3

I am expanding my raid6 array from 18x2Tb to 20x2Tb disks and after 3 weeks status shows:

Logical device Task:
   Logical device                 : 0
   Task ID                        : 105
   Current operation              : Reconfiguration
   Status                         : In Progress
   Priority                       : High
   Percentage complete            : 11  <——— !!?

I.e. I need to wait another 30 weeks for it to complete... Is there any way I can cancel expansion and revert the array to its original state?

From my past experience adding just 1 drive to the same array took only 3-4 days, I never expected that 2 drives would take that much longer.

user1559834
  • 139
  • 1

1 Answers1

3

Parity RAID is simply not meant to have that many high capacity drives in it, and my guess is that you're stuck at 11% after three weeks because the array is failed or has encountered an Unrecoverable Read Error (URE) at that point. (At the same time, it's not outside the realm of possibility that a 20 disk, 2TB array might take the better part of a year to run double parity calculations on... this is frankly one of the crazier things I've heard all year, and considering where I work, that's quite the accomplishment.)

Anyway, the good news is it probably won't take 30 more weeks to complete, but the bad news is it will be stuck in that state forever unless you do something about it. Oh, and it might be considered bad news that your array's probably hosed.

Consider it a learning experience on designing an array, and RAID in general.

As to what the best course of action is at this point, I'd hope Adaptec would know, though, as you can see on the product page, you might have to pay for a support case, depending on the age and warranty status of your card.

gravyface
  • 13,947
  • 16
  • 65
  • 100
HopelessN00b
  • 53,385
  • 32
  • 133
  • 208
  • The disks weren't degraded when the expand started, so an URE would be needed in the same place on three different of the old disks to actually lose data. But bad sectors definitely sounds like a good conclusion - maybe even one of the new disks. – Shane Madden Nov 11 '12 at 19:18
  • Thanks for your insights guys, I've contacted Adaptec, waiting for their response now. But isn't controller supposed to throw some exception/error message in case of URE or how are we supposed to find out that something bad has happened to the RAID? Status still shows that everything's healthy. – user1559834 Nov 12 '12 at 07:50
  • Last night rebuild progressed to 12%... really confusing. – user1559834 Nov 13 '12 at 10:00
  • @user1559834 Well, in that case, maybe it is just going to take 30 weeks to run the parity calculations. Really, call Adaptec, find out if you can cancel the operation, and hopefully you'll be able to rebuild the array in a more sane configuration. – HopelessN00b Nov 13 '12 at 14:37