3.48.0 fails to start: Caught std::exception: Sqlite3: INSERT INTO track_backup SELECT id, version, scan_version ...

[yurivict]

lms-3.47.0 worked fine, but 3.48.0 fails to start, and writes this into the log:

[2024-Jan-29 22:58:44.398] 87559 - [info] "config: reading Wt config file: /var/db/lms/wt_config.xml (location = '/usr/local/bin/lms')"
[2024-Jan-29 22:58:44.402] 87559 - [info] - 0x1f39b9e12000 [UTILS] Starting IO context with 8 threads...
[2024-Jan-29 22:58:44.402] 87559 - [info] - 0x1f39b9e12000 [DB] Creating connection pool on file /var/db/lms/lms.db
[2024-Jan-29 22:58:44.423] 87559 - [info] - 0x1f39b9e12000 [DB] Preparing tables...
[2024-Jan-29 22:58:44.426] 87559 - [info] - 0x1f39b9e12000 [DB] Database version = 50, LMS binary version = 51
[2024-Jan-29 22:58:44.426] 87559 - [info] - 0x1f39b9e12000 [DB] Migrating database from version 50 to 51...
Caught std::exception: Sqlite3: INSERT INTO track_backup SELECT id, version, scan_version, track_number, disc_number, total_track, disc_subtitle, name, duration, bitrate, date, year, original_date, original_year, file_path, file_last_write, file_added, has_cover, mbid, recording_mbid, copyright, copyright_url, track_replay_gain, release_replay_gain, artist_display_name, release_id, 1 FROM track: NOT NULL constraint failed: track_backup.bitrate
__cxa_thread_call_dtors: dtr 0x82cb9e3d0 from unloaded dso, skipping

FreeBSD 14.0

[epoupon]

Could you try to full rescan in 3.47.0 first and then upgrade?

[flroppy]

I had the same error when upgrading from a very old version, and doing what @epoupon suggested of doing a rescan in 3.47.0 and then upgrading fixed it. Thanks!

[epoupon]

Ok thanks for the test.
Will fix that clunky migration for the next release.

[yurivict]

lms should maintain the DB schema version and it should be able to upgrade DB correctly by itself.

[Maxmystere]

In case it's needed here is the error output

sqlite> pragma integrity_check;
wrong # of entries in index listen_user_track_backend_date_time_idx
wrong # of entries in index listen_track_user_backend_idx

Following this website doing REINDEX; fixed the crash then I did a full rescan