Ubuntu Server Keeps Locking UP

Server Version#: 1.41.5.9522
Player Version#: 4.145.1
Distributor ID: Ubuntu
Description: Ubuntu 24.04.2 LTS
Release: 24.04
Codename: noble
Kernel: 6.8.0-55-generic

For the last week my server seems to randomly “crash” where I hear the fans running but I can’t SSH into it or get it to wake on the local screen. I have to power it off and back on and have tried checking all the server logs/messages and plex logs and I can’t find anything that’s showing anything obvious. I tried rollbacking to 1.41.4.9463 and still seem to get a random “crash” or lock up. It looks like it may be mostly happening when an Apple TV client is trying to stream remotely. I can stream locally no problem from Roku and Android devices. But when an Apple TV starts playing it only makes it about 4-5 minutes and then just locks up. Not sure if this could be a hardware issue locally or a bug somewhere. I have tried changing around the transcoding settings and such with no luck. I believe it’s been uploading crash logs but not sure if those are showing anything useful or not.

Anyone have any other ideas? Should I just try reloading the entire server?

When it locks up, can you please

  1. Stop Plex
  2. Manually make a tar.gz of the “Logs” directory
  3. Attach that tar.gz here

I’ll look at it

ChuckPa,

Thanks for the response but when it happens I can’t get into the server via SSH or even local display/keyboard, it’s completely locked up. Which honestly has me wondering about a hardware issue but I can’t find any kernel panics or anything that is telling me anything. But every time it happens in looking at the logs it seems to be when an Apple TV is streaming something so makes me think it’s something related to that in Plex.

Get into the command line.

see how much free space you have df -h

See how full the / (root) partition is.

I have a sneaky feeling it’s filling the root partition

If so, we can resolve it by moving where PMS is stored.

Okay so just got the server to lock up on me while SSH’d into it already watching TOP and trying to play a couple of videos, transcoding was performing horribly and kept buffering and then all of a sudden my SSH/top locked up and the server is unresponsive now

Wow.

Sounds very much like hardware

you need to look through all the system logs (dmesg & /var/log/messages*) to see what’s going on

This is nearly impossible to diagnose without seeing something

Yeah was worried about that…This is a 16 core rackmount that I’ve been running for about 2 years now with no issues and then this just started in the last 2 weeks. If no transcoding is running the server runs fine, but just now it was when it was doing transcoding that it freaked out.

Last line of syslog before the server was rebooted was this:

2025-03-26T13:55:46.728921-05:00 dotmatrix Plex Media Server[9934]: Dolby, Dolby Digital, Dolby Digital Plus, Dolby TrueHD and the double D symbol are trademarks of Dolby Laboratories.

Can’t find anything obvious in kern.log, dmesg, journal :frowning:

Maybe I will try turning on verbose logging on Plex to see if it spits out anything?

Verbose logging won’t be helpful. It’ll spew tons of junk to you and cover about 2 minutes of elapsed time.

Stupid question time:

Dust Bunnies ?

I ask because when transcoding, there is increased heat generated (significant) compared to DirectPlay (which is a read-send operation)

Can you setup temp monitors and watch it?

ah okay yeah went through the verbose and didn’t get much…yeah wondering about dust too. Just made a SystemRescue Boot USB to test on it too but will pop it apart and make sure to clean it out

well probably not a good sign that the server is crashing on memtest running about halfway through…ugh, looks like I need to move to different hardware. Thanks for looking.

This very much smells like a hardware problem. I would begin by reseating the power, disk, and RAM connections, and crossing your fingers.

If you are not lucky and there is a hardware fault, in my experience, the power supply is the most common culprit. However it is hard to test unless you have a spare, but if your RAM is a pair of sticks, you can pull them one at a time to see if one is bad.

A hardware fault is a real PITA to figure out but you’ll get there with some patience.

Not really, memory is cheap, unslot half your ram, test again, either remove half or add half back until you isolate the faulty stick

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