CodingHorror has a blogpost titled Get Your Database Under Version Control
It is suggested that you also keep the data in version control. I am reading through the comments and I am amazed. One person writes "Once you have used ActiveRecord Migrations it is very hard to go back!"
This might work for a small or medium size database. My database is well over a terabyte, how would you keep that data under version control? Data gets modified every day. Some of the data is encrypted. Some of the data gets inserted into audit tables. Data gets replicated to other servers/databases. There are jobs that pull in data from real time systems every second. I do have different versions of DBs on staging and QA servers but only one on the production server. Changes have to go through change management, you have to open a ticket to do a change. This is not something you would do on a daily basis.
What is your opinion? Do you have the schema in version control? What about the data itself?
