A few words from Dave the Bloody Pulp after learning about NT software RAID

Last Updated on September 30, 2010 by Dave Farquhar

I’ve worked 26 hours so far this week. I’d say I hope there’s an end in sight, but then again I hoped to leave work at 3:30 this afternoon and that didn’t happen. I’m gonna go read a book and try not to think about work or much of anything else.
I learned something today. I learned that it’s possible for NT4 software RAID (mirroring) to cause a bluescreen and immediate reboot loop on bootup. Funny how the thing that’s supposed to make the server reliable can make it unbootable. One of my colleagues and I figured it out after about three hours of messing around with a downed server. An NT repair install failed; after loading a device driver for the SCSI card, it would refuse to see the drives. At one point we booted the thing up with only one of the drives present (accidentally, I think) and it started. And then we accidentally booted it up with the other one. And that was how we learned that the system was mirrored. The system was set up by some other consulting shop, in case you’re wondering why we didn’t know our own server.

And I’m sure that even a tyro (ahem) knows this about NT4 software RAID, but seeing as I’ve only ever done hardware mirroring, personally, with NT4 (and it’s been six years since I did software striping with it), it’s new to me.

Nothing in the event logs gave any indication what corrupted the mirror, but we broke and recreated the mirror and the server’s been fine.

And maybe this isn’t a tyro problem, because the configuration was unusual. There were two drives and two SCSI host adapters in the mirror. It should make it even more reliable that way, theoretically. But it didn’t turn out that way.

I’m gonna go read a biography of F. Scott Fitzgerald. That ought to take my mind of things.

If you found this post informative or helpful, please share it!