Linux software RAID 10 rebuild
Posted: Thu Sep 29, 2011 8:08 pm
I have a broken RAID 10 that I need to rebuild/reconstruct. It is a Linux software (mdadm) RAID 10, consisting of 3 drives (yes, 3, apparently you can do it in Linux), all drives are in good order. Stripe size is 64k and data distribution is like this:
A1 A1 A2
A2 A3 A3
A4 A4 A5
A5 A6 A6
Meaning you have two identical data stripes on two adjacent drives.
How do I recreate this using Virtual Block RAID in R-Studio? I tried
A1 IGNORE A2
IGNORE A3 IGNORE
with both 2 and 4 rows and it didn't work. How many rows do I need to define? Do I ignore repeated block or what? Basically, the way I see it, I need to get a block of data for a disk, skip the block of data from the next disk, get the block from the following disk and so on. Can it be done in R-Studio GUI or I have to create a definition file (the questions are the same - how many rows, do I ignore repeated blocks, etc.)
Any help woud be greatly appreciated
A1 A1 A2
A2 A3 A3
A4 A4 A5
A5 A6 A6
Meaning you have two identical data stripes on two adjacent drives.
How do I recreate this using Virtual Block RAID in R-Studio? I tried
A1 IGNORE A2
IGNORE A3 IGNORE
with both 2 and 4 rows and it didn't work. How many rows do I need to define? Do I ignore repeated block or what? Basically, the way I see it, I need to get a block of data for a disk, skip the block of data from the next disk, get the block from the following disk and so on. Can it be done in R-Studio GUI or I have to create a definition file (the questions are the same - how many rows, do I ignore repeated blocks, etc.)
Any help woud be greatly appreciated