Database Errors - Recorder Purge sqlite

[datascope11]

Hello all. Wondering how to go about cleaning this up? Ever since I updated to 115 I’m getting these 2 error in my logs, which seem related. I have no special config for recorder or database, just using default HA.

Error purging history: (sqlite3.DatabaseError) database disk image is malformed [SQL: DELETE FROM states WHERE states.last_updated < ?] [parameters: (‘2020-08-05 14:12:23.189851’,)] (Background on this error at: http://sqlalche.me/e/13/4xp6)
Logger: homeassistant.components.recorder.purge
Source: components/recorder/purge.py:101
First occurred: September 23, 2020, 4:12:00 AM (3 occurrences)
Last logged: 4:12:00 AM

Error executing query: (sqlite3.DatabaseError) database disk image is malformed [SQL: DELETE FROM states WHERE states.last_updated < ?] [parameters: (‘2020-08-05 14:12:23.189851’,)] (Background on this error at: http://sqlalche.me/e/13/4xp6)
Logger: homeassistant.components.recorder.util
Source: components/recorder/util.py:43
Integration: Recorder
First occurred: September 23, 2020, 4:12:00 AM (3 occurrences)
Last logged: 4:12:00 AM

Doesn’t seem to affect my instance’s performance, but it bothers my mild OCD. I’ve read that the solution could be to just delete the entire home-assistant_v2.db file using File Editor, and letting it get rebuilt on a reboot. Is this correct?

I should probably read up more on maria or something…

Thanks!

[tom_l]

Deleting and restarting is the easy way, with the downside that you lose your history.

There’s some information on repairing the database in this topic if you want to hang on to your history:

DB image malformed. How to fix it?

Hi, after some hussel around sd card images (i.e. restoring snapshots) and some updates this errors occures: 2018-11-01 22:37:01 ERROR (SyncWorker_4) [homeassistant.components.recorder.util] Error executing query: (sqlite3.DatabaseError) database disk image is malformed [SQL: 'SELECT events.event_id AS events_event_id, events.event_type AS events_event_type, events.event_data AS events_event_data, events.origin AS events_origin, events.time_fired AS events_time_fired, events.created AS events_c…

[datascope11]

@tom_l

Deleting and restarting is the easy way, with the downside that you lose your history.

Thanks. This worked. History gone, but not used for much anyhow… Hope this isn’t a regular occurrence (database corruption…)

[tom_l]

Updates and backups seem to do this to the default db now and then. MariaDB addon running in its own container seems to be much more resilient, to updates at least.