MySQL Voodoo

why is it that, when it comes to replication, MySQL administration is as much voodoo as it is science? Keep in mind that while I dont have a problem with this area of MySQL administration... many people do. The reason for this, i think, is that it's not a very fault tollerant system. Any sneeze can really screw up replication. When you throw in different table types, etc, you have a recipie for disaster.

The funny thing about it is that it's actually less difficult to list out the steps to resolution for the various problems (there are relatively few paths available for correcting problems) than it is to tell somebody how to go about figuring out which of those very few paths to take. Between hardware,OS, filesystem, network, etc, there are thousands of things that could go wrong... meaning that it's nigh impossible to tell someone "how to fix replication if it breaks." It's more or less "well i've seen something that FEELS a lot like this before... so I have a pretty good idea of where to start..."

Theres something you dont want to list on your resume! "Hi, my names Demitrious, and you want to hire me because of the valuable voodoo administration experience I will bring to your ops team!" Whats scary is that the only people I'd trust to work on my servers are exactly that -- voodoo administrators. Nothing scares me more than someone with a root password thinking like an M.C.S.E. >EG< with a worn out "in my perfect world" checklist of what-ifs.

Heh... "Microsoft, if it doesnt fit in our little box... you shouldnt be doing it!" and "Linux f it doesnt fit in our little box... make your own... just.. dont expect people to understand what the hell you're doing"

I wonder if I were to make a troubleshooting wiki if people in the tech community would get involved. It sure would be nice to have a place with common (or not so common) paths from ACKFOO to resolution...

Probably just a dream... Butif enough people tolf me they were interested... I would put it up in a heartbeat.

--DK