If you are responsible for managing systems, you should have some sort of disaster recovery plan. Even if you are only managing the one system you carry with you on a regular basis, you should ask the question: What would it take to destroy this data?
It’s a good question, and in the post I’ve linked above, a technologist talks about some of the failings of disaster recovery plans because they forward engineer plans to recover systems. People think of specific problems and try to prevent them. They don’t reverse engineer to find out what events would cause them to lose their system.
That’s really the key to a lot of design and architecture in computer science. You can’t think about just what you expect to happen, or even what you want to happen. You have to consider the other ways in which events could occur or the issues that could cause an problem in your system. In development, the things you expect are considered to be the happy path. A good architect or engineer will think about everything else in addition to the happy path.
I sometimes think that far too many decisions are made while considering the happy path, but ignoring or discounting the other paths available. These other paths may be unlikely to occur, but having worked with computer for over 25 years, I can tell you that I often find the least likely events occurring far too frequently.
As the author says, you should never say “I never even thought of that happening.” Consider reverse engineering problematic situations and then make a judgment of how likely is it that any of the events will occur. That way you have at some idea of what could go wrong and what events you are willing to protect against.
The Voice of the DBA Podcast
The Voice of the DBA podcast features music by Everyday Jones. No relation, but I stumbled on to them and really like the music. Support this great duo at www.everydayjones.com.