Plex Media Server crashing during scheduled maintenance

I noticed my Linux server (Intel nuc with Debian) had been hard crashing around 2am for the past few nights. This is when my Plex scheduled tasks are set to to run. I confirmed this was the cause by adjusting the time and waiting, sure enough…hard crash and had to do a turn it off and on again reset.

I downgraded to an earlier version of Plex, this resolved the problem, I will wait for an updated version before upgrading.

I moved your post to a new thread. Your problem is most likely different, as those systems were not crashing. Also, the thread was in the Windows, not Linux, section of the forum.

If you want to troubleshoot the problem, return to the troublesome version of Plex Media Server. Then, please take the following steps:

  1. Return your system to the correct date & time.

  2. Ensure Plex Media Server is configured for debug, not verbose, level log files.
    Settings → Server_Name → General + Show Advanced

  3. If you changed the logging level, restart Plex Media Server.

  4. After the next crash, restart Plex Media Server

  5. Wait 3 - 5 minutes for Plex to fully start and the log files capture the startup sequence.

  6. Retrieve the log files.
    Settings → Troubleshooting

  7. Upload the ZIP file to the thread.
    Drag the file into the message window. It will insert at the cursor location.
    Alternately, use the upload button (with the arrow).

1 Like

Might be related to the new Credits detection feature which seems to be what is causing my OMV (Debian Linux) server to hard crash. I disabled that feature and that solved the problem for me. Downgrading also solved it.

Some recent versions of Plex Media Server have also had problems with loudness and sonic feature analysis. Log files should show what is occurring when the system crashes.

1 Like

@FordGuy61 Thanks for moving this to it’s own thread!!

Alrighty, I followed your instructions:

  1. Updated to latest version
  2. Kicked off scheduled tasks
  3. Server died after about 2 minutes

I’d be happy to upload specific logs, or portions of specific logs if it helps plex devs debug (I’d prefer not to upload everything). Can you let me know what would be most helpful?

Thanks!

I already know that enabling credits detection is what causes my server to crash. I’ll wait for the next release before enabling it again.

1 Like

Plex Media Server.log through Plex Media Server.5.log

If you’re worried about personal info, those files may contain your e-mail address. You can mask it easily via search/replace in a text editor. Please do not change any other information in the files.

AFAIK they are the only log files that contain such info.

It’s 100% credit detection. Turning that on == crash, turning it off != crash. I scoped the logs and I don’t see anything obvious. There is enough potential PII in there that I do not feel comfortable posting to a public forum. Would be happy to send directly to any plex devs if that is useful.

Outside of that, I will just wait for an updated version.

Thanks all for helping me verify it was plex and not my hardware or linux install!

1 Like

Same problem. In version 1.30.2 everything works fine, in the latest 1.31.0 my Intel NUC hangs when doing a scheduled maintenance.

Disabling Credits Detection solves the problem.

I had a similar problem years ago, and it was because I didn’t have graphics acceleration enabled inside the plex container.

It seems that until the issue is fixed we cannot use Credits Detection.

1 Like

So I have a new hypothesis. . .

I think this isn’t caused by a bug with Plex, but is caused by high CPU utilization. The plex credit detection happens to use a lot of CPU and thus causes the problem.

I think the above is true because I tried running geekbench 6, and my nuc crashed when CPU hit 2.88

If true, kind of a bummer :frowning: Below screenshot is when the nuc crashed.

Memtest86+ and Prime95 stress test time – AFTER you open it up and blow all the dust bunnies out. Make certain to gently blow out the cooling fan (don’t overspin the fan bearings)

I opened it, there were no dust bunnies. I did blow some air in there just to be sure. I also adjusted the fan setting in the bios to “Cool” or whatever the loudest option was :slight_smile:

  1. Ran geekbench again, crashed around the same CPU load (~2.8)
  2. Ran prime95, did not crash, CPU load was up around 3.99

I suspect its the GPU, this tracks with the issue I’ve seen with plex and with prime95 (i think) not hitttng the GPU, but geekbench does.

@ChuckPa I’d be curious if geekbench crashes your machine… or not!

Which model NUC is this? I have 3 different models.

Also wondering about memtest86+ , extended tests, extended duration

NUC7i7BNH (Intel Core i7) Circa 2017

I will give memtest86+ a try next.

Thanks for helping !!

Before you memtest it.

  1. Take it apart again.
  2. Unclip the memory and take it out.
  3. Clean the edges of memory chip with alcohol and q-tip (they do tend to oxidize)
  4. Put it back in.

I have an i3-BNK model (same form factor). I also just cleaned it

This might still be PMS but if other tests are failing then let’s get the unit solid again

Oddly memtest is not running via the grub menu. I’ll have to see if i can sort that. Note I have never run memtest so I do not know if it not running is related or not.

I select memtest in the grub menu and it just shows the debian grub background.

I still think it has to be the GPU or the GPU overheating. The only time this happens is when the GPU is taxed.

That doesn’t work for me either.

I spin a dedicated USB of it

Alrighty. Memtest reported 0 errors.

I tried one thing in the bios, changing from performance mode to balanced. No change, geekbench crashed pretty much where it did last time.

For now I will keep plex credit detection off, and if the machine winds up crashing during normal operation (it has not thus far) I will just move everything to my m1 mac mini.

I’ve got the same issue but I found that my system was doing a hard crash when it encountered two bad video files. It looks like it handles one bad file ok but it goes unresponsive after the second video file.

I’ve removed out both files from my PMS and will see what it does tonight.

Can confirm that the hard crashing has only occurred recently and I’m running the same NUC with Debian.

How long did you let it run? Single pass?

I normally set it to full suite and run continuously until stopped.
If it runs 3-4 hours without a glitch , I’m certain memory is ok… Your memory is probably good too.

Yes, Keep Credit detection off for now. It’s the only recent significant change.