alonm15

I believe i found the issue. It seemed to happen when plex would hammer the CPU during transcoding. I didn't know this was happening since the server would crash multiple times and clear out the log (Im sure i'm doing something wrong with the logs.) But it ended up being an apparently known issue with Linux i915 kernel drivers and transcoding. Froze up the CPU. hard system crash. I have many remote plex users so I'm not always sure when it is in use. I finally captured a log file when it crashed and could see the mover moving transcoder files which gave me the clue. Hopefully this puts a bow on it. Thank you to everyone for the help.

the server started crashing repeatedly on Sunday night. I had to leave town for work so I kicked off a memtest and it ran for 8 hours with no errors and passed. Is it possible RAM is potentially an issue or should I move on to testing other hardware components? PSU is also brand new.

so just like this?

Another crash today. Though for whatever reason, the previous sys log did not show anything from prior to the actual crash. It just shows it starting up. So I have no idea if there were more nginx errors or not. Double checked my Syslog server and it seems setup correctly but still no logs from prior to the crash. just the crash itself. The strange part is, the server comes RIGHT back online. I don't have to go power it up. It crashes, then automatically restarts itself which seems to rule out a hardware issue. The PSU is now brand new and still the same issue.

Got it. Is there a way to increase the available memory?

tower-diagnostics-20260215-1127.zipHere it is!

Yeah sorry, I don't have the diags from before it crashed. I'll upload again if it happens...

syslog.zip

Server completely froze tonight and had to be force shut down and rebooted. I do not understand what the hell is going on. This is becoming infuriating. Nothing is working. Nothing is helping. This feels like it has become an enormous waste of time and money. How is there not a single log file that shows anything helpful. How is Unraid not providing technical support. What is going on here. This exact same set of hardware has worked flawlessly for years and the minute I convert to unraid, not a single thing goes right. I am beyond frustrated. I appreciate the help of the community it trying to solve this but how the hell can this be a viable product when I can't keep a sever up for more than 48 hours.

Rebooted in safe mode and so far no restarts but I am now getting the same nginx error still. Error occurred over 16,000 times all with the same timestamp Feb 14 00:22:53 Tower nginx: 2026/02/14 00:22:53 [error] 4653#4653: MEMSTORE:01: force-reaping msg with refcount 1

I am getting this exact same behavior and my machine has been in safe mode for about a day now. error below repeated over 16,000 times all with the same time stamp. [error] 4653#4653: MEMSTORE:01: force-reaping msg with refcount 1

Should I reboot and leave it in safe mode for a few days and see if it crashes? Or did you mean retest in safe mode due to the nginx errors

syslog-previous.zip I haven't done a mem test but can do that. It crashed again at some point last night/this morning. Any chance these logs have anything useful? If not I'll start swapping out hardware.

It seems to also come right back up. So it crashes, but then the server reboots itself. Does that point to anything in particular?