Yep, however I was also able to reproduce in past even when remote access is removed 
Have you any luck with the 1.23.5 Build?
23.5? that a beta build cos I have moved to 1.23.4.4805 and so far its been 36 hours so far with minor issues (mostly CPU bottle necking, think my Ryzon 5 is funky
)
Yes, 1.23.5.4801 is the latest beta build: Plex Media Server - #445 by emilybersin
Thank you for this thread, I am having the same issues as you for the past few months and I tried everything including building a new server but couldnāt figured out why.
Yeah a Server rebuild didnāt solve it for me, if and when it happens to you please do follow the Guide in my OP so that the Devās can have as much Diagnostic information as possible
Well bugger me, just 2 hours after saying things are going ok its only gone and done it 
.
@sa2000 Diagnostics sent again this time from the right server
Surly the Devās must have some clue by now, its coming up to 5 months with this!!
3 Likes
I am hoping all my troubles are now over, I had 1.23.4.4805 do this once since installing it
However I have now fully moved away from Plex on this machine and have started using my backup server as my full time server, This one has never once had this issue but I never fully used it as it has no HW transcoding capabilities.
Just wanted to add that Iām getting the same issue since about 3 days ago.
Currently on 1.23.5.4801-bc4dd58bb but Iāve tried downgrading to 1.23.4.4805-186bae04e.
If I do the full SQL dump database repair it seems to stay up for about 6-12 hours.
Otherwise once it starts it will deadlock in about 1-3 minutes after restarting.
Iāve checked the connections page and it reaches about 100 in 2 minutes.
Half the connections appear to be āpaused tokensā from users with 10+ seconds. The other half appeared to be from Tautulli, so I also tried downgrading that from 2.7.5 to 2.7.4 as that update happened around the same time but it didnāt help.
The other āappsā I use in case it helps are Ombi, Sonarr, Radarr, Tautulli.
OS: Windows Server 2019, bare metal no VM no dockers.
If it happens again Please do follow the Guide I posted in the Op to pull proper diagnostics for the team, as the Dump file of Plex before the process its killed coupled with a copy of the connections and logs will help them pin down the route āStuckā thread that then courses a cascaded effect
1 Like
I went ahead and reproduced the issue on my end by initiating a scan in Overseerr and captured some dumps, connections logs, and diagnostic logs to send over to @sa2000 on PMS v1.23.5.4801. I hope we can stamp out this bug.
1 Like
Iāve sent a PM to @sa2000 with logs, connections and dump for versions 1.22.0.4145 and 1.23.5.4801.
2 Likes
Thank You, I have analysed your 3 dumps and passed together with the connections lists and logs to the development team
The diagnostics were from:
1.23.5.4801-bc4dd58bb
1.22.0.4145-0af3a4016
1.21.4.4079-1b7748a7b
There was a regression in deadlock fixes between 1.23.4.4805-186bae04e and the first 1.23.5 beta 1.23.5.4801-bc4dd58bb
Please switch to the 1.23.5 beta that was released yesterday
1.23.5.4841-549599676
3 Likes
Iām still seeing deadlocks in 1.23.5.4841-549599676 shortly after initiating a library scan in Overseerr. Connections, dumps, and logs have been PMād.
1 Like
Iām really curious how many connections, as before I switch machines Petio (same as overseerr) would open about 50 connections when scanning, but it was actually Plex Clients like Xbox, Firesticks etc that would open hundreds of connections
I think while refreshing the connections page, the most I saw was around 280. The deadlock started around 150ish. I think there were maybe 3 or 4 active users at the time. As long as I turn off my Overseerr docker, I have no issues. I also run Tautulli on the same Windows VM as PMS and it doesnāt seem to cause any issues.
EDIT: Looking at the connections log I sent, I see Overseerr only using 40 connections.
@sa2000 Do you have instructions on how to create a dump with Unraid? Iāve follow the linux instructions and could not create one.
No offense but is there anyone else that can relay dump files and logs to the developers? One guy working part time that hasnāt looked at any of my dumps/logs supplied isnāt enough. I believe I have the same issue as @morpheus2n2 and have taken time to create dumps and logs that we were asked to provide only to have it ignored. OneDrive tells me no one has downloaded the files Iāve supplied. Iād like to pass it off to someone that will.
1 Like
Well, I thought things where to good to be true.
Seems that todays public build 1.23.5.4862 has coursed it to happen within 2 hours of my installing it, as the server is now on a remote machine I wasnāt able to pull dump file in time but will next time.
It may be worth noting all the connections where from an remote Xbox
@sa2000 Ok have had it happen again less than a hour after my last post this build 1.23.5.4862 is very broken and I cant downgrade as I donāt have the pervious update file for some reason, I have send new diagnostics to you!
@StSimm1 Any ideas as to why this build is much much worse, I tried installing the 1.23.6 build you put out a few hours ago but I sucked up 50% CPU upon launch and I was unable to access either the GUI or the Connections page so had to go down to the equally broken 1.23.5 build.
1 Like
