For a really quick solution, you could dump each table to a CSV file (ordered by primary key) and then run diff against them.