If you recently updated your mirror server to the 2022-05-16 schema change release, we’re sorry to say that a bug in our upgrade script caused aggregate genre and tag data (if you had imported any) to be deleted. If you need this data, it can be re-imported from a recent dump, and we’ve written a script to help automate that.
You can safely ignore this post if
- you’re not running a mirror server;
- you’ve never imported the derived data dump or don’t make use of genre or tag data;
- you’ve already done a full re-import of your database after the schema change release.
To restore the genre and tag data, follow these steps:
- Ensure you’ve replicated up to the most recent replication packet available. If you’re not sure, run
./admin/replication/LoadReplicationChanges. If you’re up-to-date, it should log “Replication packet … is not available.” - Run
git checkout production && git pull origin production. - Turn off any cron jobs that update the database, including for replication.
- Run
./admin/sql/updates/20220720-mbs-12508.sh. - Restart any cron jobs that you disabled.
You can verify that this process worked by checking the number of tags in the database: echo 'SELECT count(*) FROM tag' | ./admin/psql READWRITE. It should be over 200,000.
Sorry for the inconvenience, and let us know if you encounter any further issues.
MetaBrainz Blog 