22 Juni 2012 13:53
We're currently on one Exchange 2010 server. I've deployed a second one and currently in the process of configuring a CAS array with HLB. Once the VIP is configured on the HLB, I'll change the mail.domain.com A record to point to the VIP.
My question is (if the above is the correct approach):
How will the Outlook clients react? Currently, they auto-detect the mail server (let's say ex1.domain.com) without internal autodiscover. All client machines are domain joined, when we migrated user mailboxes to Exchange 2010 (ex1.domain.com) it was able to automatically detect the correct server. I'm wondering if Outlook will do the same when I configure the CAS array/HLB, and point mail.domain.com to the VIP instead of ex1.domain.com
22 Juni 2012 16:16
When you configure an RPC CAS Array, nothing happens to existing clients. That is one of the major pain points with the RPC CAS Array. For the clients to start using the RPC CAS array, one of three things has to happen.
1. A new Outlook profile is created.
2. The existing Outlook profile is "repaired".
3. The server becomes completely unavailable and the autodiscover process kicks in.
There have also been some changes around the change of server address in the latest rollup for Exchange 2010 (SP2 rollup 3), see the msexchangeteam blog for details.
Outlook clients are using internal autodiscover, whether you like it or not. That is the only way that they pick up the configuration automatically (unless you are using something to create the profile). However once they are configured, the Exchange server in use isn't changed by autodiscover, unless something changes.
The recommendation I would make is that you first configure the RPC CAS array to point to one of the servers with the CAS role directly, verify it works correctly before introducing the load balancer. Once you know it works, then change the configuration to point the DNS entry to the load balancer. If it fails, point it back for testing.
- Ditandai sebagai Jawaban oleh Zi FengMicrosoft Contingent Staff, Moderator 02 Juli 2012 3:20
22 Juni 2012 16:30
Just be aware that by default the clients will query for a SCP record in AD there is one created by Each CAS server this is where the confusion lies when people talk about autodiscover. They think if I do not have autodiscover.domain.com then i am not using it. But when a CAS is installed it creates an SCP record in AD and puts itself as the URL for autodiscover. This is why some people seem to think they have to have the server name in the cert. But you can change the autodiscover URL used by the SCP record. And have both records point to the same URL.
if you do a get-clientaccessserver |fl Server,*uri* it will tell you what URL it is using. And yes that is URi not URL in the command. once you know that then you can set them by do a set-clientaccessserver.
So what this means is that when you introduce that second CAS you may see users recieving a Cert popup Randomly until you put proper certs on it. or Change the SCP autodiscoveruri to point to the Autodiscover URL listed in the other server.
Mitch Roberson MCM Exchange 2010|MCITP:Enterprise Server Admin, Messaging 2007, 2010 |MCTS:OCS with Voice Achievement |MCT |MCSE 2000\2003 |MCSE Messaging 2000\2003
- Ditandai sebagai Jawaban oleh Zi FengMicrosoft Contingent Staff, Moderator 02 Juli 2012 3:19
22 Juni 2012 16:41Very nice info Mitch !!!
Anil MCC 2011,ITIL V3,MCSA 2003,MCTS 2010, My Blog : http://messagingschool.wordpress.com