Last week I updated the computer which hosts my plex media server from Ubuntu 16.04 to 18.04. This happened to coincide with the release of the new Samsung TV UI. Performance was horrible. I couldn’t watch a show for more than 2 minutes without what I thought was the TV app freezing. The only way to fix it would be to log into the server and through a terminal, kill the process and restart the service. Well, the TV App was rolled back and the performance issues remain, but now I also see they are not limited to the one playback device. Even if I play a video locally on the server, it will cause a freeze which requires me to kill the plex process and restart the service. How can I track down what is causing the server issues?
The best way to start (really only way to start) is by
Verify DEBUG logging is enabled in Settings - Server - General, VERBOSE logging is disabled
Restart PMS
Recreate the failure
Wait 30 seconds after stopping playback / the freeze to collect the logs (Settings - Server - Troubleshooting - Download Logs).
a. If PMS itself has frozen, restart it then collect the logs immediately.
This was created right after a freeze. I was watching a recording on my tv. When it ended, the tv app could no longer see the plex server. Went to the server and found that the plex service had frozen. I needed to kill the process and restart it.
Another interesting tidbit. Though I finished watching the whole recording, the restarted server thinks I only watched 15 minutes of a 45 minute long program.
Next time it hangs, Please force the core dump termination when you kill it (kill -15 ) (SEGV)
PMS will trap that kill signal and generate a crash dump.
When it restarts, it will upload.
We only need the cash dump report (in the “Crash Dumps” directory under PMS) to track it for you.
further to what @ChuckPa said, we would also need the connections list at the time of the deadlock.
See Diagnostics for Deadlocks on NAS/Linux in this post for details of what is needed
connections list and noting down the time, a process dump invoked via kill -segv <pid> and the logs taken after resarting the server and the crash report for the process dump uploaded
what time was the /connections request made?
what time was the kill -SEGV pid done?
I need the dmp file to be automatically uploaded so it gets processed and analyzed
could you place it back in the relevant folder /var/lib/plexmediaserver/Library/Application Support/Plex Media Server/Crash Reports/1.14.0.5465-425e4ac90
and restart the server so it gets processed please
and then get the logs to show the crash report upload happening
Also having these issues with update to Version 1.14.0.5465. Performed update today around noon and unable to really use Plex on any of my 4 Rokus since. Can this be rolled back?
I notice most of the requests are from (TV UN65KS800D) on 192.168.86,189 - can you tell what version of the Plex App and if it is the same version as you mentioned in your first post above
The deadlock fix that got released today in beta 1.14.0.5468 unfortunately does not appear to fix the deadlock you have.
I would still recommend that you upgrade to it.
The process dump you provided appears to have been invoked after a shutdown attempt. So I am waiting for fresh deadlock diagnostics - but best to get it with version 1.14.0.5468