Plex Deadlocking with to many hung connections issues for last 5 months Now!. (This ended up giving me anxiety attacks!) (Edited title for clarification)

For the last 3 weeks I have been getting random deadlocks, If I go to http://127.0.0.1:32400/connections?X-Plex-Token=############# I can see hundreds of connections some say paused for over 10 mins.
The latest one was 407 connections!!
I have lost a lot of sleep over this it is actual effecting me mentally and yes I know it shouldn’t, unfortunately it is which is why I am desperate for help!!

I Run on Windows 10 and its been fine for months/years with minimal issues until late last month.

Following this guide I have collected the Logs, dumpfile and also a copy of the connections page.

I also noticed when looking at Resource Monitor the Plex Media Server process had hundreds of items in the wait chain so its clearly locking some where.

Edit: Steps I have taking to try and fix and still not much luck

  1. Full uninstall and reinstall of plex more than once
  2. Full Nuking of Database and Metadata and rebuild (this took almost a week to do)

Edit 12/06/2021:- While this has been giving me Anxiety attacks I have sort help!, Also I am hoping another team member can hop on and let me know if there’s been any progress with the various diagnostics I have been sending and/or if anything else would be needed to help with the investigation?
While it has reduced down to once a week since update 1.23.1.4602 with the DNS fix, it dose still happen and I do believe these are connected!

I have found that killing Plex and then restarting my router will normally solve this however sometimes I have to keep plex off for about 40-60 mins before starting it back up otherwise the issue will appear within a matter of min’s.

1 Like

Is there anyone at all that can help me try and pin down what is the problem.

This whole thing actually ended up making me rather ill due to the stress!!

While it has improved it still happens at least once every 7-10 days.

Along with the full Nuke of the DB and metadata drive I even ended up fully nuking the server, everything bar the media was fully nuked and drives zeroed.

Yet I still get this, it all started about 3 months ago now I have a feeling it’s something on my network but I have no clue to to test and troubleshoot as everything works perfectly fine at full speed bar Plex when it dose this.

What I have noticed in the wait chain I mentioned in last post is it says Plex is waiting for network I/O which is why I am thinking something is holding it up at random on the network.

Good luck with your anxiety; it’s a big deal and I hope you’re getting whatever help you need. :hugs:

Do you notice any patterns in the connections? Are there are large number from specific IP addresses, or certain users, or to specific endpoints? I’m wondering if there are any clues that might help diagnose the issue, or might help you workaround or avoid the problem.

Do you have any additional security software on the Windows server? Antivirus, firewall, and network-security software is notorious for interfering with networks connections and database file access.

1 Like

Happy to help. You mentioned you have captured process dump / connections list and logs for one of these - if you can upload somewhere and send me link by Private Message - I will look into it. I only work one day a week - so if you get the diagnostics today I will look into it and if the evidence is good - i will refer it to the development team. If not today, then will be around next Tuesday. It is always best to get the diagnostics from the current version - but if you have it from a recent release that will be good enough

Please double check that debug logging is enabled

1 Like

Thank you, yeah in the end I shut plex off for a few days and basically slept for 3 days lmao.

Yeah it never seems to be the same users but normally a very similar listing of “Paused 212.961s - GET /:/timeline?key=%2Flibrary%2Fmetadata%2F229187&ratingKey=” obviously with the user and IP along with the key for the media.

Its the paused bit as that time can be from a few seconds right up to 10+ Mins.

Thank you for the complete set of diagnostics. Perfect.
I have referred the details to the development team
The deadlock was on 1.22.3.4392
Would be good if you also get a set of diagnostics from the current beta - or whichever is current when you get the issue again

Thanks

Yeah when it happens again I will defo grab all new sets for you and PM you so that you can pick it up next time your able :smiley:

Thank you so much

Oh could I ask, that I be updated as to if its a bug or something my end when more info is found out as this has effected me mentally due to the stress that I would appreciate having my mind put to rest.

Thank you so much

This has unfortunately happened a few times today again, I have sent @sa2000 a fresh set of Diagnostics.

The stress this is coursing is unreal, I don’t know why its effecting me so much and really wish it wouldn’t, guess its part of the midlife crisis😭

Hi @sa2000 Did you get my new set of log and diagnostics? Do you know if there has been any progress?

The development team want to know if this is also happening on latest version 1.23.1 (public) ?

I have not looked at the new diagnostics you sent - (I am only working 1 day a week - back next Tuesday)

So far no, I am hopeful it was the DNS thing that was the problem but not knowing what I’m looking at in the logs etc.

So I’m not 100% sure if its fixed or just biding its time, not to mention it went on for 3 months and at random.
So as you can imagine my anxiety is still rather high lol

@sa2000 Unfortunately, it has happened again in the last 30 mins (04/06/21) so I have sent you another set of diagnostics :smiley:

Please do let me know the progress as this did course anxiety attacks, I have been trying really hard this last week not to let it effect me.

However I have been able to restore it a lot faster this time round, I killed the process rebooted my router once the internet was back up I restarted Plex and it has fixed it.

I suspect this is part of the DNS fix ( #12710) that’s actually still being a problem! as since that update went out it has reduced in frequency a lot!

Thanks for the new deadlock diagnostics = logs / connections list / dump
I have referred them to the development team - they were for a deadlock on 6th June with requests starting to hang at 22:14 running on 1.23.2.4625-a83d2d0f9

Brilliant, I am pleased that its showing up and hopefully the Devs can find the course and nip it in the bud :smiley:

I just want to say, if your anxiety is as bad as you seem to indicate, I would suggest you seek out to medical professionals asap.

Letting yourself get so consumed with anxiety is no joke, the struggle is real, and the effects can be devastating for yourself and your family.

4 Likes

I have now, normally I have never let these things bother me but coupled with current situations this was the tipping point.

Thank you for the kind words :smiley:

2 Likes

@sa2000 Unfortunately, it happened again, I didn’t capture anything this time as I just shut it off and have left it, not worth the stress. I do hope all my previous captures (still available on drive) are enough for the team!

I have also noticed, it seems to be mostly Friday nights it happens (never the same time, and never the same users!)

If it helps at all, I am on BT Full Fibre 1Gbps, with the Smart Hub 2 (cos I believe this may be a facture), if the Devs need any more details let me know, I do have the logs from tonight just not the connections or dump file.

The diagnostics were looked into by the development team but have not been able to find the culprit thread that caused the deadlock

So would be good to get another set - and again please try the latest beta release. Current beta is 1.23.3.4692-c86c6f454

Thanks

No worries, will grab latest beta and pull Diagnostics if it happens again :smiley: