So here’s a new dump of logs w/ active recording issues again. The show “Killer Couples” on S10E10 recorded as 12 minutes. S10E09 recoreded for 17 mins, and S10E08 got only 2 minutes…
Latest logs uploaded - this should have the CPU pinning issue fixed as well as the network issue - not sure what else could possibly be going on - thanks in advance if you can assist at all.
Is it possible that if there’s a “blip” in the transmission from the antenna that it would just abruptly stop the recording?
Here’s another example of Good Morning America coming in at 12 minutes again today – i’m at a total loss here.
Any help is appreciated!
@pl_5309 or @FordGuy61 do either of you see anything concerning in these new debug logs?
Plex Media Server Logs_2025-03-12_09-43-40.zip (3.1 MB)
EDIT: 12 mins before the recording ends I see this in the logs
Mar 12, 2025 08:47:25.693 [23092537400120] DEBUG - [Grabber/2296109dfd6eda46f46aa33865253a5780b52d7c] Starting grab immediately for ‘Good Morning America - E81 - Good Morning America’ on channel 5fc76c732f220e002d6d6707-5fc705fba62668002dcefe2b and we’re 6445 seconds in already.
Not sure what to make of this - it was clearly recording GMA this and the final resulting DVR recording is only 12 mins long but it’s not the very tail end of the show either. I’m more confused now than before sadly.
The server is being killed off externally at what looks like a pretty regular interval of every hour.
If you have something that monitors container health or memory I would look there. May want to add LogMemoryUse="1" to your Preferences.xml with vi/nano to see if you are hitting some container limit. This will add memory consumed messages to the Plex Media Server.log files.
What log/timestamp is showing that? I’d like to take a closer look at the time it’s happening to see if I can find anything, but to my knowledge nothing is happening other than a “WatchState” container running on the hour to sync from Plex to Jellyfin but I set that up AFTER all of these issues started incase I had to pivot to Jellyfin for DVR recordings which I don’t really want to do.
Nothing else should be doing anything on the Plex side to cause an issue.
The only thing I can think of is that I’m using a script to set my GPU into a low power state if it’s not currently in use by any apps but that schedule is hourly and wouldn’t explain 12 min recordings where I’m not transcoding during recording.
Anyways, I did setup the LogMemoryUse you suggested.
The container should have over 100GB of RAM available to it as I have 128GB with very little of that being used on a regular basis. I can’t imagine it’s running out of memory but we might as well log it!
Ok I see it happens at around the X:47 mark doesn’t it…?
It doesn’t crash EVERY hour but when it does it’s always around that timestamp? How strange.
I checked the crash uploader folders but there’s nothing in them. I was hoping it would have more information for me to investigate.
Is there anything else I can do try try and track down what could possibly be happening?
EDIT: Typing my reply at the same time. Didn’t see yours until after I submitted mine.
When Plex Media Server starts, it rolls over the log files. It opens a new Plex Media Server.log. The previous Plex Media Server.log becomes .1.log, the previous .1.log becomes .2.log, etc. The log files also roll over when they hit a size limit.
When Plex Media Server starts, you’ll see a log entry of DEBUG - BPQ: [Idle] -> [Starting]. It will be in the top 5 - 7 lines of Plex Media Server.log and .1.log to .5.log.
Your log files show Plex Media Server restarting at 47:09 past each hour.
Plex Media Server.log
Mar 12, 2025 08:47:09.645 [23092863736464] DEBUG - BPQ: [Idle] -> [Starting]
Plex Media Server.2.log
Mar 12, 2025 07:47:09.719 [22772519586448] DEBUG - BPQ: [Idle] -> [Starting]
Plex Media Server.3.log
Mar 12, 2025 06:47:09.735 [23016024099472] DEBUG - BPQ: [Idle] -> [Starting]
Plex Media Server.4.log
Mar 12, 2025 05:47:09.899 [23375614356112] DEBUG - BPQ: [Idle] -> [Starting]
Plex Media Server.5.log
Mar 12, 2025 04:47:09.204 [22478737951376] DEBUG - BPQ: [Idle] -> [Starting]
When PMS restarts, the reason for the restart is usually logged at the end of the following file (because the log files roll when restarted).
So, when Plex Media Server.log shows a restart, the reason is usually listed at the bottom of Plex Media Server.1.log (and so on through .5.log).
However, your log files do not show any reason for the restart (that I can see, anyway). There are no error messages, no “ordered to stop” messages, etc. The log file just ends.
So, the quandary is “Why is Plex Media Server restarting at 47 minutes past every hour?”
It does not appear to be anything internal to Plex.
Might it be something external - an OS setting, an external program running, etc.?
Is that what is causing the recordings to end early? Unknown. However, it is unexplained and should be investigated.
Thank you, I’m going to try to look into this as deeply as I can here.
The “12.x min recordings” would line up with the server restarting at X:47:09 as a few seconds later it starts recording again a I end up with a 12:xx recording at the end. So this certainly could be related.
Now just to find out what the heck is causing it to crash.
If anyone has any insight on this I’d appreciate a point in the right direction - otherwise if I completely fail at this I may consider a wipe/reload of Plex entirely from scratch. But that also sounds awful 
I was watching docker console logs when it happened.
kill: usage: kill [-s sigspec | -n signum | -sigspec] pid | jobspec ... or kill -l [sigspec]
the console log in the plex webUI didn’t show any errors or warnings as it shut off.
Any indication of which process/user issued the kill command?
I don’t run Plex in docker. Not sure if the log files have such information.
Unfortunately not that I see in the docker container logs or in the plex logs.
I even turned on verbose logging, but nothing at the tail of PlexMediaServerLog1 at all - just normal behavior and it cuts abruptly.
Verbose logs where it happened at 6:57PM while watching a live TV stream.
Posting for good measure here. I’m not seeing anything telling sadly.
I’ve made some posts on reddit to direct people here too if they’ve got any ideas.
VERBOSE - Plex Media Server Logs_2025-03-12_19-06-57.zip (2.7 MB)
SOLVED – It was an hourly script to reduce GPU idle wattage when it wasn’t actively transcoding, set to run once an hour…
Script: GitHub - SpaceinvaderOne/nvidia_powersave: script to save power on Unraid server with Nvidia GPU
Turns out - it’s not working anymore or something… It says no containers using GPUs and it terminates Plex. sheesh.
I was able to replicate this by manually running the script and seeing the exact same behavior in Plex.
Checking for process termination eligibility...
No Docker containers are using the GPUs. Terminating non-Docker processes using NVIDIA devices...
USER PID ACCESS COMMAND
/dev/nvidia-uvm: nobody 30733 F.... Plex Media Serv
/dev/nvidia0: nobody 30733 F...m Plex Media Serv
/dev/nvidiactl: nobody 30733 F.... Plex Media Serv
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
