It’s Chris Yates (b/t) birthday! (I think he just tured 19.) And in honor of his birthday we are writing blog posts! Well, technically it’s T-SQL Tuesday again, but we can pretend we are writing for Chris’ birthday. In fact that’s the subject he’s selected. “Give yourself a birthday present.” Since my birthday is in October I’m just going to call this a birthday present for Chris.
Default database file sizes have been updated!
The other day I noticed there was an #AskSQLTiger hash tag on twitter, so I thought I would ask about database defaults. For years they have been pretty unrealistic.
No one is going to use a file size of 3mb (data) and 1mb (log) unless it’s a very very simple test. And the autogrowth of 1mb (data) and 10% (log) has caused me more grief than you would believe. I mean by the time you are working with multi gb files these growth patterns just don’t work anymore. Just think about the number of growths at 1mb at a time it would take to get to a 10gb data file.
The Tiger team responded.
Well what do you know! SQL 2016 not only goes faster it has lots of awesome new changes too.
Now 8mb initial 64mb growth still isn’t all that big. Perfect for testing, or small lab environments, but not big enough for live databases in my opinion. But guess what? The awesome doesn’t stop here!
The default file sizes didn’t used to come from the Model database. This is no longer true!
To be fair when you are creating a new database you should pay attention to file sizes and file growth. On the other hand mistakes happen, people are lazy, etc. It’s nice to know that the safety net (defaults) have been updated to be a bit more realistic.