locked
Active node disaster recovery RRS feed

  • Question

  • Hi,

    Problem Statement:

    There is a business requirement for Active Failover of the Primary Site Server role and functionality in case of a disaster recovery solution. I am aware that the site roles do not support clustering but I want to check whether there are any options to actively restore the Primary Site Server role in case of an outage without having to restore databases or re-assign site codes on clients.

     

    Thanks

    Tuesday, June 14, 2011 1:40 PM

Answers

  • Not within ConfigMgr itself.

    The two primary options used by some customers are hypervisor high availability possibly combined with geo-clustering (such as Hyper-V Live Motion) and boot from SAN possibly combined with storage replication.

    Details of how to implement even these though depends on what is included in the scope of your disaster though. Are you talking about a simple hardware failure, application failure, site resiliency, data corruption, or something else? Each of these has its own challenges that must be planned accordingly.


    Jason | http://myitforum.com/cs2/blogs/jsandys | Twitter @JasonSandys
    • Proposed as answer by Garth JonesMVP Tuesday, June 14, 2011 11:52 PM
    • Marked as answer by Yog Li Wednesday, June 22, 2011 11:49 AM
    Tuesday, June 14, 2011 2:00 PM
  • ConfigMgr was not designed for high availability.
    You can reassign clients to another hierarchy, see http://blogs.technet.com/b/configmgrteam/archive/2009/04/09/reassigning-a-configuration-manager-client-across-hierarchies.aspx
    Torsten Meringer | http://www.mssccmfaq.de
    • Marked as answer by Yog Li Wednesday, June 22, 2011 11:49 AM
    Tuesday, June 21, 2011 7:44 AM

All replies

  • Not within ConfigMgr itself.

    The two primary options used by some customers are hypervisor high availability possibly combined with geo-clustering (such as Hyper-V Live Motion) and boot from SAN possibly combined with storage replication.

    Details of how to implement even these though depends on what is included in the scope of your disaster though. Are you talking about a simple hardware failure, application failure, site resiliency, data corruption, or something else? Each of these has its own challenges that must be planned accordingly.


    Jason | http://myitforum.com/cs2/blogs/jsandys | Twitter @JasonSandys
    • Proposed as answer by Garth JonesMVP Tuesday, June 14, 2011 11:52 PM
    • Marked as answer by Yog Li Wednesday, June 22, 2011 11:49 AM
    Tuesday, June 14, 2011 2:00 PM
  • 1) We can not afford to virtualize the primary server since there are many clients assigned to the Primary Server. The load on the server is high therefore Virtualization is not the best option.

    2) I am looking at an active DR situation for:

    - Hardware Failure

    - Application Failure

    - Data Resiliency

    - Database Failure

     

    The key ask is - IN case of a complete breakdown of the Primary Server, I should be able to fail over the entire funtionality of the Primary Srver to another Primary Server at another site with zero down time.

    Tuesday, June 21, 2011 2:10 AM
  • ConfigMgr was not designed for high availability.
    You can reassign clients to another hierarchy, see http://blogs.technet.com/b/configmgrteam/archive/2009/04/09/reassigning-a-configuration-manager-client-across-hierarchies.aspx
    Torsten Meringer | http://www.mssccmfaq.de
    • Marked as answer by Yog Li Wednesday, June 22, 2011 11:49 AM
    Tuesday, June 21, 2011 7:44 AM