If you designate one database as the 'master', and the other as the 'slave', then as long as the master is up to date, then the easiest way to keeping the slave up to date would be to periodically dump the data from the master, and then load it into the slave. This would of course, require a fair amount of time - but probably less than checking line by line. Are you going to be using both databases in tandem (ie., in a cluster?) or is one going to be a backup? The dump/load method is probably the best for a backup. And I think there's software for working with dbs in a cluster. Or you could use PostgreSQL, which has lots of advanced features (you could write a quick PL/pgSQL triggered function to do it, for example).