Client has a Dell Storage array, 10 hot swappable disks in a shelf which
is split so 7 slots are assigned to one SCSI port and the other 7 are
assigned to the other port. Each port is attached to a different server,
the two servers have a PERC (spit) controller each.
After moving the servers and shelf themselves (apparently we were too
expensive) they connected the shelf back to the servers but, crucially,
connected it up incorrectly, swapping the cables around. Unfortunately
they then decided to accept configuration changes on the PERC controller
which corrupted the RAID set in the shelf.
Fortunately for them, they stopped right there and called for help.
After realising what had happened I made the correct connections but the
RAID set was gone. Options at this point were to replace the disks,
create a new RAID set and restore from backup (14 plus hours of restore
and then integrity checking, it's an Exchange store) or send for data
recovery which would be scary expensive, think tens of thousands.
With the client's blessing I tried an old trick, creating a new RAID set
of the same size as the old set on the old disks. A quick reboot then
application of a free recovery utility and I had a partition that looked
good and crucially, a store that mounted. 8 hours of ESEUtil integrity
check later and the users were able to access mailboxes again.
Lesson learned by the client? The experts are cheaper than the monkeys,
even if they charge more.
--
Clint Sharp