SCCM 2007 R3 on Windows 2008 R2 / SQL 2008 R2. Site Mode: Mixed Mode. We have no PKI. 38 servers have the connection type: Always intranet. Our two Exchange 2010 servers with a DAG (cluster) have the connections type: Currently Internet. I think (but do not know at all) that the issue with the Exchange servers might be related to multiple subnets, one for MAPI, one for replication and one for the Microsoft Failover Cluster Virtual Adapter (169.254.y.z). That's the only difference between the Exchange servers and the other servers I can see. Anyhow, there is no support for Internet clients. How do we change the connection type? Add the 169.254.0.0/16 range to the AD Site perhaps?MCTS: Messaging | MCSE: S+M

SCCM 2007 R3 on Windows 2008 R2 / SQL 2008 R2. Site Mode: Mixed Mode. We have no PKI. 38 servers have the connection type: Always intranet. Our two Exchange 2010 servers with a DAG (cluster) have the connections type: Currently Internet. I think (but do not know at all) that the issue with the Exchange servers might be related to multiple subnets, one for MAPI, one for replication and one for the Microsoft Failover Cluster Virtual Adapter (169.254.y.z). That's the only difference between the Exchange servers and the other servers I can see. Anyhow, there is no support for Internet clients. How do we change the connection type? Add the 169.254.0.0/16 range to the AD Site perhaps? MCTS: Messaging | MCSE: S+M Those Exchange servers, are they in the same DNS namespace, AD forest?

Yes, single domain in a single forest. Naturally, same DNS name space.MCTS: Messaging | MCSE: S+M

So far, the following procedure has solved the issue with the second Exchange DAG member: (1) Adding the APIPA range (169.254.0.0/16) to the subnets belonging to the site (Active Directory Sites and Services). (2) Running a Repair on the Components tab in Configuration Manager 32-bit Control Panel app on the Exchange server. (3) Rebooting the server. Step 2 turned the ConfigMgr Connection Type to Unknown. Step 3 turned the ConfigMgr Connection Type to Always intranet. And now the server shows up in the Config Mgr Console with an agent and a lot more options in Control Panel. However, this procedures has not helped with the first member of the Exchange DAG ... Right now, I'm trying with ccmsetup.exe /uninstall. What is a Microsoft Failover Cluster Virtual Adapter anyway? http://blogs.technet.com/b/askcore/archive/2009/02/13/what-is-a-microsoft-failover-cluster-virtual-adapter-anyway.aspxMCTS: Messaging | MCSE: S+M

After the uninstall, a repair and twice restarting the SMS agent, the issue was solved on the first Exchange 2010 server as well. I tried all this yesterday night for several hours without any success. This makes me believe that adding the APIPA range to the site's subnets, which I first attempted today, contributed to the solution. Still not sure, though. On the other hand, these are the only Windows 2008 clustered servers. Would be glad to hear about others who have run into these problems.MCTS: Messaging | MCSE: S+M

May be something to do with overlapping of boundaries?Anoop C Nair - Twitter @anoopmannur MY BLOG: http://anoopmannur.wordpress.com SCCM Professionals This posting is provided AS-IS with no warranties/guarantees and confers no rights.

I've only defined one boundary. It's of type Active Directory site. We do only have one site in terms of Active Directory, stretched over two locations. The MAPI and replication network was defined in AD Sites and Services, but not the APIPA network used by Windows 2008 clustering. Therefore I believe that SCCM got confused, saw a network that was not part of the boundary and thus configured the connection type as: Currently Internet. I really don't know. I'm very new to SCCM ...MCTS: Messaging | MCSE: S+M

A client will switch to "currently internet" if it cannot reach it's default management point. It has nothing to do with boundaries.Torsten Meringer | http://www.mssccmfaq.de

This explanation makes sense. Multi-homed servers have often caused problems. Neither the replication nor the APIPA network (Microsoft Failover Cluster Virtual Adapter) have a default gateway. So the management point cannot be reached from those networks. The MP can however be reached from the MAPI network, actually on the same subnet. If this is the cause, how do you make sure that the client is communicating from the right network? Could the client suddenly by its own whimsy move back to "Currently Internet"? MCTS: Messaging | MCSE: S+M

Your site is in mixed mode ... how did you install the client? Which (client push) parameters where used?Torsten Meringer | http://www.mssccmfaq.de

Is that possible that client can't reach default management point because of overlapping ?Anoop C Nair - Twitter @anoopmannur MY BLOG: http://anoopmannur.wordpress.com SCCM Professionals This posting is provided AS-IS with no warranties/guarantees and confers no rights.

Your site is in mixed mode ... how did you install the client? Which (client push) parameters where used? Torsten Meringer | http://www.mssccmfaq.de Yes, mixed mode. Did it manually: \\sccm\sms_xyz\Client\CCMSetup.exe /mp:SCCM /logon SMSSITECODE=XYZ This worked for all servers, except the two Exchange servers. MCTS: Messaging | MCSE: S+M