I'm currently on Pass 3 of Memtest without any errors.

Planning to go for the rest of the day.

10 passes, 18+ hrs and no errors.

Definitely looks like its not a memory issue.

I've replaced the flash drive, and I've disabled C states globally.

 

I'm pretty sure this is a software issue, whether thats something with my dockers, or something Ive configured in Unraid.

Unlikely that a hardware problem would cause the server to always crash at the same time.

On 1/1/2024 at 3:54 AM, JorgeB said:

Unlikely that a hardware problem would cause the server to always crash at the same time.

Agreed, I may have been too hasty in assuming it was a crash.

It happened again this morning and I just ignored it, left it to go to work and noticed it was online without requiring me to force shutdown.

 

Reviewed the logs (mirrored to flash) and I see a massive wall of stuff happening around 5-6hrs after it went offline.

I've attached the log, can anyone explain what I'm seeing there at line 2544, at Jan  1 09:13:30?

Update:

 

Well damn! Just noticed I have scheduled tasks to begin in Plex at 3AM, this is when my issues are popping up around 3:15am!!!

Referencing the syslog I can see an out of memory warning... Not sure where things went wrong but I must've accidentally mis-configured something.

Though I haven't touched my Plex settings in several months as far as I recall.
 

Any tips on what to check for these memory issues with Plex?
(Reminder, my RAM passed an exhaustive memtest.)

 

Update2:

 

My working theory is my Plex db may have corruption.

Running an integrity check on it now.

Following this guide.
Not sure how long it will take, need to sleep, will check back in the morning.

 

Edited January 6, 20242 yr by techlovin

This is absolutely an issue with my Plex docker.

 

Memory keeps ballooning during PlexScan and locking up the server until the process is killed.

Issues with my Plex db caused the container to consume all available RAM, eventually grinding my server to a complete halt killing off most functionality.

After several hours Unraid would kill the process consuming all available memory and service would be restored.

 

No issues with the Plex scheduled tasks turned off.

 

I noticed some Plex log entries about this that made me believe the scanner was getting hung up trying to match some of the files.

I deleted my hentai library which was using the hama.bundle plugin.

 

Turned the tasks back on yesterday after deleting a few libraries, and did not notice the memory issue or my server locking up.

 

Will monitor for a couple days and If no more issues crop up I'll mark this as resolved.