RAID5 reconstruction problem

Brian DeFeyter (bdf@gospelcom.net)
Fri, 9 Nov 2001 14:18:34 -0500 (EST)


Last night I had 2 drives in a 6 drive array instantly fail and cause my
array to crash. I suspect that the channel these 2 drives were on was the
cause as these were the only 2 drives on that channel.

Here's my current /etc/raidtab for the failed /dev/md7:

raiddev /dev/md7
raid-level 5
nr-raid-disks 6
nr-spare-disks 0
persistent-superblock 1
parity-algorithm left-symmetric
chunk-size 32
device /dev/sdc1
raid-disk 0
device /dev/sdd1
raid-disk 1
device /dev/sde1
raid-disk 2
device /dev/sdf1
raid-disk 3
device /dev/sdg1
raid-disk 4
device /dev/sdh1
raid-disk 5

sdc1 and sdd1 where the failed drives. I used the 'fail one drive in
/etc/raidtab' trick on sdd1 to bring the array back which worked fine.
Then remarked sdd1 as a 'raid-disk' instead of 'failed-disk' and then did
a 'raidhotadd /dev/md7 /dev/sdd1' which started the recontruction.

About 1/2 way through, /dev/sdd1 died with the failure diagnostic led
blinking. I swapped it out with a new drive, paritioned it (Linux, not
auto-detect), formatted it, and then attempted a 'raidhotadd /dev/md7
/dev/sdd1'

Unfortunately, the raidhotadd appears to fail right away. I get a couple
hundred lines of kernel messages with 'DISK' and raidconf printouts. A gut
feeling tells me it doesn't like the superblock on the new drive. However,
I do know that the drive itself is fine since I manually mounted it and
ran a few tests on it before re-formatting it and putting it into the
array. If anyone wants all the kernel messages, I'll send them along.

Thanks,

- bdf

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/