I have a conspiracy theory: IBM puts time bombs in their RAID hardware. Sounds silly right? Here me out. For the second time in 6 months, I've seen this pattern with older ServeRAID cards/drives/enclosure.
- MSSQL Database file currupts itself. CHECKDB fixes it. Machine is fine for a few days. No hardware errors about drives.
- MSSQL Database file corrupts itself again. Reboot machine. Chkdsk /XRF finds no file system or sector errors. CHECKDB fixes the issues. Upgrade RAID firmware/drivers.
- MSSQL Database file corrupts itself again. Reboot machine. Chkdsk /XRF finds no file system or sector errors. CHECKDB fixes the issues. Install new RAID controller.
- MSSQL Database file corrupts itself again. Reboot machine. Chkdsk /XRF finds no file system or sector errors. CHECKDB fixes the issues. Move everything to new machine. Problems cease.
Again, I've seen this pattern TWICE. Both times, no SQL/Windows updates or patches were applied nor were any drivers/firmwares updated before the problems started for at least 6 months.
Dear St. Murphy: Die.
So?
Ron Savage on 2007-04-05T22:29:39
Industry standard behaviour should not surprise you.
