D
dbeavon
Guest
We stopped using DF's and migrated to SQL deltas some time ago. But the idea is basically the same. Maybe the use of the SQL-style deltas is where we took the wrong turn. Maybe we need to take a step back, and just talk about applying a change to your own development database (the first-ever place you are creating your DF's). Lets say you modify the database which your are actively working with. Let's say you add a column. And you delete any previously created r-code. Do you then have to bounce the entire database to avoid errors in your applications? Or do you just restart client applications? If you aren't bouncing the database, then are you working exclusively with "shared memory" clients? That's all I'm really talking about. The "remote servers" are the things that seem to choke up on schema changes. The need to restart the database for the sake of "remote servers" is very odd. It is defeating the whole purpose of the "online schema changes". We can also stop and restart any of our databases in less than 15 minutes, but that shouldn't be necessary.
Continue reading...
Continue reading...