Well, I could not understand what you meant by 'disk array'. If this disk is in an array like EMC or XP1024, this is a whole different thing to figure out how to fix it. But if it is a disk in a locally connected disk array, like an SC10 or similar, and you are using your LUNS as whole disks and only that way (no LUNS spanning over multiple disks etc, otherwise you are in the EMC and alikes situation), then you run
ioscan -fnC disk | more
and hunt for the NO_HW device. Find where this device is located on your array. Swap it with a good, working disk.
assuming it is in vg01 volume group (if not change the vg01 below to the name of your volume group below)
vgcfgrestore -n vg01 /dev/rdsk/cXtXdX
where you will put the proper numbers where "X"s are, in the above command, referencing your broken disk. Then
vgchange -a y vg01
then if this volume was a mirror of another volume, remirror the stale volumes in vg01.
On the other hand, if you have something like an SC10 which uses bonded disk volumes to make larger virtual disks, you need to refer to the users' manual of the array, explaining how to fix it, as it may not be as trivial as what I explained above.
Hope this helps
________________________________
UNIX because I majored in cryptology...