Exchange 2010 with Hyper-V

HI,
I would like to know all the possiblities for running exchange 2010 in an virtualized environment.
Scenario: HP 460C blades -2 No:s with 2 Processor with 40GB RAM / HP EVA Storage  with 3TB/Users 750 on 3 Stoarge Groups (Currently on exchange 2007)
Proposed action: I would like to run hyper-V on the both the Blades with OS cluster for HA with 3 guest OS on each host for the following server Role
1. 2 File Servers with stoarge on SAN on both Hosts
2. 2 Exchange 2010 MB Role with Stoarge on SAN
3. 2 CAS/HUB Exchange Servers

What are the best practises I need to look forward and take care for the fully reduntant/HA & problem free scenarios??. Whether the Live migration will support the snap-shots & Microsoft Backup agent (Data Protection Manager - Agents) etc. Is there are any limitations in the near future for the upgradations / managment etc etc......????
rahimpaAsked:
Who is Participating?

Improve company productivity with a Business Account.Sign Up

x
 
jhindsonConnect With a Mentor Commented:
I would utilize CSV (Cluster Shared Volumes) for managing the Hyper-V storage to the virtual machines. CSV will reduce the number of LUNs required to enable virtual machine HA, and simplify administration. For the file servers and HUB/CAS servers I would leverage VHDs residing on a CSV.
I would suggest using pass-through disks for the mailbox databases since you want high-performing disks for the mailboxes.

Just to reiterate...Microsoft doesn't support combining Exchange 2010 high-availability (DAG members) with Hypervisor high-availability such as Live migration. See this article:  http://technet.microsoft.com/en-us/library/aa996719.aspx
0
 
kevinhsiehCommented:
2nd try, as E.E went down for when I tried posting before...

I don't see any issues with what you are trying, as long as there is a domain controller somewhere else. Exchange 2010 doesn't use any of the traditional Windows clustering technologies to achieve HA, so plan the Exchange environment just as you would physical machines. The only difference is that you don't have to worry about hardware failure, because if there was a problem with a blade, all of the VMs would restart on the other blade.

For the file servers, you can either have both file servers in a Windows cluster, or you can keep them as two separate servers and just use DFS replication to keep them mostly in sync and DFS namespace to control which server the clients are accessing. If you cluster the file servers, they need to access your SAN via iSCSI. I suggest that your mailbox servers use iSCSI as well, because that keeps all of the benefits of the SAN such as snapshots and dynamic volume expansion. If you put your data onto VHD, it will behave just like DAS. Using iSCSI from within the VM is probably a little slower, but I feel that it's worth it.

I am not familiar with how DPM works, but if you treat your file and mailbox servers just like physical ones, you shouldn't have any problems.

I would leave the total amount of RAM allocated to all of the VMs in the cluster to less than 39 GB so that if one of the blades were to go down that you have enough RAM on the remaining blade to run all of the VMs. That shouldn't be a problem with only 6 VMs, but something to keep in mind as you add VMs to your environment.
0
 
jhindsonCommented:
Also it should be noted that combining virtual machine high-availability (Live Migration) with DAG member server is not supported. The mailbox servers should not be configured to Live Migrate to another host server when participating in a Database Availability Group.
0
 
rahimpaAuthor Commented:
To brief the solution

I will have 2 host Systems with HYPER-V on windows 2008 Enterprise Core installation with Windows cluster and hyper-V added to the HA resource

* File server cluster with Shared LUN on the file server nodes ( V1 on Host1 & V2 on Host2)

* 2 HUB_CAS servers with VHD disks with pass through disks on SAN with seperate LUNS ( V3 on HOST1 & V4  
   on HOST2) with NLB

* 2 Mailbox server with Seperate Luns on the storage for the DAG with out any guest clustering (V5 on HOST 1 & V6 on HOST 2)

Concerns
1. Whether DAG is supported in this Scenario
2. The Virtual guest will automatically switch to the available server  and mount database w/o any service disruption

Over all comments on this design
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.