• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 529
  • Last Modified:

Hyper V Cluster Shared Volumn Performance

Hi,
I have 2 host and an ISCSI SAN configured a Hyper V 2008 R2 cluster.
I have 3 options to deploy the VMS.
Option1 : create a big LUN in SAN , and only volumn1 in CSV , put all VMs in there

Option2:  Cteare 2 big LUN in SAN , then have volumn1, volumn2 , put all VM configuration and      Disk C in volumn1 , put Disk D, E... into volumn2

Option3: Create 3 big LUN in SAN , get volumn1, volumn2, volumn3 , put all
VM Configuration files in volumn1
Disk C in volumn2
Disk D,E... in volumn3

What will the best option to do , is there any evidence show the performance different?
EG : is there an chart to show the performance different , like X% faster or slower?
0
GordonLiq
Asked:
GordonLiq
  • 3
  • 2
3 Solutions
 
jakethecatukCommented:
One thing to consider above disk performance is spreading the load and high availability.

If memory serves (and I haven't looked at HYPER-V for about 12mths), you will need to have each VM in a separate LUN to allow for fail over and load balancing.  If you go with option 1 or 2, only one host will have access to your VM's due to the disk configuration.
0
 
GordonLiqAuthor Commented:
Hi Man,
this is past , the Hyperv 2008 R2 resolve the problem. all 3 options works as well, just dont know about performance
0
 
jakethecatukCommented:
You're right - CSV's claims to resolve the problem I hightlighted.

Performance then - how many spindles do you have in your iSCSI SAN as knowing that will help.
0
Creating Active Directory Users from a Text File

If your organization has a need to mass-create AD user accounts, watch this video to see how its done without the need for scripting or other unnecessary complexities.

 
jakethecatukCommented:
oh, and what RAID level are you thinking and what will your VM's be doing?
0
 
GordonLiqAuthor Commented:
It is HP P4300G2 ISCSI SAN , I will hold about 10 VMs with Windows 2008 R2 and Linux.

SAN configured RAID10
0
 
kevinhsiehCommented:
If all LUNs go back to the same SAN and same disks, there shouldn't be a significant difference no matter how many CSV volumes you use. NTFS cluster size on the NTFS is going to make a bigger difference.
Som
There is a potential that if you have so much IO to a volume that the command buffer for the LUN will fill, which will cause you to have less than is you had multiple LUNs, each with their own set of buffers. This isn't something I worry in my environment. I suggest you setup your CSV in a way that makes sense to you from an administrative standpoint. For example, I use multiple LUNs for different snapshot and replication schedules. I also use multiple LUNs when I feel that a LUN is getting to big for me to be able to easily handle (about 800 GB is my cutoff choice).  
0

Featured Post

Get expert help—faster!

Need expert help—fast? Use the Help Bell for personalized assistance getting answers to your important questions.

  • 3
  • 2
Tackle projects and never again get stuck behind a technical roadblock.
Join Now