Over the time, I’ve seen/had several questions on whether this is a supported scenario or not. A two-member DAG with a member server in the primary datacenter and one in the backup datacenter is an attractive solution for especially SORGs that want to deploy a solution where mailbox data are replicated to another site, but where local HA for the HT, CAS, and MBX roles is not a requirement.
The short answer to the question is yes, this is a supported scenario.
The long answer is yes this is a supported scenario, but you should bear in mind that datacenter activation coordination (DAC) mode is not supported with only two members servers in a DAG. This means that you will not be able to benefit from the automatic failover mechanism provided by DAG and that you cannot use the site resilience tasks built into Exchange 2010 (Stop-DAG, Restore-DAG, Start-DAG) to activate the DAG member in the backup datacenter. When you need to bring the DAG member in the backup datacenter online, you must instead use the Windows Failover cluster tools.
First, you need to stop the cluster service in the primary datacenter followed by a force cleanup:
Net stop clussvc
Net stop clussvc cluster <DAGName> node <DAGMemberName> /forcecleanup
Then you must stop the cluster service in the backup datacenter:
Net stop clussvc
Finally you must force quorum in the backup datacenter:
Net start clussvc /forcequorum
For more information on DAC, datacenter switchovers, see the relevant TechNet documentation at below links:
http://technet.microsoft.com/en-us/library/dd979790.aspx
http://technet.microsoft.com/en-us/library/dd351049.aspx
Cheers,
Henrik Walther
Technology Architect/Writer
MCM: Exchange 2007 | MVP: Exchange Architecture
