NeilBrown wrote:
This typically happens because /dev/sda and /dev/sdb have md superblocks
on them and mdadm attempts to assemble an array from those and in the
process tell the kernel to forget about the partitions.
Can you:
blockdev --rereadpt /dev/sda
blockdev --rereadpt /dev/sdb
mdadm -E /dev/sda* /dev/sdb*
and report the result?
Hi Neil,
Thanks for the reply, but due to time constraints I have had to abandon
this, although I would be interested in your comments on what I found
after this my mail and before giving up.
The 2 1TB Seagate drives were taken directly from a 4 drive Windows
Vista striped (RAID0) array - they have never had md RAID installed.
I discovered that for whatever reason /proc/partitions, although it
showed entries for sda and sdb, it did not have entries for the
partitions sda1 and sda2 and I guess this is why mdadm had trouble.
Thinking it might have been an fdisk limitation, I tried partitioning
using parted - same result.
I then tried building using sda and sdb, instead of partitions and this
time the mdadm errors were along the lines of "device busy" for both
drives - even after rebooting.
At this point I gave up. The drives were attached to an Adaptec PCI SATA
card, which along with 2 SATA drives on the onboard controller and a
PCI-X SCSI adapter with 2 drives attached, seemed to cause much
shuffling of sd devices during bootup, so I am wondering if this
contributed.
Regards,
Richard
--
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
