Old deadlock bug has reappeared in the last 2 public builds

A bug from 2 years ago that also took almost a full year to fix has reappeared (this one here Plex Deadlocking with to many hung connections issues for last 5 months Now!. (This ended up giving me anxiety attacks!) (Edited title for clarification) )

It’s been going on about 2 weeks randomly but always with out fail between midnight and 1am, problem is I can’t grab all the info needed as I work nights, so first I know is when what ever I am listening to or have playing in the background on my server stops, when I remote in to see my connections list I have over 200 active connections to my server (as of writing there’s 410 open)

1 Like

Thanks for reporting this. I do not have any other report of this yet.

Lets hope we can get to the cause quickly this time. You know the drill as you were involved previously with the similar issue back in June-August 2021

So please run the server with debug logging enabled.

Make sure you know what the server security token is to include in the request for the connections list when the issue arises

May need to increase the number of log files from the default - the server setting is LogNumFiles - see Advanced, Hidden Server Settings | Plex Support

If running at default, double the number of logs to 10

When the problem arises, get the number of connections on the server through

http://127.0.0.1:32400/connections?X-Plex-Token=xxxxxxxxxxxxxxxxxxx

Try to access the server libraries - as confirmation that the server is inacessible

Wait for few minutes and repeat the connections list test to see that is growing
http://127.0.0.1:32400/connections?X-Plex-Token=xxxxxxxxxxxxxxxxxxx

Save the output from the two connections lists requests in text files - suggest using filenames that indicate the time the request was made

Force a process dump for the Plex Media Server.exe process in Task Manager

Restart the server and get the server logs zip

Zip the dmp file created earlier and upload with the other files/zip to dropbox / google drive etc and send me a private message with a link

Please run with the latest available version of Plex Media Server - at this moment that is 1.31.3.6868-28fc46b27

Thank You

Thank you, I am looking in to best way to remote access my server when at work for when this happens so I can grab the bits and save them till I get home to post here lol

@sa2000 @ChuckPa (first of all sorry to tag you on this ChuckPa since you weren’t active in this thread) - not sure if on the two latest builds you have seen/heard of deadlock issues or unresponsive servers with the changes for credit detection functionality? I have a few plex servers and have just been doing some testing with it on (as media is added & as a scheduled task) vs off with the same media.

It seems when credit detection is happening even though performance is not bound by the CPU (it is not near maxed out, nor RAM of course) - the server pretty much locks up. Only noticing this in the last build or two as mentioned.

Also, Chuck I wanted to say thank you for your work on the DB repair scripts! I had my own for Windows, but yours is just a bit cleaner :smiley:

Do you have credit detection enabled (or “Generate credits video markers”) and if so, have you noticed whether it is/was processing files for this functionality when it gets locked up?

NO I have turned it off, did this when it first happened and saw a post about it, however the deadlocks still happen with the intro and credits both disabled.

Just curious, what about intro detection?

We’re going to need DEBUG logs (verbose OFF) ZIP files which capture it becoming non-responsive.

If it takes a long time to manifest, increasing log retention count is the best course of action (LogNumFiles="10")

note: glad my script work helps. Here in the forums is a thread where we can discuss it if you wish.

Like I said it’s turned off.

I’m starting to get a little annoyed with the fact it seems to only deadlock when I go to work, been here and hour and it’s just died, did a connections test to find 200 open connections. Have no way to remote in to grab the logs or dump file sonyet again it’s going to have to sit dead for next 7 hours till I’m home grrr

If you are always away when this happens, suggest you temporarily enable Windows Remote Desktop on your PC and try and gather the diagnostics whilst being away

Please run with the latest beta / public release - currently 1.31.3.6868-28fc46b27

The diagnostics are summarized in my earlier post Old deadlock bug has reappeared in the last 2 public builds - #2 by sa2000

Well thankfully this has suddenly disappeared in the latest public build.

Been a almost a week without this issue popping up

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