Consolidation / VPAR (98 Views)
Reply
Advisor
javedk
Posts: 35
Registered: ‎11-26-2005
Message 1 of 3 (98 Views)

Consolidation / VPAR

We are getting a new application and the application vendors suggested 2 * RP3410 servers 2CPU each. We have couple of RP8420's and we were looking at the option of accomodating this on the RP partitions by adding more CPU's / Memory but the price for the CPU in RP8420 is 3 times expensive than that CPU on RP3410. The CPUS are the same Dual core PARISC 8900 @ 1Ghz. For the price of one CPU is RP8420 we can buy two RP34XX servers. What I would like to know is

1) Is there any difference in performance between the CPU's in RP8420 / RP34XX.
2) How do we acheive consolidation with this kind of pricing. Consolidation seems to be more expensive.
Honored Contributor
Steve Lewis
Posts: 725
Registered: ‎05-18-2000
Message 2 of 3 (98 Views)

Re: Consolidation / VPAR

1. Not much difference in overall performance; theoretically slightly better i/o and cpu-memory capability in the rp8420, but we have not found it to be noticeable in the real world.

2. Well consolidation is one of the main sales drivers these days, but your point about the tiered cpu pricing is a good one.

3. What you will also get with the rp8420 is:
3.1 redundant power configurations;
3.2 more flexibility if you use vpars or WLM as well, but you pay for those facilities on top of the basic cpu price;
3.3 a lot more i/o slots;
3.4 a lot more memory if you want it.

Esteemed Contributor
Ian Vaughan
Posts: 293
Registered: ‎07-24-2000
Message 3 of 3 (98 Views)

Re: Consolidation / VPAR

Howdy,
Just to add my 2p...
How hard are the existing rp8420's worked?
You mention VPAR in the title of the post, do you already have vpars on the rp8420's?
You have 2 servers, are they a ServiceGuard cluster and is one server just a standby for the other?

I'll make a sweeping assumption and assume that they are in a cluster and one is worked harder than the other and you have vpars. If this isn't the case then maybe a few parts below will be of interest if not all of it..

In the above situation you already have your procesors for your new application (the existing ones in your lightly loaded chassis) however you will need to reconfigure your vpars (leave one cpu in your standby vpar and create new vpar for the new app) and purchase more memory. Then you should buy some icap processors at 25% of the price of perm cpu's and turn them on using temporary icap (or use new "Global" icap feature) for processor resource in the event of a failover or when you need some more resource (end of month?). Your provision of processors between your vpars can be driven by your package startup scripts if necessary or through Global Workload manager if finances allow it.

This approach would get the most out of the hardware you already have without adding additional power consumption and additional cooling costs to your budget.

Have a look at this as well -
http://docs.hp.com/en/B9073-90101/index.html

Hope this helps
Ian
What's yellow and dangerous - shark infested custard
The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation.