Server Version#: 1.15.6.1079
Player Version#: All, across multiple devices
Having a strange issue where the servers (I run two) are crashing/hanging. Not sure what to do - perhaps someone could shed some light?
We would need to see the DEBUG logging (not VERBOSE) to understand what it’s doing.
Running with DEBUG on does not utilize any additional disk space and is the default setting.
If, after you optimize your database (I found this), Please set the logging to DEBUG enabled and restart.
May 19, 2019 23:09:36.244 [0x7f4a1e7fc700] WARN - SLOW QUERY: It took 210.000000 ms to retrieve 1 items.
May 19, 2019 23:09:36.257 [0x7f49f17fa700] WARN - SLOW QUERY: It took 240.000000 ms to retrieve 1 items.
May 19, 2019 23:09:36.257 [0x7f49f0ff9700] WARN - SLOW QUERY: It took 260.000000 ms to retrieve 1 items.
We can then examine to see what’s happening and potentially causing the busy database issues which are showing up. It is entirely possible, with the slight evidence of transcoding activity you are showing in these logs, the CPU can’t keep up with the load. It only has 4500 Passmarks (about 1/2 of a 2012 i7-3xxx series)
https://www.cpubenchmark.net/cpu.php?cpu=Intel+Xeon+E5530+%40+2.40GHz&id=1244
Okie dokie - enabled “Debug” mode and rebooting the servers. This wasn’t an issue with the same hardware prior to a couple of releases ago. Same hardware, same users, same load on server. Will attach logs the next time it crashes.
Thanks!
There have been a lot of changes and additions in the most recent versions including the new CI as of 1.15.0.
Understood - totally appreciate the work the team does! Just a waiting game for me now as to when it’ll dump. Thanks so much for your time during this.
I’m also having a problem with crashes, same symptoms as you. I’m running a big ol server w/ ubuntu 18.10.
Finally crashed again, here are the logs.
Thanks, I see where PMS stops responding.
I also see 17 plug-ins?
Did you add some plug-ins of your own?
May 21, 2019 21:23:57.268 [0x7f2f63fff700] DEBUG - Starting plug-in /usr/lib/plexmediaserver/Resources/Plug-ins-78232c603/OpenSubtitles.bundle.
May 21, 2019 21:23:57.268 [0x7f2f84ff9700] DEBUG - Starting plug-in /usr/lib/plexmediaserver/Resources/Plug-ins-78232c603/TheTVDB.bundle.
May 21, 2019 21:23:57.268 [0x7f2f637fe700] DEBUG - Starting plug-in /usr/lib/plexmediaserver/Resources/Plug-ins-78232c603/TheMovieDB.bundle.
May 21, 2019 21:23:57.268 [0x7f301dffb700] DEBUG - Waiting for 17 plug-ins to finish starting.
May 21, 2019 21:23:57.279 [0x7f2f63fff700] DEBUG - Perfect cache hit, we don't need to start [com.plexapp.agents.opensubtitles] yet.
May 21, 2019 21:23:57.279 [0x7f2f63fff700] DEBUG - Starting plug-in /usr/lib/plexmediaserver/Resources/Plug-ins-78232c603/PlexMusic.bundle.
May 21, 2019 21:23:57.283 [0x7f2f84ff9700] DEBUG - Perfect cache hit, we don't need to start [com.plexapp.agents.thetvdb] yet.
May 21, 2019 21:23:57.283 [0x7f2f84ff9700] DEBUG - Starting plug-in /usr/lib/plexmediaserver/Resources/Plug-ins-78232c603/Media-Flags.bundle.
Thanks for the logs.
Please leave them here for now. I’ve asked one of the team members who specializes in this to take a look.
Thanks, Chuck!
I only have the “services.bundle” - see screenshot.
Any traction on thus? Not impatient, just curious 
I have been following best I can and they are working on the crashes.
Linux is normally the most stable so this is getting attention.
IIRC, it is a timing problem where two tasks get out of sync and one breaks the other (but i might not be 100% accurate on that)
Fantastic, if it helps identify a problem for others then I’m happy to help diagnose. Thanks for your replies.
Looked at the logs - the first were without debug logging but the second set had debug logging. The server has been crashing with what appears to be similar crashes as reported on this forum topic Ubuntu 18.04.2 LTS - Plex Server Crashes - Double free or corruption (!prev)
The issue is still unresolved. It appears to arise sometimes when getting to the end of playback from a transcoded video. It is a timing bug where a couple of threads are dealing with requests from the app notifying end of playback - and some plex clients (eg Roku Plex app) send this indication in more than one way.
The last set of logs show a crash at 11:23:10 am on May 21st when playback stopped for remote user sewxxxxwithxxxndi on a Roku. A crash report was generated but the server process did not terminate and continued to run in cripple mode - eventually deadlocking and stopping to handle requests with the last completed request at 12:30. You eventually stopped the server at 21:21. The crash report was uploaded to our server at 21:23
The cause of this issue is yet to be identified. Attempts to reproduce in a debug environment have failed so far - I am liasing with the development to help identify the bug.
Fantastic, please let me know if there’s anything else I can do to help.
I wonder i today’s release is to attempt to address the issue:
- (Database) Scheduled database backup job could cause a crash (#9993)
I expect to have a development build later today with potential fix for the crash that I looked into from your server
Excellent news, glad to be of some limited service
I sent you a download link for Linux for development build 1.16.0.1181-9505422cd
Please let me know if the crashes stop. Let me have logs if you continue to have crashes - or if you spot any new issue
Thank you
Will do, external users will have to wait for the current randomness to be corrected before they can connect again, but will report when able.
