N
Nigel Allen
Guest
Antipodean salutations I've been tasked with bringing a rather large system under control using subversion. The code side of it I have no problem with and have a mud map already of the path that changes will take when migrating through multiple workspaces (devel, test, system test, preproduction, production). I have to admit some puzlement though as to how we can control any database changes that need to be migrated along with code changes that are implemented. Some of the puzzles I'm trying to solve are based around: How do we ensure that only relevant schema changes are migrated forward? How do we roll back schema changes if testing fails or if requirements change? How do we deal with data during database changes? I'm thinking about maintaining a copy (empty) of the databases from each environment and doing incremental database updates but this only solves some of the challenges. Can anyone share how they have handled this type of scenario with a hand-rolled system? (And please don't say "Roundtable" - if I were able I'd install it in a flash). Thanks and Regards Nigel.
Continue reading...
Continue reading...