[MUSIC PLAYING] [WHISTLING]
Can you give me three considerations for planning a successful database migration? Just three. What are your top three?
Well, the first one is going to be just that. Planning. Spend most of your time in planning. Plan to spend most of your time in planning. Figure it out ahead of time. Do all the planning and find all the gotchas and all of the things that are going to bite you ahead of time.
And the next one kind of goes along with that, and that's modeling your data before you move it. Make sure that you have at least a logical model so you understand the relationships between the data and you can build physical models that reflect both the old and the new environment, and you're going to uncover any hidden dependencies or any things like that before you actually start into your migration.
And the third piece, third thing is test, test, test and test again. If you go into a migration and you say, well, based on some copies I did two years ago, I think the copy might-- I think it's going to take me six hours to copy this terabyte of data from one database to another. Right? And you get into your migration and it's a Big Bang migration and instead of taking six hours, it takes 12. Well, now what do you do?
First of all, you better have planned for it. And secondly, if you are planning and testing, you would have known that it's going to take 12 hours instead of six. OK, I'd better plan for that 12 hours instead of six. So they all kind of tie together. But again, planning I think is the most important.
And the more time you spend planning, the less time you'll spend doing and actually making it happen. And the actual make it happen is going to be. yeah, OK. We planned for that. We know that. We knew that was going to happen, or we knew this was going to go on, and that was going to go on. And then you move forward.
I got to say, I was debating on whether you're going to say testing or automation on that last one.
Yeah, well, automation is important, and it's good to automate everything you can. But the first thing is you can't automate chaos. Right? If you haven't planned right, and you don't understand your environment, if you haven't modeled your environment so you understand it, automation is not going to do you any good. Same garbage you had before you're going to have in the new system, right? And you don't want that. You want to have at least a little bit of improvement and a little bit better system. So automation is good, but that's going to come after, I think, planning and modeling and probably testing is.
02:58