Why Database Names Matter

There is a reason users shouldn’t be allowed to name their own databases. Actually, there are probably several thousand reasons. I have a new one for you–It makes my life more difficult.

Lets say you work for a company called Acme. And lets say your manufacturing people created an in-house Microsoft Access application with a SQL database called AcmeDatabase.  And lets say that your support organization created a database on a different server called Acme that’s part of their CRM system. On top of that, the ERP system (yet again on a different server) prefixes all four of its databases with Acme_.  As the DBA, you’d never permit this to happen unless you inherit this mess.

Welcome to my world and the mess I inherited. I have my company name (or some form of it) in countless databases. Lets make this just a little more interesting, though. We’re going to be consolidating database servers in the near future and potentially putting all of these databases on the same server.

These are databases that could potentially end up living on the same server:

• AcmeDatabase
• Acme
• Acme_App
• Acme_Data
• Acme_Forms

Which one s are ERP, CRM, or Manufacturing?

When we consolidate database servers, we’re also going to be renaming some of these databases. My application owners won’t be too happy about this. They’re going to have to get over that. I inherited this mess, and I’m going to make it right.

Rookie Mistakes: Named Instance

I hate rookie mistakes. I hate them even more when I’m the one making them.

My coworker was installing Microsoft OCS Server, and he said that he needed a SQL instance, and he told me what the instance needed to be named. Since we already had a SQL server dedicated to this project, I created a named instance on said server.

In retrospect, I should have looked a little more closely instead of just assuming I needed to create a named instance. After I looked at the screen that asked for the SQL instance, I knew that we could use the default instance and not have to go down the named instance path.

Fortunately, we’re still in the test phase with virtual servers and the servers we’re using will be replaced with physical machines. That means I’ll be able to go back and get it right the next time around.

That’s why we have test environments, right? I’d much rather make a rookie mistake in a test environment than in production.