Server Version#: 1.27.2.5929
Windows 10: Version 10.0.19044 Build 19044
Player Version#: All clients
Old Plex installation, has been working fine for at least a couple of years.
Last week (may or may not have been related to PMS update) it was unavailable for all clients one night, web interface only showing “503 Service Unavailable”. PMS was running (icon showing).
A PMS restart helped (=watched some TV episodes), but the night after it was the same.
And same thing happens every day since.
Logs included (Plex update Service.log copied and renamed because 7zip refused to add it to archive because it was locked. Also, installation logs removed due to 900+MB size) Logs.zip (11.8 MB)
What can be the problem? And even more important, how to fix it?
The issue that you are having is called a “deadlock”. Myself and others have this problem and have been working with a Plex employee (sa2000) in this thread: Plex Server Crashing Randomly - #106 by changelin
Here are the steps to get the data Plex employees need:
When the problem occurs (and please wait for sufficient time before capturing diagnostics or restarting eg 10 minutes of requests to the server not getting through) , get the output from a specific browser request on the server. In a browser, go to this url http://localhost:32400/connections?X-Plex-Token=xxxxxxxxxxxxxxxxxxxxx where you put the server token string instead of the xxxxxxxxxxxxxxxx. Copy the displayed output to a text file and save
Next is to force a process dump for the Plex Media Server.exe process. We can do this with Windows Task Manager - but on a 64-bit windows and with Plex Media Server.exe being a 32-bit application, you would need to load the 32-bit x86 version of windows task manager. So load this program instead of the default task manager C:\Windows\SysWow64\Taskmgr.exe
Using this 32-bit Taskmgr.exe find the Plex Media Server.exe process, select it and right click on it and choose Create Dump
Check that the dmp file has been created
Restart the server and capture the zipped logs
Zip the dmp files and upload to dropbox or similar service
Send to @sa2000 by private message link to the zipped dmp files and send the saved /connections output text files and the zipped logs
Feel free to PM me if you need any help with the diagnostic steps. Logs alone will not fix this, we need to provide all of the info outlined above to Plex for them to properly diagnose it.
Thank you for the superdetailed instructions. “Hopefully” PMS will crash again before I leave for vacation tomorrow morning, but I’ll get back as soon as possible.
Meanwhile, at the end of the log I noticed a lot of requests from a specific user, could he and his client be the reason for the deadlock?
I don’t think that requests from a specific user have anything to do with it, but I could be mistaken. The developer that I have been in contact with told be this in a PM:
He said that they have data from 4 users so far and are trying to diagnose it, but so far the cause is unknown. I am hopeful that more data from more users will help them, but I really don’t know.
Thanks for the logs - i can see a deadlock arising
We are seeing if 1.28.1 fixes this - an alpha build is being trialled with few users that had the deadlock - i am including you as well - scroll through the message thread for latest alpha download - which is now 1.28.1.6018
Update for any readers of the thread. It took a few days for the server to crash again, but I have now sent to debug data. I am offsite on vacation and thus do not want to try alpha stuff right now, so I’ll wait a week or so and see what has happened once I’m back.
Thanks for the diagnostics. Yes it was a deadlock on version v1.27.2.5929
The potential fix that was made available in the alpha 1.28.1.6018 I made available to you is now also released in the latest beta 1.28.0.5999 which will be on public release by the time you are back rom vacation
So please switch to 1.28.0 when you are back - no need to try the alpha now
