I like the aggregation idea. If you had a table with an entry for all other tables (perhaps, even itself) and it contained a checksum of some aggregate of the respective tables, you could be reasonable sure that the databases were in sync. The updating of the checksum table should be an ongoing process to minimize checksumming each entire table each time.
in reply to Re: Verifying 2 databases are in sync with brute force
in thread Verifying 2 databases are in sync with brute force