On Saturday February 4, oren@xxxxxxxxxxxx wrote:
> Hi all,
>
> I recently had a machine (Debian 2.6.7) with a raid array have a failure,
> the disk controller card died partially ( so a couple of disks went
> offline) then died permantly taking a further two drives out of the array.
>
> The array (/dev/md1) was made up of /dev/hda3, /dev/hdi3, /dev/hdk3,
> /dev/hdl3 and /dev/hdj3) the array was made up of the first three and then
> the fourth was added (I believe).
>
> Mdadm.conf looks like:
> ARRAY /dev/md1 level=raid5 num-devices=5
> UUID=aeb44e5d:14388f60:8c2f39ec:9488facd
> devices=/dev/hda3,/dev/hdc3,/dev/hdd3,/dev/hde3,/dev/hdf3
>
> Due to the controller death I've moved the disks across to another
> machine, which has relabelled the disks as hda, hdc, hdd, hde, hdf.
>
> I've attached the output of mdadm -E, below and have at least worked out
> what the mapping is from old to new. (ie hda-hda, hdi-hdc, hdk-hde,
> hdl-hdf, hdj-hdd).
>
> When I try to assemble the drives I either get a 3 disk array mounted or
> the following output:
> phineas:/etc# mdadm --assemble /dev/md1 /dev/hda3 /dev/hdc3 /dev/hde3
> /dev/hdf3 /dev/hdd3
> mdadm: failed to add /dev/hdf3 to /dev/md1: Invalid argument
> mdadm: failed to add /dev/hdd3 to /dev/md1: Invalid argument
> mdadm: /dev/md1 assembled from 3 drives - not enough to start the array.
>
hdf3 and hdd3 have bad checksums somehow, but otherwise look fine.
You should be able to fix this with
mdadm --assemble /dev/md1 --update=summaries /dev/hd[acefd]3
If that doesn't work, could just recreate the array:
mdadm --create /dev/md1 -l5 -c128 -n5 /dev/hda3 /dev/hdc3 /dev/hde3 \
/dev/hdf3 /dev/hdd3
NeilBrown
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
