I've got a web application project that actually works with quite large database (over 5GB). Data within the database is partitioned by project. Each project takes roughly 1GB, which is a small looking for application to operate (we all do some math information that spread across this dataset and getting rid of an element of the dataset isn't a choice).
As part of the daily build we deploy our application right into a test atmosphere. To do this, a builder runs a custom DB update utility that updates the present DB towards the appropriate version. But we have essential to help keep all of the daily develops to ensure that QA team could "go in time" and compare calculation is a result of different develops. Around we attempt to possess a backward compatible data schema it is sometimes very challenging and time intensive. So, hence the question:
What method of the database versioning can you use should you have had essential to help keep previous daily develops ready to go and want to handle large database when you are performing daily develops?
We use SQL Server 2005 and ColdFusion with a few Java in front-finish whether it makes any difference.