Convert mirrored raid drive back to stand alone? Stuck in SATA raid reboot loop!

[OutToLunch]

AAARGH. If I hadn't had a bad enough week already, this morning, 1 server's Intel matrix console showing failed drive in a mirrored set on a vanilla 2003 server install. Shut the server down, unplugged failed drive, plugged in a new sata drive, restarted, added drive back into volume set and left it rebuilding. A while later, about 80% through rebuilding, beep beep beep from UPS as power drops and it runs briefly before shutting down the server. As far as I know, this shouldn't do any harm to the working drive which the mirror was being rebuilt from?

 

As soon as the power's back on properly, restart the server again expecting it to continue rebuilding or restart from scratch on the new HD. I never get far enough to find out - the server is now rebooting immediately after the first POST screen before the raid controller option rom kicks in. If I unplug both drives, it will show the screen (with no drives, obviously) and continue to fail to boot since there are no drives present. If I plug the original working drive back in, reboot loop.

 

I'm not 100% familiar with where all the raid volume info is stored - I know I could plug the drive into another machine and rescue the data from the volume but I don't want to have to do a whole server rebuild Couple of q's for those more up on this than me -

 

If I clone the working drive, will that pick up everything it needs so that I have a working copy as an initial backup?

 

Is there any way to get the raid 1 info off and reset the drive to a single sata HD so I can attempt to get the server back up in a hurry then worry about reRAIDing later?

 

Any other ideas?

 

Motherboard is an Intel 865 chipset, onboard sata raid was being used (not my choice, stuck with the servers we had) so I'm not sure whether I could track down an identical one on ebay and whack the drive on that to continue if the raid chip has died somehow?

 

Already had a failed router, dead proxy and a million other things go wrong this week, not happy!

[Busybub]

Bit late but you never know...

 

As far as I know, this shouldn't do any harm to the working drive which the mirror was being rebuilt from?

 

Bad luck with the UPS

 

Shouldn't have done any damage unless you were unlucky enough that the shutdown caused a mechanical glitch on the working drive that was bad enough to corrupt a few critical sectors. Not likely though.

 

If I clone the working drive, will that pick up everything it needs so that I have a working copy as an initial backup?

 

Yes, assuming that the disc is undamaged.

 

Is there any way to get the raid 1 info off and reset the drive to a single sata HD so I can attempt to get the server back up in a hurry then worry about reRAIDing later?

 

Wouldn't have thought you would need to as RAID1 doesn't stripe data, thus no configuration data for the array, the card should just copy identical data from one drive to the second. In my experience of multiple RAID 1 failures (8 or 9 at the last count), I've always been able to bypass the RAID controller, plug the working drive in a SATA socket, and boot from it.

 

Any other ideas?

 

First, get a hard drive dock, check that you can actually see any critical data and copy it off the drive - better safe than sorry.

 

My usual approach to a RAID1 failure is to unplug the degraded disc, image the working one whilst still plugged into the RAID card, then obliterate the array and start from scratch. Plug the other drive back in (or use two new ones, keeping the original as a backup), create a new array, then spit out the image onto the working array.

 

However, this depends on having a server which is known to boot up!

 

Worth double checking that the BIOS on your server hasn't reset itself. It may have picked up the dead drive in the array rather than the good drive or array itself (either of which would have to be selected in order for the server to boot from the remaining disc if you are using a built-in RAID controller).

 

If you're having problems booting from the other drive, might be worth doing a repair on the server operating system to see if that will fix it - after you have cloned the disc or taken copies of the data.

 

Motherboard is an Intel 865 chipset, onboard sata raid was being used (not my choice, stuck with the servers we had) so I'm not sure whether I could track down an identical one on ebay and whack the drive on that to continue if the raid chip has died somehow?

 

Already had a failed router, dead proxy and a million other things go wrong this week, not happy!

 

I would check the boot order first.

 

Worst case, you can plug a working drive into a different chassis, and run a repair install on it to get it working. It will reset the server to a basic, bare install (you lose drivers, printers, updates) but any file data is left intact.

[OutToLunch]

Got sorted in the end, think it must be a wierd and wonderful glitch in the chipset on this board - for reference if anyone googles this up, I ended up plugging the failed drive back into the sata raid on its own which stopped the reboot loop and displayed the expected post POST raid config screen. Bizarrely after that it quite happily copied data from the 'failed' drive to a new blank drive to recreate the array!