Need help to set up Cluster for linux servers
 

Hi Gurus

My environment has 8 number of linux servers ( 4 X RHEL 6.2 64 bit and 4 X OEL 5.5 64 bit) where it is getting space for applications and database from a HP EVA 8400 SAN Storage.

I have been recently assigned to this project and asked to setup a production environment (OS Redundancy).

We have 14 number of HP VM machines running in a HP BL890c servers too.

What could be the best the way to bring these servers under the umbrella of cluster. Mostly i would go for multiple 2 node clusters.

Do i need to suggest my customer to for HP Service Guard Suite or Red Hat cluster suite?

Please help

There is no 'best' answer...you need to thoroughly analyze the clients needs, and tailor the environment to meet them, the same as you would for any other service. If you're in a corporate environment with an HP SAN, using Red Hat Enterprise and Oracle, then you have resources from all three vendors to ask for details about how they would best approach things.

You don't say what services you need to provide, or what your goals are. Do you want to provide failover/HA? Round-robin load balancing? Both? Uptime on the services, per the SLA? There are detailed clustering guides you can find on Red Hat's website...start there, and by contacting your vendors, and finding out what your clients needs/goals are.

Since you say you're a "Linux Specialist" with Red Hat certifications....shouldn't this be fairly self-explanatory for you?

Hi TBOne
Thanks for your quick reply. Customer is expecting High Availability for their application and Database. Customer don't have much experience on Linux and Unix and hence asked me to come with the proposal. As i am looking this as an opportunity/challenge to improve myself, wanted to give the best solution to my customer. I tried my best but failed to turn as an solution architect in overnight. Hence decided to take help from gurus like you.

I don't want to mislead the customer but wanted to give a simple yet best solution to them.
But as i am not getting much details on the license part, i have put this question on our forum.

Note: I will never use the term "Linux Specialist" (at least for some time) as i couldn't success in this matter.

Again, that is a broad term. Go back and re-read ALL of the points asked before. Things like failover time, acceptable downtime per SLA, etc. You also have to take into account the dependencies of things...if the application depends on the database, you need to make sure BOTH fail over, and in that order.
How did you 'fail'?? What did you do/suggest, before posting here? And if they don't have experience in Unix/Linux, why would they buy such a large amount of Unix/Linux hardware/software, and who maintained/installed it before you got involved???

Again, there is NO WAY anyone here can answer the question without a LOT of research/questions. And if we're going to do the consulting job for you, you need to post this into the LQ Job Marketplace, and tell us how much you're willing to pay.
And you're not going to get details on the license part in forums. You need to CALL THE VENDORS, that is Oracle and Red Hat, and ASK THEM. Each customer can have different license terms.
What terms you use are up to you...but if you've taken a job you can't do, you should be honest with your client. If you've got an RHCE, you should already have knowledge about how systems interact, and be able to perform basic research on your own. Asking about how to configure a two-node cluster and licensing terms doesn't show that effort.