The easiest way is to simply change the internalautodiscoverServiceURI value to the same value on all servers. This value could be one specific server or a general name. As long as it resolves and the host name is in the SSL certificate, it will work fine. Simon.Simon Butler, Exchange MVP Blog | Exchange Resources | In the UK? Hire Me.

Thanks for your input Mr Butler, Just so you know, we use the CAS array "or to be more precise RPC CAS array" to manage MAPI traffic but use WNLB for smtp, pop imap and web as well and it has worked pretty well so far. If you didnt notice Kenneth was concerned if the cas server would be able to handle the re-directed traffic from another CAS server so my suggestion to introduce a CAS array was for resiliency purpose. Last time i checked, you cannot create a cas array without NLB and my suggestion was create a cas array and load balance the traffic NOT create a cas array to load balance the traffic. I don't know what information you are reading, but it is wrong. There is no requirement to use any kind of load balancing with an RPC CAS Array. The RPC CAS Array is just a DNS entry and modification to the database configuration, therefore you can (And should) use an RPC CAS Array with all deployments, whether single or multiple server. The reason I use the RPC CAS Array name is because of the confusion over its role. It is JUST MAPI traffic that uses the RPC CAS Array, nothing else. While you can use the same host name, it is not recommended. The fact that it works for you doesn't mean it is the correct thing to do. Microsoft actually recommend AGAINST using the same FQDN for everything as it has an impact on Outlook Anywhere: "It's important that the (FQDN) specified in the command be only resolvable internally. If the name is also resolvable externally, these external clients will attempt to connect to the array via a TCP connection instead of HTTPS. " http://technet.microsoft.com/en-us/library/ee332317.aspx Not picking on you, but there is a lot of misinformation about the role of the RPC CAS Array and its functionality, and when something wrong is posted, I will correct it to ensure that others do not make the same mistake. Simon.Simon Butler, Exchange MVP Blog | Exchange Resources | In the UK? Hire Me.

fair enough.. i think i used "CAS Array" term incorrectly so thanks for pointing that out.. What i meant was to load balance autodiscover traffic which i don't think is not recommended. (Not talking abt FQDNs, never did) We use same FQDN for RPC, pop , imap and outlook anywhere. Works like a charm.. and if you have split DNS, then even though the client may try to connect using TCP first (which the will do anyway regardless of the FQDN), they will use HTTPS eventually and the delay caused by same FQDN is not worth not implementing.

Can I use set-clientaccessserver -internalxxxx but point to other server, can it handle it?

The easiest way is to simply change the internalautodiscoverServiceURI value to the same value on all servers. This value could be one specific server or a general name. As long as it resolves and the host name is in the SSL certificate, it will work fine. Simon.Simon Butler, Exchange MVP Blog | Exchange Resources | In the UK? Hire Me.

Is it possible choose which CAS server to handle autodiscover We have 3 CAS server in the same site, we dont want to user connect specific 1 CAS server. I know Outlook will use SCP to get the CAS server from AD, but I cannot find how to remove 1 record. Actually I think we simply shutdown that CAS server, user also wont point to that server also and would pop up alert msg

Yes, use set-clientaccessserver -internalxxxx and change the internalautodiscoverServiceURI. If you have 3 CAS servers, why not create a CAS array and load balance the traffic?

Hi, 1) go to ad sites and services, 2) expand services (to view services, click view and select view services node.) 3) expand microsoft exchange 4) expand your org, 5) expand administrative groups 6) expand exchange adm.. group. 7) expand servers 8) expand your cas server, then protocols, then select autodiscover. 9) right click service connection point and choose properties. 10) in attribute editor tab, you can view/change/remove it. To check where your outlook is connecting, open outlook, press control key and right click outlook icon in the taskbar and select test email autoconfiguration. uncheck guessmart option and click test (enter your pass as well). then click the log tab for info.

Yes, use set-clientaccessserver -internalxxxx and change the internalautodiscoverServiceURI. If you have 3 CAS servers, why not create a CAS array and load balance the traffic? You don't understand what a CAS array is used for. The CAS array, or to be more precise, the RPC CAS array, is used for MAPI traffic only. It should not be used for anything else. Furthermore an RPC CAS Array on its own doesn't do any kind of load balancing. You could use a load balancer or WNLB to manage the traffic and this can use the same IP address as the RPC CAS Array, but it should not use the same host name. The host name used for the RPC CAS array should be unique to that function and should NOT resolve on the Internet and as it isn't used for HTTPS traffic it does not need to be in the SSL certificate. Simon.Simon Butler, Exchange MVP Blog | Exchange Resources | In the UK? Hire Me.

Thanks for your input Mr Butler, Just so you know, we use the CAS array "or to be more precise RPC CAS array" to manage MAPI traffic but use WNLB for smtp, pop imap and web as well and it has worked pretty well so far. If you didnt notice Kenneth was concerned if the cas server would be able to handle the re-directed traffic from another CAS server so my suggestion to introduce a CAS array was for resiliency purpose. Last time i checked, you cannot create a cas array without NLB and my suggestion was create a cas array and load balance the traffic NOT create a cas array to load balance the traffic.