DRS: e3-1230 and x3440 work together?

peanuthead

Supreme [H]ardness
Joined
Feb 1, 2006
Messages
4,699
I am just looking for a quick answer to help clear up my confusion. Will the E3-1230 and x3440 work together for DRS? Or should I build my second box with a c202 chipset and a e3-1230 xeon to match my c204 chipset server?

I should be okay from what I can tell, but I wanted to check with the veteran users first.
 
Last edited:
Its not as simple as that. Which part of DRS do you want to use. If you use DPM you might run into problems. When you design your cluster you should set the highest Baseline of your hosts and then apply a policy accross.
You can run different generations of CPU's as long they are from one manufacturer.
ie: Host 1 CPU = Core2
host 2 CPU = i5
host 3 cpu =i7
Host 1 acpi power = S3
host 2 acpi power = s3
host 3 acpi power = s3

A default policy should be made to install all guests to use the highest baseline of core2 and acpi power s3
This will ensure that only features of core2 are used when installing a guest, this will ensure that all Vmotion will be succesful when DRS is called on. Definetley Resource pools and resource allocation to high priority vms need to be setup correctly. Its a bit more complicated but thats kind of the base idea behind it.

As far as c202 and c204 chipsets being a neccessity I don't think it will be a gigantic problem
 
Thank you, Wrench00. That is the gist I figured out by reading. The highest level that I can go is "Nehalem" due to the Lynnfield processor. Is there any reason to consider (advantages or disadvantages) having by baseline set to "Sandy Bridge" level IF I replaced the Lynnfield setup with a SB setup?
 
If you install a guest os that uses certain features of a CPU ie. SSE3 but one of your hosts only has the capabilities of SSE2 you will not be able to do VMotion (VMWare will tell you that)to the Host with SSE2 which kinda defeats the purpose of DRS in the first place. DRS is more or less an automated VMotion configuration. Basically they did this so you can migrate your Clusters to newer platforms with out breaking backwards compatibility. Granted most guest instlls like windows from 2003 will not use Sand Bridge features because the os is too old to know the difference but an OS that you can install and upgrade the Kernel features that recognize CPU features ie. BSD/OSX or any nix version will break your DRS. Some programs will do this as well.

As far as advantages disadvanteges go, well usually its all related to speed and performance gains.
 
Thanks! Makes perfect sense. All the hosts need to "look the same" when they look at each other, hence the use of EVC.

I also read that the processor must be from the same family and manufacturer. For example, all CPUs must be from Intel and a Xeon in all of the hosts. Nothing like an Intel Xeon (lets assume E3 variant) in one host and an Intel i3 in another host. Is this correct? I would think they should work due to the same base instruction sets other than turbo boost, etc.
 
Yes I have had success running certain consumer versions of Intel CPU's mixed with Comercial ones. They still use the base design and instruction set. Granted tho this was a test enviornment not a production one. AMD was actually the same result. I have an i5 750 and a xeon 34XX working together with vmotion.
If you got Intel stick with intel if you got AMD you stick with AMD.
EVC is the key.
And yes you are correct. Hope that helps.
 
Back
Top