DAG Exchange replication topology.

Please refer these articles completely.

DAG Repliation happens only between DAG memebers (i.e. mailbox role servers which are part of configured DAG). We can select which database to be replicated to which server.

Mentioned  below articles will explain DAG features as well as setp-by-step how to set it up.

http://www.msexchange.org/articles_tutorials/exchange-server-2010/high-availability-recovery/uncovering-exchange-2010-database-availability-groups-dags-part1.html

http://www.msexchange.org/articles_tutorials/exchange-server-2010/high-availability-recovery/uncovering-exchange-2010-database-availability-groups-dags-part2.html

http://www.msexchange.org/articles_tutorials/exchange-server-2010/high-availability-recovery/uncovering-exchange-2010-database-availability-groups-dags-part3.html

http://www.msexchange.org/articles_tutorials/exchange-server-2010/high-availability-recovery/uncovering-exchange-2010-database-availability-groups-dags-part4.html

Ideally you will have 2 NICs (or 4 with NIC teaming to make 2) on each server.  
Setup one set of NICs to be on the same network as the clients.  The other NIC is just for replication.  Give it a unique IP subnet just for the DAG.
Click on the DAG in EMC.  In the bottom pane you will see the DAG Networks.  You will see Replication Enabled for the different networks.  Only turn on replication on the unique subnetted NICs.

SERVER1              SERVER2              SERVER3    
LAN NIC              LAN NIC              LAN NIC
REP NIC    <<>>   REP NIC    <<>>   REP NIC

Word of advise, teaming replication nics is not recommended and unsupported. You can team the public mapi network but not the replication nics. If course you can use additional replication networks.

Ideally the replication network should be an isolated vlan using a seperate subnet, if possible use a seperate switch for replication traffic to prevent contention related failures and to mitigate against single points of failure.

Even node dags need an additional file share witness to maintain quorum, odd number dags don't need this for operation but a fsw is required In order to create a dag.

the instructions I got said to have a second nic for replication. I made a direct peer to peer conection and assigned them 150.100.100.0/16

So I should have replication disabled for the MAPI

and enabled for the other subnet?

Correct, replication disabled for MAPI and enabled for the replication only NICs.

Also Teaming as I recommended IS supported.

"Additional Replication networks can be added, as needed. You can also prevent an individual network adapter from being a single point of failure by using network adapter teaming or similar technology. However, even when using teaming, this does not prevent the network itself from being a single point of failure."

Taken from Technet article Planning for High Availability and Site Resilience found here:
http://technet.microsoft.com/en-us/library/dd638104.aspx

I just tried to create the DAG the MBX2 joined succefully but MBX1 had the following error.

The operation failed CreateCluster errors may result from an incorrectly configured static IP address

Has anyone out there know why?

How many servers are you setting up in your DAG?  What other roles are you setting up on the same machines?

2 servers with Hub Mailbox and CAS

no other roles.

Also, please note that you should get some kind of hardware load balancers to balance out the CAS Array.  We used Kemp Loadmaster 2000's which were very easy to setup and not very expensive.

I have a do have a witness server only runing HUB transport, and it found it when I created the DAG, but when I add tried to add the 2 mailbox servers to the DAG  MBX1 and MBX2 only MBX2 is added sucessfully, and MBX2 is mounted to the on premise server MBX1
MBX1 gets the error that CreateCuster errors may result from incorectly configured static IP

also I noticed that there was a warning the failover clustering computer account is not online and to check that the IP address configuration fro the database availability group is correct.

Temporarily disable your replication NICs in Windows and see if it works then.

You have to do a witness server for the DAG because you only have 2 servers.  If you have 3 servers in the DAG you don't need a witness.  The steps I provided should setup everything for you.

Ok I think I found something both MBX1 and MBX2 can not resolve to the DAGs IP address. How can I put a static IP on the DAG account?

My DAG is not on a subnet with DHCP. So if My DAGs name is "testdag" What would the commands be to assign an IP through power shell?

Or could I assighn it a static through clustering management?

I am worried for you that you are not referring to the instructions I provided.  From step 4 above:

Set-DatabaseAvailabilityGroup -identity “TESTDAG” -DatabaseAvailabilityGroupIpAddress 192.168.1.4

You didn't say what IP so I just made one up.

Yes I followe your instructions and it worked thanks!! Everything is looking good except I am not sure exactly how this works. MBX2 has a copy of MBX2 mailbox and it mounted to MBX1s database. The preference is set to 1 on both.

So all email on MBX1 is replicated to MBX2 database. If MBX1 goes down MBX2 takes over.

Is that the whole Idea ?

Yes, the idea of the DAG is that both servers have all the same data and if one server goes down the other takes over.  

If you divide your data up into multiple information stores, for example let's say 4 stores.  Then you could set 2 databases per server to be active to increase performance by limiting the load on each server.  So in addition to being redundant, you are also distributing work load during normal operation.  If a DAG member goes down then the other server will bring those database online.  When the other server comes back online then you manually need to verify that the databases are in sync and then you can set the database to be active on the other server again.

Sounds like you are in good shape now, please don't forget to mark the accepted solution.

Regards,
Mark

Thanks!! Is there a way that I can be notified when the DAG is broken?

Take a look at this reference for how to monitor your DAG.  Note you could also use SCOM for more proactive notification.

http://www.ntweekly.com/?p=531

Another reference, this one from Microsoft:  http://technet.microsoft.com/en-us/library/dd351258.aspx

You could schedule the scripts in the above technet article to run and send the results to yourself in an email.  PowerShell Send Email- http://technet.microsoft.com/en-us/library/dd347693.aspx