In Exchange 2010, the Client Access Servers now handle all client traffic now (MAPI, Outlook Anywhere, IMAP, POP, OWA, etc.). In Exchange 2007, you could create a hardware load-balanced array of CAS servers, but MAPI connections would always go directly to the mailbox server where the user resided. With Exchange 2010 Database Availability Groups, a user’s mailbox can be moved around between mailbox servers within the DAG. To enable this functionality without causing a major disruption to clients, they connect to an array of CAS servers. A new feature with Exchange 2010 allows you to set a Client Access Server array at the database level. The need for these CAS arrays is critical if you plan to fail between mailbox servers in the DAG.
The following are some quick steps to setting up your CAS Arrays:
- Configure your Client Access Servers behind a load-balancer (HW or Windows NLB). We use F5 Local Traffic Managers (BigIP) in the same site and Global Traffic Managers across sites (3DNS).
- Create an A record that resolves to the Virtual IP Address (VIP) of the CAS load balancing array. For example, outlook.servusinc.org.
- Configure your load balancing array to load balance the MAPI RPC ports:
- TCP 135
- UDP/TCP 6005-65535; or set static ports
- Using the New-ClientAccessArray cmdlet create a new CAS array (More info):New-ClientAccessArray –Name “DFW CAS Array” –Fqdn “outlook.servusinc.org” –Site “DFW”
- Once this is complete you need to configure each Exchange 2010 database that is not configured with the CAS Array. The RPCClientAccessServer property on the database needs to be configured with the CAS array FQDN. By default, the first Exchange 2010 Client Access Server is assigned to the RPCClientAccessServer property on all DBs.Set-MailboxDatabase DB1 -RpcClientAccessServer “outlook.servusinc.org”
For more on this topic click here.