Monday, April 5, 2010

RPCClientAccessServer attribute

In small networks where Exchange 2010 needs to be configured with redundancy , there is manual DNS works needs to be done in order MAPI clients to connect to “mounted database” in case of database fail over.

Of course if you can purchase two more servers and dedicate resources you can set CAS Array and make the outlook users MAPI experience completely automated if anything happens to “mounted” database and “healthy” database becomes “mounted” database. We are talking about small environment there are no Load balancers and no other resources, all we  have is two Exchange Server and we will make our messaging environment “redundant”


  • When you introduce first CAS server ( where CAS Array has not been created) The attribute called “RpcClientAccessServer” will be set to first CAS server in the AD site.
  • Each Exchange Database does contain this attribute  “RpcClientAccessServer” = Server Name
  • If you cannot afford to have CAS array still create the CAS array with common name and give IP address as one of your mail servers.
  • This will allow you tagging common name to RpcClientAccessServer attribute and you will control this with modifying DNS (A record) entry.
  • In small environment you don’t have to worry about messing with TTL when you make changes to this A records

here is what I put together to make this bit clear


If database fail over occurs , all you have to do it to change the A record in DNS and pint the exiting A record to second server IP address.

Normal Operations DNS Entry will look like this = (mail1)

If database fail over occurs , the DNS will look like this = (mail2)

image image

Outlook clients are connecting to =

Exchange databases pointing to

Best regards,

Oz Casey , Dedeal

MCSE 2003, M+, S+, MCDST
Security+, Project +, Server +

Http:// (Blog

Http:// (Blog)

Http:// (Blog


Frank Carius said...

I did a similar thing. (german description at But i simply added a third IP-Adresse as Cluster Ressource and if the "active" server fails, the IP-Adresse is transferred to the remaining node. So the third IP-Address is using the "alias" name of the cas array. This allowes an automatic failover.

Brian said...

Two things:

--> TTL is still important in a small environment

--> A DB failover doesn't necessitate a change in RPC endpoint. A complete server failure however (in your diagram), does.

Anonymous said...

1.Needs your help regarding configuration
I create a CAS array name as for internal and external use, set the databases with this cas array.Set the RPCclientserver
create a DNS entery as with same IP address of Server
Owa is working fine but MAPI clients are unhappy and did not connected
Any suggestion ??