Last week I came into the office to find that a drive on our domain controller had failed.  No big deal, right?  I overnighted 2 drives, and installed them both the next day; one as a replacement for the failed drive, and another as a hot spare.

I was hurried into work this morning by a voicemail from K.  She informed me that noone was able to log into the Exchange server.  I know I had everything running last night before I left, so this really caught me off-guard.

I rush into the office (fortunately I’m only about 5 miles from the church) and sit down to begin diagnosis.  My tests led me once again to the domain controller I had just worked on less than a week ago.  I tried to click on the start menu, and the whole system froze up.  Wow, the same symtoms that I experienced last week.  I winced, hit the reset button on the machine, and sure enough, the RAID controller’s BIOS shows a rebuild in progress.  Once the server was back up and running, I could see in the event log that the failure had taken place at 7:30

I was partially relieved to see that the drive I had just installed wasn’t the culprit, but was another one of the original drives for that array.  Wow…2 drives in under a week’s time.  These were both Maxtor DiamondMax hard drives that had failed on me, and they’ve been in use for less than a year.  The controller is a Promise Technology TX4310, and we’ve had no issues up to this point.

.

 

Another weird point is that the rebuild actually started BEFORE the machine locked up.  The hot spare stepped in just like it was supposed to, but the controller shows 5 device timeouts before the system gave up on it.

Hopefully this is a drive issue and not a controller issue.  Nonetheless, the task of getting a second DC up and running just went up in priority by many notches.