Can an Exchange 2010 server belong to two different DAG's?




I have a DAG configured with three servers, one of those in another AD site. I want to install another Exchange in the remote site and create another DAG. Can I use the server in that site that belongs to the current DAG to be member of the new DAG?

Thank you!


No it can not. But you can stretch your DAG across subnets and AD sites with no problem.

Use the Add-DatabaseAvailabilityGroupServer cmdlet to add a Mailbox server to a database availability group (DAG). A DAG is a set of Mailbox servers that provide automatic database-level recovery from a database, server, or network failure. When a Mailbox server with replicated databases is added to a DAG, it works with the other servers in the DAG to provide automatic, database-level recovery from database, server, and network failures.

To add a Mailbox server to a DAG, the Mailbox server:
Must be running Windows Server 2008 Enterprise Edition or Windows Server 2008 R2 Enterprise Edition; Must not belong to any other DAGs; Must be in the same Active Directory domain as all other Mailbox servers in the DAG; and Must not be configured as an Active Directory directory server.

AndyD_ [MVP]

Even if you could, it sounds like a disaster in the making!



What about the following scenario:

Three AD sites. Each site with two Exchange Mailbox. I would like to have the two servers on the site containing one DB serving to the local users of that site. I also want to have a HA, standby copy of the DB on the other server of that site. Finally I want a lagged DB copy on a server in another site.

What do you recommend? Can I have all those DB copies, all those servers of the three sites in one DAG? Or is better to have one DAG per site and create another mailbox server in another site for the lagged copie?

Thank you!

Brian Day MCITP

Are you envisioning something like this?

Site-A: MBX1 (DB1-Active), MBX2 (DB1-Passive)

Site-B: MBX3 (DB2-Active), MBX4 (DB2-Passive)

Site-C: MBX5 (DB1-Lagged), MBX6 (DB2-Lagged)

You'll also need a 7th server holding a file share witness due to having an even # of servers in the DAG. Let's say it is in Site-A for now.

The very first thing you need to do is ensure you have a full mesh network topology between all 3 sites. No one backhoe can take out a single piece of copper or fiber that drops your connection to two remote sites or else the local DB will as the local servers will have lost quorum.



DAGs can easily cross AD sites and are no more than a logical replication boundary. You can have up to 16 servers in a DAG. So if you want to replicate a database to another server, then it needs to be a member of the same DAG. One other thing to note is that you can only have one copy of a database on a server (because it must share the same database location path). This means that if you have a lagged copy of your database you cannot have an HA copy of that database on the same server. One issue that you have to work out is client access. Becuase users connect to the CAS role and not the mailbox role, you need to design the CAS appropriately. One thing to note, if you create a CAS array, it cannot cross AD sites. You will need a CAS array in each AD site. This means that if a site completely goes offline and you failover to another site, then your will have to update the RPCClientAccessServer paramenter on the database or users will not be able to connect.

So to answer your question, I would make all the servers a member of the same DAG (since you want to replicate database between sites). You should not consider a lagged copy a viable copy to activate as it will take a while to replay the logs. Lagged copies should be used for point in time recovery or a backup alternative. Design your required HA solution and then add some lagged copies if you really need them.