It’s strange that I find myself sitting down to write this post on what should be a completely obvious statement. Honestly I’m not trying to campaign for the Captain Obvious job.
I got an e-mail today from an old co-worker of mine that I hadn’t spoken with in some time. He reminded me of a situation that we had at a company that we both started at around the same time. They will remain nameless for obvious reasons, but I can tell you that they are a fairly large financial firm. When we started the job we were told that there had been a DBA working there prior to us being hired and that said DBA had taken care of the firm’s database servers for a number of years. We expected to find all the usual bits and bobs in place and we would just begin building a system of standards and roll it out to the servers we were responsible for. What we found what something quite different.
All of the firm’s primary production SQL Servers were being “backed up” by copying the .mdf, and .ldf files onto a file server nightly. When the IT Director told us about this system you can image the looks of shock painted on both of our faces. We both stammered trying to explain to him why this wouldn’t work. These servers were responsible for processing and recording financial transactions with amounts in excess of 100 million dollars.
So the thing my old co-worker wanted to e-mail me about was a new job that he just started where he has now run into the exact same situation.
I wish that I could say that that one time long ago was the only time I had ever encountered this problem myself, but sadly it wasn’t.
Look, I understand that backing up the databases properly and making certain that they are recoverable maybe isn’t the most interesting part of the job, but it is arguably your primary responsibility as a DBA. Before you get yourself involved in any other projects you should be designing a backup system that assures recoverability of the databases under any and all circumstances.