Wednesday, April 25, 2012

Reactivate Mirror immediately caused Failed Redundancy?!

WINDOWS MIRROR IS NOT SO SMART
I just spent a frustrating couple of days trying to upgrade the SATA controller on my Windows 7 workstation.  The controller upgrade went smoothly but, in the process, my existing software mirror (RAID-1) broke and I could not re-establish it!  Everything I tried to google was unhelpful, i.e., I never found anything beyond, "Delete the mirror and recreate it."

Both drives were still visible and with identical partitions/folders, but Windows Disk Manager reported "Failed Redundancy."  As soon as I saw there was a problem, out came the external hard-drives and I spent those first 24 hours updating my offline backups.   I then tried breaking the mirror, deleting the volumes on the 2nd drive, and then re-adding them to a new mirror.  No go.  No matter how I tried to do it, the new mirror immediately came up saying, "Failed Redundancy."  Checking the Windows Event Viewer merely reported, "FT Orphaning : A disk that is part of a fault-tolerant volume can no longer be accessed."

FYI - I was also SUPER CAREFUL to consistently remove only the 2nd drive, i.e., so that I didn't accidentally nuke the copy I was keeping.  (The original partitions on that 2nd drive had long-ago been nuked.)

MORE THINGS I TRIED

  • The only suggestion I found online (that I hadn't already tried) was to click Reactivate Volume on the newly recreated (and 'failed') mirror.  But this simply triggered a new "FT Orphaning" error in Event Viewer.
  • I had originally upgraded from a 2-port controller to a 4-port Sil-3114 controller.  Out of the box the 3114 was running the latest/final v5.403 RAID firmware.  When my continued attempts to re-add the mirror kept failing I began to suspect the card.  So I re-flashed it with the non-RAID firmware, v5.500.  (I was able to do this from within the Windows driver!  Very cool. I had to install the non-RAID driver afterwards, too.)  Unfortunately this didn't help.
  • I had already tried deleting everything on the 2nd drive from Windows Disk Manager but I wondered if maybe there was some left-over MBR data that had to be deleted.  So I booted a copy of FreeDOS and ran the following command:  fdisk /clearall 2
    Unfortunately, that didn't help either.

WHAT FINALLY WORKED :-)
I have no idea why but re-arranging the drives' positions on the controllers did the trick!  Specifically, I took one of the drives connected to the motherboard (not my boot drive!) and swapped it with the 2nd mirror'd drive.  Other than that I made no additional changes to the drives, partitions etc.

It's disappointing, really, that after all these years the Microsoft software mirroring system still has such deficient error-reporting.
_______

3 comments:

Arik said...

Hi Toni,

Did one of these disks was part of a Sil-3114 RAID before you tried making it a windows software raid?

Did you try clearing the first disk or its mbr? Or only the second disk?

Regards,
Arik.

Technical Tony said...

@Arik,
I never used the Sil-3114 "hardware" mirroring. I had everything setup and working properly via Windows (software) mirroring. The only reason I changed anything was that I wanted to install more drives.

Maybe the Sil-3114 numbered the drives differently, somehow, and the problem always ever was the order?

Arik said...

Hi Toni,

I tried connecting those drives to the on board controller and making a mirror. This time I got "All disks holding extents for a given volume must have the same sector size". http://support.microsoft.com/kb/2714829

Seems that one of the disks is "advanced format" disk and the other is a regular/legacy one. Still trying to decide what to do but maybe this is something that is relevant.

Arik.