locked
Best Practices and Recomendations RRS feed

  • Question

  • I have an opportunity to setup a new corporation and enterprise infrastructure from scratch. I want to do the latest/greatest Microsoft Full Meal Deal.  This includes System Center, Forefront (Stirling), Office Communications Server 2007, MOSS 2007, SQL 2008, and Dynamics AX 2009 for starters.  My thought is to use Dell 2950's with dual quad core and 32GB of RAM and use Hyer-V.  I am wondering what are the best design and setup practices for Redunancy and Disaster Recovery.  I do not need 5 9's or instant failover but what are my options or recomendations for architecting DR.

    Can two servers be clustered and then have VM's run on top?

    Should the VM's just get backed up and mirrored to the redundant server to be brought on-line if there was a failure?

    Anyone have a problem with any of the Servers I have listed running virtualized?

     

    I am an InfoPath MVP and just returning to systems management since I have an opportunity to set up the latest environment without being limited by legacy systems and integration.  But I want to do it right the first time and hope some of you can give me your opinions.  Or point me at the existing resources I should be looking at.

     

    Thanks so much.

    Clay

    Tuesday, May 13, 2008 6:46 PM

Answers

  •  Clay Fox wrote:

    I have an opportunity to setup a new corporation and enterprise infrastructure from scratch. I want to do the latest/greatest Microsoft Full Meal Deal.  This includes System Center, Forefront (Stirling), Office Communications Server 2007, MOSS 2007, SQL 2008, and Dynamics AX 2009 for starters.  My thought is to use Dell 2950's with dual quad core and 32GB of RAM and use Hyer-V.  I am wondering what are the best design and setup practices for Redunancy and Disaster Recovery.  I do not need 5 9's or instant failover but what are my options or recomendations for architecting DR.

    Can two servers be clustered and then have VM's run on top?

    Should the VM's just get backed up and mirrored to the redundant server to be brought on-line if there was a failure?

    Anyone have a problem with any of the Servers I have listed running virtualized?

     

    I am an InfoPath MVP and just returning to systems management since I have an opportunity to set up the latest environment without being limited by legacy systems and integration.  But I want to do it right the first time and hope some of you can give me your opinions.  Or point me at the existing resources I should be looking at.

     

    Thanks so much.

    Clay

     

    As far as HA goes, Ben Armstrong posted an article linking to how to set up highly available virtual machines...

     

    http://blogs.msdn.com/virtual_pc_guy/archive/2008/05/13/hyper-v-clustering-step-by-step-guide.aspx

    Wednesday, May 14, 2008 11:40 PM

All replies

  • I'm looking at something similar for some time now, just wainting for Hyper-V to go RTM to start. I wouldn't touch Dell again from previous experience with servers. If something goes wrong you'll really in trouble. Stick with name brand hardware, you'll be surprised at how little the price difference is now. HP isn't as good for options as one would expect which pretty much leaves you with IBM but more options than you'll ever need.

    Think about using a SAN solution with two hardware platforms for virtualisation. If two hardware platforms will do it for you, you have the option of using SAS for the SAN solution which is a lot cheaper than Fibre channel that would be required for more than two hardware platforms.

    DR can then be a mattter of loading the VMs on the surviving hardware from the SAN. You can also make backups of the VMs from the SAN and restore to another hardware platform with the same virtualisation resources as a another DR option.

    VMs seems to make DR a really str8 forward process which it has never been in the past. Just wondering what I'm missing, surely it's too good to be true....

     

    Cheers,

    Mark.

    P.S. Would really like to know what you end up using hardware wise.

    Tuesday, May 13, 2008 11:29 PM
  •  Clay Fox wrote:

    I have an opportunity to setup a new corporation and enterprise infrastructure from scratch. I want to do the latest/greatest Microsoft Full Meal Deal.  This includes System Center, Forefront (Stirling), Office Communications Server 2007, MOSS 2007, SQL 2008, and Dynamics AX 2009 for starters.  My thought is to use Dell 2950's with dual quad core and 32GB of RAM and use Hyer-V.  I am wondering what are the best design and setup practices for Redunancy and Disaster Recovery.  I do not need 5 9's or instant failover but what are my options or recomendations for architecting DR.

    Can two servers be clustered and then have VM's run on top?

    Should the VM's just get backed up and mirrored to the redundant server to be brought on-line if there was a failure?

    Anyone have a problem with any of the Servers I have listed running virtualized?

     

    I am an InfoPath MVP and just returning to systems management since I have an opportunity to set up the latest environment without being limited by legacy systems and integration.  But I want to do it right the first time and hope some of you can give me your opinions.  Or point me at the existing resources I should be looking at.

     

    Thanks so much.

    Clay

     

    As far as HA goes, Ben Armstrong posted an article linking to how to set up highly available virtual machines...

     

    http://blogs.msdn.com/virtual_pc_guy/archive/2008/05/13/hyper-v-clustering-step-by-step-guide.aspx

    Wednesday, May 14, 2008 11:40 PM
  •   There will also be a webcast next week on high availability with Server 2008 (including Hyper-V related features).

     

     

    Thursday, May 15, 2008 4:04 AM