On 2011-11-08 16:48, Chris Purves wrote:
Hello,
I have a software RAID 5 consisting of 5 disks. One of the disks (sdc) failed
and could no longer be recognized by the BIOS. I replaced it with a new disk
and while the array was rebuilding, another disk (sdd) experienced a read
error, which caused the disk to be reset and the array to be taken down. I was
able to determine that disk sdd has bad block, which when read, brings down the
rest of the array. I am able to assemble the array in a degraded state without
sdc, so as of yet I have not lost any data (other than the bad block), but I
need to be able to rebuild the array using sdc before I can replace sdd.
What I have tried:
The first thing I think I need to take care of is the bad block on sdd, so I cloned sdd
onto another disk using ddrescue and plugged the sdd sata cable into the clone and tried
to assemble the array using "mdadm --assemble --force /dev/md1". This does not
work. mdadm assigned the sdd clone as a spare and says it doesn't have enough disks
available. If I connect back to the original sdd I can assemble the array, degraded, but
functioning. So mdadm doesn't accept those two disks as the same.
Help, please.
After discussing on the linux-raid mailing list, I tried again to clone the
disk with the bad block:
ddrescue /dev/bad_block_disk /dev/new_disk
Once that completed (and reported only 4 KB of bad sectors) I unplugged the bad
block disk and ran:
mdadm --assemble --force /dev/md1
This time it worked and brought the array up in degraded mode with four disks.
Perhaps I hadn't stopped the array the previous time or something stupid like
that.
--
Chris Purves
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/4ebc6a11.10...@northfolk.ca
