It wasn't me. You can't prove anything.


2009-01-20

Diagnose This

Nat watches a bunch of the medical shows. One of them is House. I hate the show. I don't like the primary character. He is a needless pill-popping jerk. I like holding doctors to higher standards. Drunks.

I had a box that one of the graphic interfaces was not working properly. I ran through the entire setup for all our products and still no luck. I eliminated all our stuff. This took the better part of a day. I then switched this morning to stuff that is not our code. I reinstalled the video driver and regenerated the xorg.conf. It worked. Fixed the problem. If I had started there, I might have fixed the trouble in ten minutes. No one died. One of my fellow employees didn't have access to that machine as soon as she needed it. I wish now that I had started at the right place.

I had messed with the xorg.conf, but not replaced it. I hadn't reinstalled the driver. One of those steps did the trick. As it turns out someone else had taken a look at the X configuration as well. I wonder if our combined effort is what caused the problem to begin with.

How do you know where to begin? How do you know when to give up and start over? How do you know when to declare the patient dead? This box is the third I tried to get to work. One machine had the wrong slots. Another had bad hard drive configuration. This one had a robust enough motherboard and plenty of ram and drive space. I'm committed to getting something to work.

Thank goodness the people who asked me to set this system up allowed time for problems. Knowing all those answers is called experience. I'm getting experience every minute I'm at my current job. When I was at the old huge employer a couple years ago, they had everything documented and procedure-written out the wazoo. It sapped people's innovation and stifled creativity. Everything was predictable. There are arguments and advantages both ways.

Next time. . . One of my buddies told me that this kind of error is preferable to a real "There is a fundamental flaw in the design." kind of error. I suppose he is right. I should learn and move forward. Besides, it wasn't our stuff that was messing up (Nvidia!).

No comments: