Logs Attached Logs.zip (3.4 MB)
The logs attached we’re taking after a crash but before the restart.
I’m running plex on a docker container on Unraid from the official plex repository.
So starting about 2 weeks ago my server goes down every night without fully crashing, the libraries become unreachable and the only remedy is a restart of the Plex container.
I have tried:
Repairing the database using SQLite commands
Changing every setting known to man
Removing all plugins (originally i thought subzero subs we’re doing it)
Rebuilding the entire container from scratch save the .DB files (rebuilt metadata)
My logic with everything I’ve changed is that it has to be the .DB but it’s not something that the SQLite commands are able to fix. This is a fairly old DB with lots of watch status and custom collections that have taken me countless hours to procure, please any help is welcome. I’m really hoping there is a fix that isn’t starting over from scratch.
Because it crashes in such a subtle way, there is no way to set up any auto-restart parameters, so I basically have to wait for one of my users to text me saying “hey the server is down again” before I know.
I have the same issue with Plex timing out every few hours but not completely crashing. Downgraded to 1.22 and it helped with the crashing but Plex still times out almost nightly. I moved Plex to a new unRAID server. It works great for a day or two and now back to the timing out again.
My db files are 8gb and 2.2gb. I’ve even done a restore from a previous month and still have issues
My Plex docker (running on Unraid) is also crashing. Very similar symptoms as Snowblind182.
Plex becomes unresponsive but does not 100% crash. From inside Unraid it looks like Plex docker is still running. The Plex docker will not shutdown using Unraid ‘normal’ stop command.
@snowblind182 do yo have a suggestion on how to kill / shutdown the crashed Plex docker?
Based on looking at your logs, I did see a few INFO entries on July 8th about SQL statements aborting because the database schema has changed, so I’m thinking maybe a database update didn’t possibly complete successfully. I’m also seeing what looks to be some network issues and possible file permission issues, but as I’m just another user, I can’t say for certain. It also looks like you might have tried to optimize the database as well and that seemed to have failed and there are a lot of slow query entries as well.
One thing I did notice is that it looks like you might have verbose logging on (or at least had it on), as I saw a few lines stating VERBOSE as the level. I would suggest going to Settings and then under Settings click General and then make sure the checkbox for “Enable Plex Media Server verbose logging” is unchecked. This way if someone from Plex asks for new logs Verbose won’t be on, as they only want that turned on if they specifically ask for it as it can hide the problem with the amount of data that gets logged.
I don’t remember who the Docker expert is for the Plex members that reply here, but I’ll try to ping @anon18523487 at least and hopefully he’ll see this and be able to give some assistance or at least ping the appropriate person.
@snowblind182@nbguy Do you happen to be running PiHole or Adguard? I noticed “analytics.plex.tv” was the top blocked domain in Adguard Home. I uninstalled the adguard home docker and changed the LAN’s DNS back to 8.8.8.8 yesterday afternoon and so far no “unresponsive” issues so far.
I’ve only just installed PiHole 2 days ago and immediately ran into issues with unraid getting routed through it. So i had to manually configure the DNS to 8.8.8.8.
So it couldn’t of been this as it’s my latest change to the configuration, and was added weeks after the soft crashes from Plex.
Do you have the DB backup from version 1.22? 1.23 did database updates so downgrading the database from 1.23 to 1.22 still made my server unstable. I had to find an old db backup that was never upgraded to 1.23.
I’m currently using Repository: linuxserver/plex:1.22.3.4392-d7c624def-ls44
Makes sense. I believe that the default number of logs kept is 5 but I think there is a way within the preferences.xml file that allows you to increase that number which could potentially help with this as you’ll have more logs to look at. You’ll want to stop the Plex docker first and then update the preferences.xml file I believe. If you update it while the Plex docker is running then I think it will get overwritten or just not take (I know I had a little trouble getting it to take I believe, though that could have been another variable I was trying to update in that file). Just some extra info.
As it’s only been 1 day since I pinged the employee, I’ll check back in a few more days to see if they popped in, otherwise I can try and ping someone else. I just want to give them time to see the notification as they are probably busy.
Hello, @Weeflix I don’t use PiHole or AdBlocker but I do have pfBlocker running on my Firewall, but that has been running for 12+ months. My Plex issue started about 2 week ago.
Yeah I had one DB from 1.22, rolled it and the Plex back to 1.22 and it still had the same issues. Is there a way to have plex keep more backups. In my case it only keeps 3 at a time and then starts replacing the oldest one.
You’d have to either manually copy the databases over to another location or if you are decent with scripting, write a script that runs from a cron job to do that. As far as I am aware, there are no settings to increase the number of logs that are kept.
Snowblind182 said “the only remedy is a restart of the Plex container”
My symptoms seem very similar but I am unable to restart the Plex container using Unraid commands. The Plex docker will not stop or restart using Unraid GUI commands.
I also tried “docker kill Name-of-container” without success.
I am required to attempt a shutdown of the Unraid server, but the server will not go down 100%, hard power off server & reboot. Now a parity check is forced when starting.
Plex is the only Docker I keep running (on a normal day). Now, I can not keep Plex running because of the lock ups.
I am really looking forward to a new version of this docker.
My only other idea would be to uninstall the container entirely and rebuild it, pointing it to the same folder in your appdata. This won’t do anything for your Plex crashing but might fix your container acting up.
This might be something you want to post to the unraid discord/forums. Sounds like you have two issues. 1 issue we share with Plex and another unrelated unraid problem.
Back to square one. My plex server went unresponsive yesterday around 10pm CST and 10pm CST today as well. I had to restart the docker a few times to get it going again.
I have four servers running the Plex Docker container on CentOS with one of them experiencing issues similar to that stated in this thread for almost two months now. Plex deadlocking randomly without fully crashing. Dashboard becomes unresponsive and a forced restart needs to be performed.
When the issue happens it’s still possible to input a URL like this: http://127.0.0.1:32400/connections?X-Plex-Token=############# and it will display hundreds of deadlocked connections. Not sure if any of you guys are having the same issue but it sounds very similar.
So far the crashing has not been as frequent with the newer versions and now running on 1.23.5.4801.
@sa2000 This thread sounds similar to what we are trying to investigate.
Thank you for the reply, @kegbeach , I will keep an eye out tonight on the connections url to see how many connections I have. Currently I am 29 connections with no one actively using Plex this morning.
I updated the Plex docker on July 13 but I only left the docker running for a short time then stopped it. I did not leave it running over night.
Mid day July 14 I turned the Plex docker on. I stopped and started the docker twice during the day. All was good. I left the docker running overnight.
Today, the morning of July 15, I was able to successfully stop the Plex docker.
Time will tell, but I was able to survive one night (so far) without a lock up.
Can I have some help with the URL mentioned by @kegbeach ? Correct the IP, I get that. Do I need a ‘token’ ? Where do I find the ‘token’ ?
Thank you
