Problems connecting to Exchange over a WAN

Our problem is that the users in our remote office lose connectivity with our Exchange server, which is located in the main office. The only way they seem to be able to reconnect is to reboot their computer.

We have a small WAN – just  2 offices located in different towns.  They are connected with a point to point T1 using Cisco 1720 routers.

Our problem is that the users in our remote office lose connectivity with our Exchange server, which is located in the main office.  The only way they seem to be able to reconnect is to reboot their computer.  Our Exchange server is running Windows 2000 Server and Exchange Server 2003.  We were running Windows 2000 Server and Exchange 2000 but recently updated to Exchange 2003.  When the remote users lose connection to the Exchange server they can still access their documents that are located on a Windows 2003 server in their building, they just cannot access resources in the Main Office.

The users in the main office do not have any problems maintaining a connection with the Exchange server.  Your help would be appreciated.

– Bryan Lawrence

At first glance, the problem you are describing an RPC (remote procedure call) problem between the workstations and the Exchange server.  Since everything was fine prior to the upgrade to Exchange 2003, we will need to go step by step to isolate the cause of the problem. 

I will assume that you have the latest updates to Windows 2000 and Exchange 2003 applied to the server in question.  The same applies to the workstations in terms of the version of windows and Outlook that they are using.  See if there have been any other changes made lately that would be unique to the remote office that would seem to point the problem to that end.  Start keeping a log as to the time of day you are seeing the problems, that might also help identify the source of the problem.  Look at the event log on the Exchange server when the problem is occuring to see Exchange is seeing the problem.

When the problem first manifests itself and before you reboot the workstation, try pinging the servers IP address to make sure you have a good communications path between the workstation and the server.  If that works, next try pinging by NetBIOS name and by fully qualified DNS name (for example. mail.mynetwork.com).  If you are still OK, it would seem to indicate to points a RPC problem.  

I would suggest looking at TID 175496 – XCON: Using RPCPING To Troubleshoot MTA Connections.  This TID and the rpcping tool that is mentioned will give you additional tests to run to further isolate the problem.  The tools this TID and similar ones talk about are found on your Exchange server CD in the SUPPORTRPCPING directory.  There are other RPC related TID’s that you will be able to find using Google or other search engines that may also help shed some light on the problem you are having.

Another thing to look at is how your workstations talk to the local server at their office and the Exchange server at the main office. 

If they are having to pass through the local server to get to the Exchange server at your office, try setting things up so that they can go direct without having to pass through the remote server to get to mail. 

Telnet into the Cisco router when this problem is going on to see if there are any errors being reported that might indicate the cause of the problem.  Try the route print command at the remote workstations before and during the problem to see if the IP routes are getting changed and causing the problem.