Not knowing anything about your project (aside from
some technical details), it's impossible to give you specific
advice.
What's your timeframe for
implementation?
Do you have any of this built or are you starting from
scratch?
What is your level of expertise in the technologies you
mention? Each of these are complex on their own. Do you have
help? How much of this have you done before?
If I were going to attempt such a project myself, I'd
look for no less than a six-month engagement to design, configure, test and
deploy a setup for production use. During that time I would take the
opportunity to learn each technology in detail, run experiments, tweak and push
them to their limits. Then always do the fun part--pulling cables to see
what happens in a redundant cluster when components really do
fail.
I'd also pick an order to configure and verify each
component in turn, something like:
- Hardware
- OS
- Networking
- Cluster
- SAN
- GFS
- Oracle
-
application
with the appropriate stress testing and proofing at
each step. Depending on timeframe and budget, I'd also give serious
consideration to training and professional consulting.
Good
luck.
-Jeff
From: linux-cluster-bounces@xxxxxxxxxx [mailto:linux-cluster-bounces@xxxxxxxxxx] On Behalf Of Mad Unix
Sent: Saturday, December 15, 2007 3:48 AM
To: linux clustering
Subject: RH Cluster
I'm in the process of setting up a 2 nodes ACTIVE/ACTIVE cluster
with PE2950 16 RAM + RHEL5 + GFS + RHCS+ SAN EMC CX10.
Each of my nodes are RAID 1+ 0 and have the regular OS tree
Each of my nodes are RAID 1+ 0 and have the regular OS tree
and share SAN storage Raid5, am planning to build it for Oracle10g
Database
_without_ using Oracle cluster ware CRS and ASM (I believe the cluster will
be done much better via the OS not DB)
to serve about 1000 users the application based on client/server
concept..
Can someone please give me a pinpoint of where I start ?
Regards,
--
madunix
-- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster
