Server Version#: 1.32.6.7557
Player Version#: N/A, server side issue, web player is at 4.108.0
Plex is preiodically becoming unresponsive. I cannot tell why from the logs myself but I have posted the timeframe from all logs using Promtail → Loki → Grafana scraping the Plex Logs folder. Let me know if you need more from a different timeframe.
Things seem to be running fine until I start seeing “busy database; giving up” and “database is locked” in the logs. I ran a pragma integrity_check which returned ok on the db and blob so it doesn’t seem corrupted. It seems like a thread is locking the database and then dying and never releasing the lock but I’m not sure how to determine that from the logs.
First DB Busy:
2023-09-28 21:53:27.154
Sep 28, 2023 21:53:27.037 [23299443452728] ERROR - [Req#1abb7a] Waited over 10 seconds for a busy database; giving up.
First DB Locked:
2023-09-28 21:53:51.475
Sep 28, 2023 21:53:51.410 [23299253861176] ERROR - Got exception from request handler: sqlite3_statement_backend::loadOne: database is locked
Thanks for the comment! I left things setup tonight like they were when the issue happened. I’ll see if I can reproduce it and fine and simularities in the log.
my windows plex player will randomally freeze/circle spin until I close it and reload it and it works fine, that did start like a month ago… wonder if it’s releated
I also get this error: Waited over 10 seconds for a busy database; giving up. I have tried the DB repair tool, reverting to previous docker versions of plex and deleting the db blobs but still to no avail. I have 5 folders, TV, Movies, TV 4K, Movies 4k, and anime. All folders scan fine EXCEPT TV, which when checking logs has the error:
Waited over 10 seconds for a busy database
and Saving activity history aborted with soci exception: sqlite3_statement_backend::loadOne: database is locked
Once these errors pop up, i cannot open any libraries in plex. This is annoying as I want to add more TV shows but seems I can’t…
I wasn’t able to reproduce the issue last night but it’ll surly happen again as it’s fairly consistent.
Usually hits when Plex Meta Manager is applying overlays. They suggested I jack up my db cache then someone Reddit suggested I tone it down to 100MB since too much cache can be a bad thing. I’ve got plenty of RAM and SSD space so not sure what to set that to, currently at 1024MB.
If you look in the troubleshooting section of the WebUI there are a couple compact database/empty trash/clean bundles options. Since you are adding a large volume of media you can leverage those options.
Still havn’t tried the special script that I linked here GitHub - ChuckPa/PlexDBRepair: Database repair utility for Plex Media Server databases but I will if I run into trouble again. The only thing I can think of is that Plex Meta Manager was doing a TON of overlays at first and that in conjunction with something like a credit scan that Plex does hit a deadlock in the database and hung the server application. Now all of my stuff has overlays and PMM is only doing a minumal amount of what has been added / changed since the last time it has ran and I am less likely to hit the issue. I will keep checking back each morning after maintenance and PMM run together and see if if it’s hung.
Keeping this alive. It happened again to me. I configured PMM to use an obscure value for a Plex timeout so I could easily see when it happened. I woke up yesterday and Plex was unresponsive again. I didn’t have a chance to look at it until now, logs are gone but I still have some in Loki.
Here is the first hit I got from the 5th. It kept timing out calls for an hour or so before the script finished running:
What is different this time is the timeouts happen every 7 minutes. The end is PMM logging all of the timeouts, the little blocks evenly spaced are the 7 minute timeouts:
I pulled the PMS logs around the big spike of database is busy error messages. I can’t tell what what was going on that would cause that but I can see it’s scanning new content.
Errors start around here:
2023-10-05 06:49:36.722 Oct 05, 2023 06:49:36.497 [22716651576120] ERROR - Waited over 10 seconds for a busy database; giving up.
Is there a more formal way to submit issues to Plex for review? The logs read like something is locking the database indefinetly and when that happens it brings down the entire server because nothing can interact with the database. This would be an issue on Plex’s side, not properly cleaning up whatever resource is hung with the database locked (or dead). Would be good to get their opionion on what is going on in the logs. My first set of logs is a bit more complete. I can pull small chunks from Grafana for a wide time range but my file system only seems to hold about a half of days worth with whatever the roll over policy is configured for:
This seemed to have fuzzled out so I will probably open a new thread if I want to reproduce the issue. I’ve had good luck only using PMM to manage playlists & collections in an automated fashion. If I want to add overlays or do anything with posters and ratings mass operations I wait until Plex isn’t scanning anything and run the operation. So far so good but Plex shouldn’t lock up the way it is and I’d appreciate some support since it’s a product that we do pay for.
I don’t use PMM but I use my own Python scripts with the API and if I change any files in the library while I’m running my scripts it causes crashes and/or long-lasting database locks. I can edit the metadata for a single file and when the Plex Scanner runs concurrently with API scripts, the next API call fails with a timeout and the web interface becomes mostly unresponsive. Sometimes I can see the “Waited x whole second for a busy database” messages but sometimes even console crashes. I can’t really find a pattern but it all started when they replaced the scanning engine (last year?). Typically the status message in Plex Web is “Scanning Music: whatever file I changed” and is stuck at 100% of the progress wheel. Following in case I can help with some logs.
Same issue here. I’m running Linuxserver container on Unraid, version 1.32.6.7557. Anecdotally, It seems to only happen when there is at least one person watching something (may just not have enough data to confirm if otherwise) - if a new movie is added to my library, it will scan up to 99.99% (The circle in the dashboard looks completely closed) and just stay there and say “scanning insert movie name”. It will generate credits and chapter thumbnails, but the application will hang. I won’t be able to go into any media library and I can’t play any files on any Plex Client until I restart the container.
If I don’t restart the server, the console will repeatedly say “Waited over 10 seconds for a busy database; giving up.” I could leave it for as long as I want and it will never fix itself. This is very frustrating because I just transferred my server to Unraid. All my drives are good, my appdata is on its own nvme, the database was built from scratch, it passes all integrity checks, and I have done all the things you can do with Plex SQLite but nothing fixes this issue. Hopefully, this gets addressed in the next update because it’s borderline unusable unless I’m constantly monitoring.
