Whole server keeps crashing and have to power-cycle

Server Version#: 1.16.3.1402

Hello. Something has been happening with my linux headless server I’ve got running Plex on and virtually nothing else. It is running Ubuntu 18.04 LTS and is regularly updated.

I know the current server version needs an update and I will do it right now but this problem has spanned lots of updates over the last year or so. Basically, a few times a week the entire server locks up and becomes unresponsive. It has only happened a few times while watching Plex, but it did happen in the middle of watching something the other day (should be the most recent set of logs, from 10 August 2019). The entire server becomes unresponsive, I can’t ssh in, no Remote Desktop, no Plex, no ping. I have to power cycle the server.

It’s gotten to the point where I have the server plugged into a smart outlet so I can power cycle remotely since it’s tucked away in a corner of my basement.

I’ve looked at the logs but I’m not versed enough to make sense of them unfortunately. Hoping someone here can help me figure this out. My wife is tired of having to ask me to reboot the server when I’m not home remotely so she can watch something! Thanks in advance.

Logs from the two most recent crashes attached.

Plex Media Server Logs_2019-08-05_19-10-42.zip (2.5 MB) Plex Media Server Logs_2019-08-10_11-43-00.zip (2.9 MB)

Please open the computer covers and check the cooling. Look for dust bunnies. Having a machine lock up when Plex is stressing the CPU is a sign of something overheating… Usually the processor.

Popped open the case. It’s a Cooler Master HAF+ with 4 fans - My cable management isn’t perfect but I don’t think it’s impeding anything. Seems pretty clean in there.

If the fans are all operational, and airflow is good, I’d bet on hardware, starting with the power supply, then memory.

Would any of the machines system logs confirm that? Should I look for a hardware diagnostic program? Just don’t want to replace one or both of these on the hope of I can narrow down the problem. I might have a functioning PSU from an older build that I gutted a few years ago that might work…

That’s going to depend on the real failure, and logs, if local, depend on a fair amount of OS activity before they get written out to disk (remote logs have a better shot in some circumstances.)

Might be a good idea to run memtest from OS install media or a thumbdrive. You will have to boot from the media where memtest is installed, hence PMS won’t be available for this extended test.

If you have a thermal monitor, might also be beneficial to see if there is perhaps any kind of latent irregularity at the processor.

I’ll see what I can figure out - so do my Plex server logs look normal then?

Yes, other than abruptly halting, they look fine.

Well shoot! Here I was blaming the server and maybe it’s a hardware problem this whole time. I’ll see what I can figure out, thanks for all your help!

Sometimes, yes it’s just a burned out light bulb :joy:

In my case, the burned out bulb is usually behind the keyboard…

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.