2-node NW65SP6 (NSBS6.6) cluster with NW65 server as iSCSI target. More
disk space was needed; however I managed to buy 2 new HP 72GB disks that
both failed just over 1 week later, leaving the RAID5 array U/S.

So now I need to rebuild the server from scratch, then restore the data
from tape (I just hope the tapes are OK). The plan is:

1. Make another server the master of root;
2. Remove the failed server and objects from the tree;
3. (Re)Install the iSCSI target server into the tree, selecting to
install the iSCSI software;
4. Apply SP6 and post-sp6 patches;
5. Create the iSCSI partition and assign trustee rights to the node servers;
6. On the node servers, edit the ION.NCF files so they connect to the
iSCSI partition (I assume the name will have changed) and create pools
and volumes;
7. Install backup software and restore data.

The bit I'm not sure about is step 2. Usually, to remove a server from
the tree I would remove directory services from the server; however the
server isn't bootable so I can't remove it that way. TID 3338221
"Removing a Crashed Server from the NDS Tree" does not mention NW65 - is
this still the recommended procedure?

Will the cluster resources still work correctly? I don't see any reason
why they shouldn't, but then I haven't done this procedure before.

Any other input appreciated.