Once CB-51 is complete we will need to create a script that will be called from cron periodically to dump the data.

This database dump system should work on the principle of whitelisting columns to be included in a public dump. Additionally, it should list known private columns and raise an exception if it encounters a column that isn't defined either as private or public.

The public data dumps should be compatible with the existing musicbrainz-server data dumps. They should contain:

1. Contain the CC license text in the file COPYING
2. A timestamp when the dump was created called TIMESTAMP
3. Contain a subdir called cbdump
4. Contain a one file per database table in the cbdump subdir, compatible with the file format of the musicbrainz data dumps. Each of these files should be the name of the database table
5. SCHEMA_SEQUENCE that contains the monotonically increasing number of the current schema version. The CB data dump importer, should refused to import a data dump that does not match its internal schema number.

To complicate matters, our database contains clearly delineated data from multiple CC licenses. To avoid confusion among our data consumers, we need to not conflate data from different licenses into one file. Therefore we need to multiple data dumps files – this requires more discussion. Please do not implement this ticket yet.