Sign in
Microsoft.com
 
Microsoft
 
 
 

locked
DR for OCS RRS feed

 > 

  • Question

  • Question
    Sign in to vote
    0
    Sign in to vote

     

    I'd like to hear what others are doing to provide DR to OCS. What I mean by DR is offsite standby servers that you can fail-over too in case of complete primary datacenter loss. I assume one option would be to create a complete secondary install of OCS in the DR datacenter. In event of needing to faillover just update the SRV record. Anyone else want to chime in on this?
    Wednesday, March 12, 2008 6:24 AM

Answers

All replies

  • Question
    Sign in to vote
    0
    Sign in to vote

    You can only have An Enterprise pool with load balancers.

    If there are Load Balancers that span the WAN it might work?

     

    Don't think this is possible.

     

    Johan

     

     

    Wednesday, March 12, 2008 11:44 PM
  • Question
    Sign in to vote
    0
    Sign in to vote

    to provide HA into OCS please make sure considering items like:

     

    - make sure your SQL server is hosted on a Windows Cluster

    - make sure using hardware based load balancers http://unified-communications.blogspot.com/2008/03/ha-into-your-ocs-infrastructure.html

    - make sure your connection to yor basic media gateway is supported by minimal 2 mediation servers (one hot standby)

    - take a look at my weblog http://unified-communications.blogspot.com/2008/02/it-showcase-deploying-microsoft-office.html (see the topic on load balancing)

     

    Load Balancing

    Microsoft IT wanted to ensure that users experienced a high quality of service with Office Communications Server 2007. Although sizing servers properly, distributing the topology to multiple data centers, and ensuring adequate bandwidth provided a sound foundation for a high quality of service, using load balancing at multiple points gave Microsoft IT yet another level of control for optimization. Microsoft IT decided to use both hardware load balancers and Domain Name System (DNS) round robin in its design. Specifically, Microsoft IT uses the following strategies with these server roles:

    ·        Server pool   Microsoft IT uses a third-party hardware load balancer for the front-end servers in the server pool. This is the case with both the expanded and consolidated configurations. Using two server pools in each data center in the consolidated configuration provides load balancing in addition to redundancy.

    ·        A/V Edge servers   Microsoft IT uses DNS round robin to balance the media traffic that the A/V Edge servers handle. Microsoft IT chose to use round robin (an unsupported configuration) initially instead of a hardware load balancer for expediency in order to deploy A/V Edge servers without the need for the additional training and support requirements associated with a hardware load balancer. In the future, Microsoft IT plans to implement a hardware load balancer for A/V Edge servers.

    ·        Director servers   Microsoft IT uses a third-party hardware load balancer for servers that host the Director role.

    ·        Access and Web Conferencing Edge servers   The servers that host the Access Edge and Web Conferencing Edge roles are load-balanced with a third-party hardware load balancer.

    -Joachim Farla

    Sunday, March 16, 2008 10:25 PM
  • Question
    Sign in to vote
    0
    Sign in to vote

    Thanks for the response, however I'm not speaking strictly HA but more along the lines of DR as in complete loss of data center. Obviously one could use GeoClusters to cluster the SQL database but you would also need to span your front-ends across the WAN as well. I'm wanting to know what others are doing to bring OCS back online in a secondary data center when the primary burns to the ground.

    Monday, March 17, 2008 1:36 PM
  • Question
    Sign in to vote
    0
    Sign in to vote

    We're having the same discussion now as well.  Our difference is we want load balance the connections across multiple wan sites.  Wer're currently configured as a single site edge topology, but would like to move towards the multiple site edge topology.  Ideally, we want our Pool to span these sites, hosting one front end server in each location.  The CDR server is self explanatory, but the Edge role  brings up a few questions however. 

     

    Has anyone inserted a load balanced Edge server (conf server as well) after the fact?  What about PIC?  If anyone has done this, did you load balance the FQDN of the original Edge server Role, and re-configure the setup of the server role? 

     

    What are folks using for Load Balancing?

    Monday, March 17, 2008 2:26 PM
  • Question
    Sign in to vote
    0
    Sign in to vote
    We're undergoing the same assessment now. What would be the best solution with OCS (Std now but may have to go to Ent), with one site in production and one in DR? (Warm Standby, not load balanced if possible)
    Wednesday, March 19, 2008 5:38 PM
  • Question
    Sign in to vote
    0
    Sign in to vote

    There will be a webcast handling DR in OCS

    See the list of webcasts

    http://www.pro-exchange.be/modules.php?name=News&file=article&sid=833&mode=&order=0&thold=0

     

    TechNet Webcast: Disaster Recovery in Communications Server 2007 (Level 200)

    Thursday, April 10, 2008
    8:00 A.M.–9:00 A.M. Pacific Time

     

    Wednesday, March 19, 2008 8:25 PM
  • Question
    Sign in to vote
    0
    Sign in to vote

    Thanks..I will definitely sign up for this webcast.

     

    Microsoft IT just released a whitepaper on how they deployed OCS at Microsoft. The whitepaper discusses that they configured the client boot-strap to be able to fail-over to a secondary datacenter if the primary was down. I suspect this will be discussed during the webcast. The link to the whitepaper is here:

     

    http://technet.microsoft.com/en-us/library/cc297283.aspx

     

     

     

     

    Sunday, March 23, 2008 3:45 AM