Database locked after Image phash generation

[max_must]

I recently downloaded the latest release (v0.31.0) of Stash and migrated my database to check out the new features. I started by running a selective generation task and calculating phashes for a few images, which worked fine.

I then started a phash generation task for my entire image library. Two days in, I got a bit concerned that it was still running, so I checked the /config folder and saw that the SQLite database’s -wal file had grown to 112 GB. I decided to cancel the task and see how many images had been updated.

As I monitored the -shm, -wal, and database sizes, I opened the image tab and got an error message: “Error loading items beginning transaction: database is locked.” The -shm file grew, and the -wal file shrunk to around 105 GB. I figured it might take some time to commit the generated phashes to the database, so I let it sit overnight.

This morning, I found that the -wal file was still 105 GB. The -shm file grows to around 220 MB, then resets to 0 and grows again, while the -wal file stays the same size.

Can you help me figure out what’s going on and how to recover from this? What steps can I try to resolve the issue? I’d really appreciate some advice.

[feederbox826]

sqlite3 stash-go.sqlite
> VACUUM;
> .quit;

make sure stash is not running. that should flush your WAL to your database, otherwise you might need Recovering Data From A Corrupt SQLite Database to forcibly unbork your database

[max_must]

Thanks for the reply!

Before your suggestion I decided to restart the container. That worked all by itself.

The restart took longer than usual but after stash was up the wal was emptied and the database accessible again.