Two dag member no cas array

Question

I have two exchange servers (exchange 2010) in a dag with two databases. The production exchange server and a dr exchange server. Both of the exchange servers have a cas and I have not setup a cas array. I have set the  -RpcClientAccessServer to the
production exchange server.
Is this a valid configuration?
Is it possible to prevent connections to the second exchange server? I find on email headers and on outlook 2010 connection status that there is connections to the second exchange server constantly, even though all mailboxes are on the production server
and  -RpcClientAccessServer is production server.

Answer 1

I will look at creating a CAS array then. I thought i would be able to get away with out it. I think what happened was that after adding the second exchange server i did not update the rpclientaccesserver property to point back to the production server,
for some reason this automatically changed itself to the new exchange server. This means for a few weeks the users were connecting to the DR exchange server over a slow 100mbit link. As soon as i realized that i changed the rpcclientaccesserver property back
to the production CAS on both databases.
I found this http://www.mikepfeiffer.net/2011/04/determine-the-number-of-active-users-on-exchange-2010-client-access-servers-with-powershell/
Very handy powershell script that tells you client connection status to the CAS. currently we have 85 users on production CAS and 20 still on the DR cas even though i set it back to correct CAS over a week ago. Now i need to look at moving these users over
to the correct cas. But thinking might as well wait till we set up CAS array and then change all users that way.
the reason i dont have active active is because its over a 100mbit link and is not as fast as the gigabit we have to the production exchange and is not on as fast storage or hardware. The DR exchange just sits on local storage on one esxi box with one other
server. The rest of our vms are snap mirror across with netapp snap mirror to third party datacenter. In the event of a dr they will supply some new servers for use with the snap mirrored vms.
In the event of a DR situation there is the understanding that it may take a day before we have everything up and running and this is accepted by the client. We are looking at improving this though and would be interested to hear any suggestions regarding
improving exchange dr.
I had some problems with the DAG last month. I had problem seeding one database and finally managed to get that workign and recently had to move the witness to another server as that stopped working with strange error. But now the DAG is 100% healthy with
no issues, i think the 20 random email problem was caused by stuck emails in the users outlook 2010 on xenapp that he never reported, which is another known issue when migrating from outlook 2003-10. Still very concerning however.
It would be very handy if there was some automated way to force outlook to the correct CAS only, say by changing a registry key with group policy. If anyone knows of any way to do that, would be helpful thanks.