RAID set 'cannot rebuild'

[Sprout]

I have a A7N8X with the built in Silicon Image RAID controller.

Unfortunately, I've had a problem with my poer supply, which caused
the PC to switch off without shutting down correctly. This has
corrupted the RAID set.

Now when I go into the RAID configuration screen I get this...

* Set0 SiI Mirrored Set <PM> 78166MB

1 Maxtor 6Y080M0 Current
0 Maxtor 6Y080M0 Dropped

When I setlect the option to 'Rebuild Mirrored Set', I get the message
'You cannot rebuild !'

How can I rebuild the RAID set? - obviously, I don't want to start
from scratch and reinstall the OS.

 

[E.P. van Westendorp]

I'm using RAID1 for about 7 years, in recent years via the build in
Promise chip of an Asus A7V133 motherboard and a Promise PCI RAID card
used on another board.
I would delete the current array, create a new RAID1 array and write the
contents of the undamaged disk to the disk of which the contents is damaged.
In my experience all brands of RAID offer this option in the RAID
configuration when creating a new RAID1 array.
Take care you copy the right disk, take your time!!!

 

[Roger Hamlett]

I'm using RAID1 for about 7 years, in recent years via the build in
Promise chip of an Asus A7V133 motherboard and a Promise PCI RAID card
used on another board.
I would delete the current array, create a new RAID1 array and write the
contents of the undamaged disk to the disk of which the contents is damaged.
In my experience all brands of RAID offer this option in the RAID
configuration when creating a new RAID1 array.
Take care you copy the right disk, take your time!!!

Yes.
I think the problem here, is that the second disk has been 'flagged' by the
system as faulty. Hence it is not considered a fit target for a rebuild. If
the drive was faulty, the normal proceedure at this point would be to
disconnect the drive, and attach a new one, then rebuild to this. It is this
proceedure that the user has to 'mimic'. I think I'd probably disconnect the
second drive. At this point the RAID software, should report only one drive
present, and you can verify that this is the 'good' one, before proceeding
any further. I'd then attach the drive to a second machine, and use
something like 'wipedisk', to remove the flag (clearing everything). At the
same time, I'd run a surface check to really test that the disk is not
faulty. If the disk passes, it can then be returned to the original machine,
which should now see it as a 'new' disk, and allow a rebuild.

Best Wishes

 

[E.P. van Westendorp]

I don't think there will be any problem when using the corrupted disk as
a target for the copy procedure when creating a new RAID1 array.
It doesn't matter what is on that disk, everything will be overwritten
anyway.

I use mentioned procedure monthly because I rotate three disks for added
security.
Each of my computers has three mobile racks, one is not connected and
used as spare position, to enable quick change of harddisks.

Disk number 0 in the array becomes spare for a month.
Disk number 1 becomes disk number 0.
Disk that has been spare becomes disk number 1 in the array.

 

[Roger Hamlett]

I don't think there will be any problem when using the corrupted disk as
a target for the copy procedure when creating a new RAID1 array.
It doesn't matter what is on that disk, everything will be overwritten
anyway.

I use mentioned procedure monthly because I rotate three disks for added
security.
Each of my computers has three mobile racks, one is not connected and
used as spare position, to enable quick change of harddisks.

Disk number 0 in the array becomes spare for a month.
Disk number 1 becomes disk number 0.
Disk that has been spare becomes disk number 1 in the array.

I'm not so sure.
I have seen a couple of RAID controllers, where if you 're-attach' a drive
from another array, the disk ID is still recognised, and a 'corrupted' flag
is also still seen, and is not cleared if the array is deleted. Also
(realistically), though the original poster blaims a power fluctuation for
the problem, I would not feel happy to re-attach a disk that had been
flagged as bad, without carrying out a full surface scan. Basically, there
must have been power to the drive for the flag to be set, and it is possible
that the fluctuation has coincided with, or triggered a drive problem.
I have also seen controllers where if you create an array, the partition
table is destroyed. Hence I feel it is safer to make conditions be as close
to the situation where a drive has failed, and simply been replaced, without
deleting the array, and re-creating.

Best Wishes

 

[Sprout]

I'm using RAID1 for about 7 years, in recent years via the build in
Promise chip of an Asus A7V133 motherboard and a Promise PCI RAID card
used on another board.
I would delete the current array, create a new RAID1 array and write the
contents of the undamaged disk to the disk of which the contents is damaged.
In my experience all brands of RAID offer this option in the RAID
configuration when creating a new RAID1 array.
Take care you copy the right disk, take your time!!!

Thanks for your reply.

If I delete the existing RAID set, then create a new one, will it get
rid of the data?

 

[E.P. van Westendorp]

The Promise menu also shows a choice "Delete Array".
This doesn't mean the data on the disks will be deleted but after you
use this choice the disks are no longer marked as belonging to an array.
Therefore you can create a new one.

I have only actual experience with RAID from three manufacturers, yours
could be different!

 

[leia176]

I have the same board as you and deleted an array. No the data is not lost,
but make sure you copy the correct disk. I made sure by writing the serial
#'s down and unplugging one..shutting down etc. Then I knew what drive was
connected to what port and which was the good one.

When you recreate the mirror it will ask you if you want to keep the data or
not so pay attention and answer correctly. Then choose the right disk to
have the data copied from.