In no particular order...
9. Good old boys
8. "It's just like a spreadsheet" database design
7. Out of date statistics
6. Improperly partitioned tables (see 7 above)
5. Good old boys
4. Microsoft patches
3. Vendor software
2. Internally developed software
1. Good old boys
I couldn't believe the number of disk related problems that kept coming to mind while putting this nod-disk related list together - especially when considering the SAN as a black-box hard disk.
You may have noticed that Good old boys made the list slightly more than once. Unfortunately, that is based on years of direct observation. I'm even going to manufacture a statistic to support its frequency in the list:
Nearly 60% of production database outages are, directly or indirectly, caused by the incompetence fostered by Good old boy cronyism... and that's NOT the disk.
No comments:
Post a Comment