During the software development process, whether you’re cutting edge Agile developers, or traditional waterfall method folks, your code changes are periodically accompanied by database changes. For instance tables have particular rows and columns. When developers add new columns (ie fields on an index card), or create new tables, relationships, indexes or other database objects, all of these are lumped together as database changes.
Version control systems have brought great manageability gains to software projects, even ones involving only a single developer. That’s because they allow you to rewind to any savepoints, just like you can undo and redo changes in a word document. With database changes, however the picture because more muddied.
Database Change Management Best Practices
During deployment, operations folks and/or a DBA must still have their fingers on the trigger. Some frameworks such as Ruby on Rails include Migration scripts. Do not for example allow rollback scripts to run automatically. This is a recipe for disaster.
Above all use common sense, and always second and/or third guess yourself. Better to be safe than sorry when juggling your crown jewels.
Quora discussion by Sean Hull – What is database change management and why is it important?