue to warn you

Jellyfin or Immich would be my guess, so start with those. In the meantime, your VPN credentials are exposed in the diagnostics file, you should remove it from your post and change your username and password @JorgeB

Looks like Frigate is spawning a bunch of ffmpeg processes and running the host OOM repeatedly - fix the config so it's not writing to RAM or limit the container's memory allowance

I don't get it, either - sorry I couldn't help and good luck!

Also make sure network isolation is not turned on in the router

Can your unraid box ping any of the clients?

And your clients are all pulling IPs on the 10.10.1 subnet?

Nooo, not for local. What subnet are your clients on?

Just a WAG, but try disabling Armor on your router And I'm assuming you've tried pinging the server from your clients?

In Plex- go to settings, scheduled tasks