Situation (Community College with almost 24 x 7 requirement):

4 CLUSTERED servers, Compaq DL380 G2 and G3 (2 of each)
MSA1000 (server FC HBAs are Compaq/HP using QL2300 driver v6.80.06)
HP SAN Switch 2/16V
NW6.5 SP2 with NSS2b
eDir 8.7.3.4
GroupWise 6.5 SP4
GroupWise WebAccess
GWAVA 3.52
NetStorage
NetMail 3.5.2
NDPS/iPrint

Project (minimum) - do all 4 servers:

NW6.5 SP3
OS3a patch
NSS3a patch
NS patch

Probable extension to project - do all 4 servers:

TCP659J
iPrint - either 12b or 12c (???)
NW65SP3NCP
Syscalls3a
65AFP239 (from v2.02.02 to v2.03.09)
Wsock6g

Plan:

1. Move clustered resources away from one server (starting with Master of
Root).
2. "Cluster leave" on target server.
3. ??? Comment out clustering startup on target server ???
4. Apply minimum list above, restarting server as required.
5. Once complete with minimum, add node back into cluster and start on next
node using
same plan as first node.


Questions:

1. Critique of plan?
2. Any harm in clustered (physical) node access to NSS volumes (cluster
enabled) on SAN with temporarily different NSS versions on various nodes
during the process?
3. Any feel for how important the "probable extensions" are?

Thanks for taking the time!

Rick