LinkedIn | FaceBook | Twitter
I think half of what I do as a DBA is "spelunking" - exploring the situation at hand to see what the real problem is. Nine times out of ten, the thing that was presented to me isn't the thing I end up working on.
So where do you start with this process? Well, I normally grab as much info as I can on the situation, and start peeling back the issues to find the root cause. Lots of tools can help you do this, like PBM, the Management Data Warehouse, and even PowerShell, but before I touch the tech I make sure I know just where the issue is. I know, it sounds like the 101 thing to do, but in the heat of the moment it's often too easy to fall back on "doing something" before I really know what to do. I use OneNote a LOT, so I put a troubleshooting checklist in there and I use it religiously.It's saved my bacon multiple times.
The technical "stuff" can come easy if someone has the capacity to learn.
The "soft" skills are often overlooked and made overly complex by hunting for the toughest first. I recently posted something about this on my blog. It's a pet peeve of sorts.. Boiling down an issue to a business problem or simple problem statement is a key first step.
Anyway, great blog. Are you ever willing to share your checklist? I have one that I normally run through but always helpful to see what others look at. Helps build dialog and find holes in our own "spelunking" processes.
The temptation to just try something is pretty strong.
I've worked with engineers a couple times. They're problem strategy is definitely a trial and error. And in a lot of cases it's effective for what they do. They'll try to jiggle the cable or check the fuse and it does the trick (Okay, fine, what they actually did was a lot more sophisticated than that).
With software, finding the "real" problem often involves being more of a Sherlock Holmes than a random tinkerer (Gepetto?). I'm always inspired by Richard Feynman who impressed people when he was young "He fixes radios by thinking!" http://www.cs.cmu.edu/~pattis/misc/feynman.pdf