Create DAG for DR or site resiliency RRS feed

  • Question

  • Hello all

    Because we are currently already heavily invested in SAN's, and because there is a lot of built in redundancy with our SAN environment, i am thinking that having 2 copies of all databases that are located on the same SAN and same AD site is more redundancy that i need, and at the same time will consume 2x the storage. Below is an example of what i am considering for DR purposes.

    #1 in Site1 create two mbx,ht,cas servers. Both mbx servers would be connected to the same SAN, and the two CAS serves would be configured in a nlb array using an F5

    #2 in Site 2 create one mbx,ht,cas server

    #3 setup a DAG between the mailbox servers located in site 1 and the mailbox server located in site2

    #4  create copies of all the databases that are located in site1 to the mailbox server located in site2. The two servers in site 1 would not replicate databases to eachother (again same san and built in san redundancy)

    #5 the mailbox server ( or DAG member) in site 2 would be used only in the event of a database failure or hardware failure on either one of the servers in site1. There is a SAN in site2



    My questions are

    #1 in the event a database gets corrupted or a mailbox server goes offline in site1, and i failover the one database that got corrupted or all the databases (server failure) to site2, would the CAS server thats in site1 notice this change and proxy the mapi request to the CAS sever in site2?

    #2 If I have to restart a server in site1 i dont want exchange to  start the process of trying to find a suitable copy of a database that's located  in site2. I want site2 to be used only in event of a server or database failure.

    #3 Is there a way to control how and when Exchange fails over or the process for when it starts to look for copies of a database?

    #4 Because I will have 3 servers in the DAG, and one of the members of the DAG is located in another AD site with a different ip subnet, then what kind of cluster will the DAG create, will it still be a FSW? and if so how many witness servers should i use?


    Bulls on Parade
    Friday, April 23, 2010 6:06 PM

All replies

  • #1 If the mailbox server went offline the CAS server in site 1 would proxy the request to site 2

    #2 you can't have it both ways by default, if you want automatic failover during a failure, the same would happen during a reboot.  Before rebooting the site 1 server, you could take the databases in site 2 offline

    #3 not exactly sure what you are asking

    #4 since you have an odd number total, you won't have a FSW


    Joseph Durnal

    Joseph M. Durnal MCITP: Enterprise Messaging Administrator
    Friday, April 23, 2010 8:05 PM
  • What about implementing DAC on the DAG this sounds like i can use it to control failovers to a mailbox server in a different site?


    If i have an odd number of servers in the DAG, then what type of cluster will Exchange configure? If I configure a FSW on a member server, then doing this will add another server to the DAG, so at this point i would have an even number of servers

    Bulls on Parade
    Friday, April 23, 2010 9:04 PM
  • Hi,

    If we have internet-faced CAS server in site1, it will proxy request to site2.

    Failover means When a failure occurs that interrupts availability to one or more active mailbox databases, the DAG automatically initiates corrective action to restore service.  I think it hard to control.

    We can try to monitor the Exchange Server and then manually switchover the database.

    Switchovers and Failovers


    For DAC mode,  it should only be enabled for DAGs with three or more DAG members that have been deployed in a multi-datacenter configuration. DAC mode is used to control the activation behavior of a DAG when a catastrophic failure occurs that affects the DAG (for example, a complete failure of one of the datacenters).


    With Odd member, it could be Node Majority quorum model.


    For FSW, it cannot be the member of DAG. The requirements for the witness server are as follows:

    • The witness server can't be a member of the DAG.
    • The witness server must be in the same Active Directory forest as the DAG.
    • The witness server must be running Windows Server 2008 R2, Windows Server 2008, Windows Server 2003 R2, or Windows Server 2003.
    • A single server can serve as a witness for multiple DAGs; however, each DAG requires its own witness directory.

    We recommend that you use an Exchange 2010 Hub Transport server in the Active Directory site containing the DAG. This allows the witness server and directory to remain under the control of an Exchange administrator.

    Managing Database Availability Groups





    Tuesday, April 27, 2010 9:20 AM
  • If you have odd number of servers then exchange will use Node majority. If we have even number then exchange will use Node majority with FSW.



    Monday, October 18, 2010 2:38 PM