On Wed Sep 23, 2009 at 09:56:52AM -0400, Sunpyo Hong wrote:
> I actually know the physical order of each HD. I was able to pull them out
> of the NAS in the order specified in the NAS. (Each HD enclosure was
> labeled) In the actual sata ports, this is what the hds are in.
>
> SATA 0: HD1
> SATA 1: HD2
> SATA 2: HD3
> SATA 3: HD4
> SATA 4: CD-ROM
>
> I think that linux reads the sata ports like.. 0 = sda, 1=sdb.. etc. So I
> assume that HD1 = /dev/sda (missing), HD2 = /dev/sdb, HD3 = /dev/sdc, HD4 =
> /dev/sdd so a create command should look like this:
>
> #mdadm -Cv -level=5 --raid-disks=4 missing /dev/sdb4 /dev/sdc4 /dev/sdd4
>
> This is exactly how I wrote the create command. Again I knew the physical
> order of the Raid and put them together in that order. Tell me if I'm doing
> something wrong.
>
> I'll check out testdisk as well..
>
The array order detected by the initial --assemble is (unless you have
incredibly strong reason to believe otherwise) most likely to be the
correct order, in which case the correct create command should be:
mdadm -Cv -l 5 -n 4 /dev/sdd4 /dev/sdc4 /dev/sdb4 missing
I'd suggest re-creating the array in this order (ignoring the irrelevant
"physical" ordering) before attempting any other recovery.
Incidentally, does the --create command report that any of the disks
contain an ext2/3 filesystem? This is usually reported for one of the
drives.
However, given that the initially-assembled array failed to mount, I
suspect you're hosed, and that whatever the ShareSpace did in attempting
to rebuild the array has actually broken it completely.
Cheers,
Robin
--
___
( ' } | Robin Hill <robin@xxxxxxxxxxxxxxx> |
/ / ) | Little Jim says .... |
// !! | "He fallen in de water !!" |
Attachment:
pgp0QY5UPQbED.pgp
Description: PGP signature
