DAG (Database Availability Groups)-nabbit!

If you like your CCR Cluster and SCR Targets in Exchange 2007 you are going to love DAGs (Database Availability Groups) in Exchange 2010. I was just watching the High Availability for Exchange Server 2010 four part BLOG series by Scott Schnoll. This is a fantastic introduction into the new high availability options for Exchange 2010. Watching this really inspired me to play around with this in the lab and to write this blog. I have personally been working with Exchange Clusters for a very long time. I have seen a number of improvements in Exchange 2000, Exchange 2003 and Exchange 2007. However, I can’t tell you just how much I am impressed with the new DAGs and other HA achievements in Exchange 2010. What makes it so impressive? How ridiculously easy the Exchange team has made it now to support highly available databases. Microsoft describes the new high availability features including DAGs as Database Mobility. Database Mobility key features are:

  • Incremental Deployment
  • Database Availability Group (DAG)
  • Mailbox Database Copies

What is also great about database mobility is the real flexibility it provides in deciding when to utilize HA in your organization. You see in the past if you wanted high availability from Exchange you need to design and deploy the clustering technologies before you deploy the exchange servers. Microsoft has introduce the concept of incremental deployments whereby we can deploy a standalone Exchange server today with whatever roles we need to deploy (Mailbox , Hub, CAS and UM.) and then deploy a DAG at a later point in time to enable high availability. And you can keep your other roles at the same time on the same server too!  DAGs are comprised of two or more servers that have the failover clustering services enabled on them. The failover clustering is enabled and configured automatically when you add the server to a DAG. You can add up to 16 separate Exchange 2010 mailbox servers to a DAG. Note that you will still have to have a witness server that is not a member of the DAG just like the file share witness of the CCR cluster in Exchange 2007.

Once you have added servers to a DAG you will be able to replicate copies of your databases to the other server(s) in the same DAG using the new and improved continuous replication built into Exchange 2010. Key enhancements include using native TCP sockets for the TCP/IP stack rather than SMB file transfers for the replicated logs. Administrators can also enable Encryption and compression on the replication traffic for security and performance. It is possible to have up to 16 copies of a single database in a DAG one for each of the possible 16 servers in a DAG. Both manual and automatic failover options are available when a production database or server fails.

I guess you can tell I’m pretty excited about this new feature but even I’m still playing around with mailbox mobility in Exchange 2010. I plan to blog some more about it once I have had a chance to really see it in action. If you are curious for more information I would encourage you to check out Scott Schnoll’s blog (see link above). I think you will find this to be a very compelling reason for many to move to Exchange 2010 when it is released.

Author: Richard Luckett

In this article

Join the Conversation