"The new Netware 6.0 service pack is here! The newNetware 6.0 service pack is here!"
You would be happy too if you had the afternoon I had. This should fix the crash that starts the process that we haven't gotten to not lead to in at least a few minutes of the cluster going "Wha?" and today forced us to reboot the whole damn thing. It may not actually fix the meltdown issue itself, but if I can fix the trigger, that's at least a step in the right direction.
Do most people realize that systems administration, most of the time, is like trying to fix a car by going up underneath the car, sticking your hand up into the engine, and feeling around until you find something you want to change, and then try and replace it without the right tools? Oh, and did I mention it's completely dark and you're wearing mittens? You depend on experience and intuition, and if you haven't written the whole thing yourself you're at the whim of those who have (and don't give me that open-source crap--do you understand any of the code you've compiled beyond a cursory glance, really?) and you have to hope they have the same problems you do and know how to fix them.
Sometimes it's amazing things run at all. Knowledge is power and all that. Read another book, play with another website, look at logs and try and integrate it all in your head. That's all you can do, and put things together with some semblance of logic so you can at least say, "Hey, that's what broke." You may never know why, you may not care, as long as you can get another one of those that might be made better, or figure out how to fit it in better.
I get a little freaked out when I think about all the pieces of equipment I'm ultimately responsible for and that well over 99% of the time it just works. I guess I do know what I'm doing, largely, but not because I know what every single line of code and piece of equipment is doing.