RPC/HTTP Ports
I have a CAS running Exchange Enterprise 2007 SP2 on Windows 2008 Enterprise SP2. I enabled RPC/HTTP on the CAS server but I could not connect using AutoDiscover when I tested it from an Outlook 2007 client outside of our network. I found out about the Windows 2008 issue with IPV6, so I disabled it as suggested. However, the problem seemed bigger than that because when I ran netstat -a, I saw that the server was not listening on ports 6001, 6003 and 6004 as they should be.As it turns out, the problem had nothing to do with the CAS server, but with the ISA server publishing rule because I had not added "autodiscover.company.com" as a public name. (If I had been able to test it on an Outlook client on the LAN I would have known much sooner.) So everything is working now, but I wasted time pursuing the CAS server as the problem because it apparently wasn't listening on the correct ports. I'm just wondering if anyone knows why it wouldn't be listening on ports 6001, 6003 and 6004. I looked at a different CAS server with RPC/HTTP working, and it was listening on those ports. The difference is that the OS is Windows 2003 and I believe Exchange 07 is at SP1.Thanks in advance if anyone has the answer.
January 7th, 2010 12:31am

You will certainly see the mailbox servers listening on these ports:RPC Endpoint 6001 (Exchange Information Store)RPC Endpoint 6002 (Referral Interface)RPC Endpoint 6004 (NSPI Proxy Interface)But as far as I know, a single CAS server will only listen for this traffic on port 443 and pass it on to the mailbox servers.This is also illustrated here (BTW, highly recommended):Configuring Outlook Anywhere for Exchange 2007 SP1http://www.exchange-genie.com/2008/02/configuring-outlook-anywhere-for-exchange-2007-sp1/ I don't think it is related to OS or Exchange patch level.However, on CAS servers with NLB, you'll see the traffic (Win2008 / SP1, Exchange 2007 /SP1 RU9):This traffic is not from an outside client but from the NLB interface to the NIC of one node in the NLB cluster (netstat -na):Proto Local Address Foreign Address State TCP 10.13.69.81:40179 10.13.69.28:6001 ESTABLISHED TCP 10.13.69.81:40180 10.13.69.28:6001 ESTABLISHED TCP 10.13.69.81:40181 10.13.69.28:6001 ESTABLISHED TCP 10.13.69.81:40182 10.13.69.28:6001 ESTABLISHEDTCP 10.13.69.81:40185 10.13.69.28:6004 ESTABLISHED MCTS: Messaging | MCSE: S+M | Small Business Specialist
Free Windows Admin Tool Kit Click here and download it now
January 7th, 2010 2:00am

Thanks for the complete answer. I should have realized that the connections to the 600x ports would be on the Mailbox server, but I was concentrating on the CAS as the problem, especially because I read about the IPV6 problem as being on the CAS server. Also, the RpcProxy Reg key with those ports listed exists on the CAS server. Though if I had taken a closer look I would have seen that the server name listed next to the ports is the name of my mailbox cluster.The reason I could see those ports in netstat on the Windows 2003 Exchange server is because that server has all of the Exchange roles.Cheers.
January 7th, 2010 3:27am

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics