Crashing servers

I am in the process of moving my Plex service from a Mac to a linux server but experiencing problems with both servers. They are on the same local network (about 3m apart) and both running the latest PMS (updated to latest yesterday, 1.32 I seem to recall). The problems are obviously ‘connected’, but I suspect not related, however I thought it made sense to ask about both together.

Mac is a MiniServer 5,3 (Mid 2011) that has been running PMS successfully for quite a few years. However I am moving duties to an Intel i9 based dedicated media server running OpenMediaVault (latest 6.3 something) which is basically debian (bullseye). This involved moving most of the media files from the Mac to the OMV server and it seems since then, PMS on the Mac crashes every day at just after 02h00 in the morning. Obviously a scheduled task, but what? I have tidied up the database and removed everything that no longer exists on the Mac and it all looks fine. But next morning, I cannot access its web server, although ‘bootstrap.py’ seems to be still running. However in the MacOS GUI, it is no longer running. I can start it again and all is good - until the next morning.

The OMV server now has possibly 4TB of media to deal with. Not a problem until I started setting up libraries in PMS. Then the ‘Credits checking’ starts and after some time (hour or 2) the machine appears completely HUNG. This is linux so somewhat surprising. First of all PMS’s web interface stops responding, then I notice the NFS shares have disappeared from the Mac (Pro) on which I am trying to administer both PMSs, SSH connections are broken and even the server’s actual monitor is black and keyboard non functional. It is still running, but so severely stuffed nothing can be done except hit the power button, which I hate to do, but no alternative.

After it hung the first time, I restarted and with PMS showing no ‘Activity’, it seemed ok. Till I created another library which kicked off the ‘Credits scanning’ again and within an hour, completely stuffed again.

Logs on both seem to show things working, but then suddenly, no more entries until the restart (PMS on Mac and whole machine in case of OMV server). However, I don’t think it’s the same problem on both machines, even though problems occurred only after moving the files from one to the other.

I’m sure the logs will reveal all to those who wrote the code that creates the logs, but so far I’ve been unable to figure it all out.

This is all rather annoying at the moment, with 2 PMSs acting up and would appreciate some assistance in tracking down these problems. Like some real support from Plex themselves perhaps - for a fully paid up Plex Pass holder?

I have been trying to fix the crashing problem of PMS on the Mac. It occurs shortly after 02h00 which is when the scheduled tasks start. I changed the start time to 03h00 and the crashing then occurs just after 03h00. So I think that conclusively proves it is one of the scheduled tasks causing it.

So I’m turning them off one at a time. Still crashing, but still a few to go. Hopefully eventually it will stop crashing and I’ll know which task is causing it.

However, it is again ‘crashingly’ disappointing that something as serious as the company’s main server product crashing every night is not considered serious enough for Plex to provide any form of support - to a fully paid up customer I might add.

Mine is also crashing (on a 2019 Mac Pro) I shut off all of my scheduled tasks and that seemed to stop it, but then it was still randomly occurring. Turns out one of my users has been downloading media, and it crashes whenever he downloads something now. I haven’t had time to troubleshoot myself, but It may be related to transcoding, or optimizing the media. I hope you get your crashing sorted.

PMS 1.32.0.7002 is BETA.

PMS 1.32.0.7002 has known issues.

We are working on them.

Please downgrade to Public 1.31.1.6999 until the new BETA is posted.

Thanks for response, but I do not use betas. Mac is currently on 1.32.1.6999.

There are still some issues with Credit and Intro detections which run in scheduled tasks.

We’ve applied those fixes and fixed the regressions which occurred in 1.32.0.
also fixed Nvidia transcoder breakage.

A lot going on todqy. Everyone is testing. Please give us a bit.

I am happy to wait, as long as I know you’re on the ball. It’s when things like this are occurring and there’s just silence from you guys, like you don’t care.

The Mac is crashing during scheduled tasks, but the linux PMS literally hangs the machine when doing the credit check after adding new libraries.

I’ll let you get on with it.

I’m sorry for being more silent than usual.
I’ve been on some heavy antibiotics and steroids these past 3 weeks but only now turning the corner.

Well this is how it is on my Mac PMS. I have changed the time of the Scheduled Tasks and the crash changes time too. I have been individually turning off the tasks one at a time. Here’s how it was yesterday:-

Screenshot 2023-05-19 at 12.07.54

It crashed again. So none of the processes shown are the problem. There are obviously other routines run at that scheduled time, but which are not shown to the user.

The last log entries are:-

'Plex Media Server.1.log'
May 20, 2023 02:18:15.756 [0x700003b66000] DEBUG - BPQ: onConsiderProcessing: Idle (true)
May 20, 2023 02:18:15.756 [0x700003b66000] DEBUG - BPQ: [Idle] -> [Processing]
May 20, 2023 02:18:15.758 [0x700003b66000] DEBUG - BPQ: generating queue items from 0 generator(s)
May 20, 2023 02:18:15.758 [0x700003b66000] DEBUG - BPQ: generated 0 item(s) for queue
May 20, 2023 02:18:15.758 [0x700003b66000] DEBUG - PlayQueue: 0 generated IDs compressed down to a 20 byte blob.
May 20, 2023 02:18:15.758 [0x700003b66000] DEBUG - PlayQueue: 0 generated IDs compressed down to a 20 byte blob.
May 20, 2023 02:18:15.758 [0x700003b66000] DEBUG - BPQ: [Processing] -> [Idle]
May 20, 2023 02:18:15.758 [0x700003b66000] DEBUG - BPQ: garbageCollect - version - forced - -1
May 20, 2023 02:18:15.758 [0x700003b66000] DEBUG - Versions: garbage collecting
May 20, 2023 02:18:15.759 [0x700003b66000] DEBUG - Versions: garbage collected in 0.0 seconds

and then

'Plex Tuner Service.1.log'
May 20, 2023 02:18:37.854 [0x7fffaf609380] DEBUG - Parent process no longer present; exiting
May 20, 2023 02:18:37.855 [0x7fffaf609380] INFO - [I] device_manager_t::term
May 20, 2023 02:18:37.855 [0x7fffaf609380] INFO - [I] device_manager_t::term. Stop all devices
May 20, 2023 02:18:37.855 [0x7fffaf609380] INFO - [I] device_manager_t::term. Shut down all device auxes
May 20, 2023 02:18:37.855 [0x7fffaf609380] INFO - [I] tvbutler_aux_t::term
May 20, 2023 02:18:37.960 [0x7fffaf609380] INFO - [I] web_server_t::stop. Web server is stopped
May 20, 2023 02:18:37.960 [0x7fffaf609380] INFO - [I] device_manager_t::term
May 20, 2023 02:18:37.960 [0x7fffaf609380] INFO - [I] device_manager_t::term. Stop all devices
May 20, 2023 02:18:37.960 [0x7fffaf609380] INFO - [I] device_manager_t::term. Shut down all device auxes
May 20, 2023 02:18:37.960 [0x7fffaf609380] INFO - [I] broker_t::shutdown. Shutdown is completed.

I suspect the latter is due to the former. It’s always the same. It all stops about 20 mins after the ‘Scheduled Tasks’ start time.

Hope that might help diagnose the problem. Not sure what else I can do to determine the real reason for the crashing.

@ChuckPa , any update on the problem and possible fix?

Somewhat surprisingly, PMS Mac did not crash last night, even though I had changed nothing. The continuation of the log is as follows:-

May 22, 2023 01:58:50.288 [0x70000408f000] DEBUG - Versions: garbage collecting
May 22, 2023 01:58:50.290 [0x70000408f000] DEBUG - Versions: garbage collected in 0.0 seconds
May 22, 2023 02:00:01.202 [0x700003f89000] DEBUG - EPG[cloud]: Next thing to start/end is at 2023-05-22 02:05:00 (in 300 seconds)
May 22, 2023 02:00:11.593 [0x700004630000] DEBUG - [ViewStateSync] Starting
May 22, 2023 02:00:11.593 [0x700004630000] DEBUG - [ViewStateSync] Fetching user tokens
May 22, 2023 02:00:11.593 [0x700004630000] DEBUG - [ViewStateSync/HCl#f42] HTTP requesting GET https://plex.tv/api/v2/server/users/services?auth_token=xxxxxxxxxxxxxxxxxxxx
May 22, 2023 02:00:12.342 [0x7000045ad000] DEBUG - [HttpClient/HCl#f42] HTTP/2.0 (0.7s) 200 response from GET https://plex.tv/api/v2/server/users/services?auth_token=xxxxxxxxxxxxxxxxxxxx

So after the 2 lines regarding ‘garbage collecting’ that was previously the last entry before it crashed, last night it continued with EPG stuff. So it rather suggests that the EPG gathering process was causing the crash. However, why not last night?

Is it possible Plex have changed some things their end that allows the EPG process to complete without causing the crash?

Needless to say, a LOT of entries continuing from the above. Previously, there’d be nothing until I started PMS again.

I am seeing SQLite errors about schema having changed. However they don’t cause a crash. Cannot help wondering though why there is this problem with the db schema.

Just after that last post, thee was an update, which I installed and since then Mac PMS has not crashed. This version of PMS seems to be ‘constructed’ differently as there are many more plex based processes running.

Yesterday I turned all the scheduled tasks back on and still running fine this morning. But…

The Plex Mac app ( Version 1.69.1.3755-e00242c8) now hangs when quitting. Same thing yesterday and today. Works perfectly, but when I try to Quit, just get a spinning beachball and it remains ‘App not responding’ and I have to Force Quit.

Regular web access is no problem and closing that browser window does not seem to be a problem.

Does the new version of PMS require a new Mac client app?

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