SQLite3 DB becomes locked on PMS startup past v1.23.3.4707

[ghoste]

Server Version#: v1.23.6.4810-15ce0e21a (currently on 1.23.3.4707-ebb5fe9f3 since anything above is unusable)
Player Version#: Plex for Windows v1.33.0.2444-a220eae4
Host OS: Ubuntu 21.04
‘docker inspect plex’ output: dockerinspectplex.txt (15.6 KB)
Log: Plex Media Server.log (461.3 KB) (If Plex staff wants the full .zip, I can DM it)

Description:
If I update to any PMS version later than docker tag v1.23.3.4707-ebb5fe9f3, the SQLite3 DB becomes locked roughly upon PMS start and every item on the server times out with “something went wrong.” I can downgrade to v1.23.3.4707 from v1.23.6.4810 with zero DB modifications and it begins working perfectly again. Upgrade again, it breaks, downgrade again, works again, rinse repeat. Tried removing any synced/downloaded items I had on all devices, did the “repair corrupt database” process, rescanned and reanalyzed all libraries, cleaned bundles and more importantly ran ‘optimize database’, none of which made any difference. The PMS directory is on a high-end gen3 NVMe drive which had near zero activity so that’s definitely not the issue.

Any help would be appreciated!

[Hale_House]

I seem to be having the same problem on Windows with version 1.23.6.4810.
I’ve tried verifying the database, repairing it, and reverting to a backup database.
I’m trying an optimize right now but it took a long time to start and it had to try twelve times to begin.
While I’ve been typing this it’s finished and I’m still seeing database locked errors in the log.

[ghoste]

The thread here is the main thread regarding this issue now so that would be the best place to post. Good news is they’ve found at least a temporary workaround and a build should be pushed soon. A more deep dive and long term solution will be worked on next monday.

If a plex staff member comes across this thread, can you lock it? It’s linked on the main thread so it can be referenced if needed.