When opening Outlook 2003, users have to wait as much as one minute before being connected to the Exchange server.As soon as Outlook is in connected mode, everything works fine as designed. Restarting Outlook result in an instantaneous connection time after time.The problem only happen when Outlook has been shut down for a while. Starting Outlook after aprox. 2 hours always result in the same connection issue... I have to wait almost 2 hours before each try. Don't know how to make it happen faster.In the computers event log, Outlook ID 19 is generated each time after the connection issue happen: RPCCall(RfrGetNewDSA) to server (...) Failed with error code (5b4) after waiting 45000mS... I also found randomly in the computers event logs some Outlook ID 26 messages: Connection to the Microsoft Exchange Server has been lost. Outlook will restore the connection when possible..Very often, Users are experiencing delays while accessing a network share (using DFS).Here is the hardware details:Both Exchange Server 2003 and the domain controller are running on a virtual Windows 2003 Enterprise R2 server sp2 (VmWare 2.0) and are hosted on a HP Proliant DL380 G5 Fileserver.Workstations are WinXp sp3 with Office 2003.Network is made up of 4 sites. Each site is composed of a physical server hosting a virtual Domain Controller (global catalog), Exchange server and a Data Server (Oracle) and 30 workstations.Copying files to and from the Exchange server is fast. Not a bandwidth issue. Same for the Domain Controller.Host server: NIC teamig, TOE, RSS have been disabled.Drivers up to date (I know about these NIC issues and everything is fine.)Tried different NIC settings without success: Force 100Mb/sec etc etc...Problem is happening on different sites with different network infrastructure and switch.Tried different NIC on the test workstation.Removed Anti-Virus sorfware onto Exchange server, Domain Controller and test workstation.I tried a LOT of things without success and almost ruled out the network layer issue. I'm running out of time. Please Heeeeeeeeeeeelp!

So image that Outlook has a list of GCs that it can talk too. The command RfrGetNewDSA is making a call to refresh this list thus talking to every GC in the list trying to determine the protocols each GC will support. Since you are running Exchange 2003 SP2 the communication is not only happening to GCs within the same ADSite as the clients but it will also talk to any GCs with ADsite connectors to the ADSite the client is a member of. Now you ask why does this RfrGetNewDSA happen so much? Well this is an uncommon, when the RfrGetNewDSA is called normally something changed like a DC coming online and then going offline (or the network connection issue to GC at another ADSite). More information of RfrGetNewDSA http://msdn.microsoft.com/en-us/library/ee218508(EXCHG.80).aspx Having worked with Exchange and Outlook for many of years the one word that makes me cringe the most is Teaming. I can’t tell you how many times during a support call we have looked at network packets and noticed random packets are getting dropped between the Exchange server, Client and the DC because of network teaming. I’m not saying that is your issue but it is a good place to start. You will want to get a network capture between the Outlook Client, Exchange Mailbox Server, and DC at the same time. Here is a good blog that talks about setting up a network capture https://cbfive.com/blog/post/Taking-a-long-network-capture.aspx and this blog talks about Network Teaming https://cbfive.com/blog/post/Considering-Network-Teaming.aspx. If you need help looking over the network capture you might want to ping Jared via his blog, he is a former SME (subject Matter Expert) at MS on network issues. Even if is not a Teaming issue, it could be a host of other issues from the network so getting a network capture is still a good idea. Other options - Take a look at your ADSite topology and see if there is a known issue with one of the site links or a DC in another site. you can look for event 2080 on the Exchange server to see what DCs Exchange and Outlook will talk too. - Change the DSAccess servers http://www.msexchange.org/tutorials/Closer-Look-Directory-Service-Access-DSAccess-Part1.html. Changing this setting will limit the DCs Exchange and Outlook will talk to. - Change the preferred DCs for the Outlook client http://support.microsoft.com/default.aspx/kb/319206 Note: If you choose the last option please let me know I will write a quick blog at our site http://cbfive.com/blog on how to push this change on all computers If you have any questions please let me know Chris cbfive.com

Great answer!Effectively, remotes sites could be unavailable during some time: These are connected via a unreliable and slow satellite link who lose connectivity many times a day. (no other option...) I was expecting to keep the "local" traffic in his respective site... These links are only used for replication purpose and everything is fine about that.Currently under test...Thanks

Did that work out for you?Chris cbfive.com

Hi,Did you test this issue on another machine for the same user?Please click Start, click Run, input Outlook /RPCDiag and press Enter, the Microsoft Exchange Connection Status UI will be displayed, check which connection is slow in the Microsoft Exchange Connection Status and confirm whether the currrent connected Server name is in your site.ThanksAllen

Solution still under test... (Changing the DCs Exchange and Outlook will talk too)Actually it works very well on my test site. I'm about to deploy the solution on a remote site and confirm that everything is fine.I will let you know pretty soon about the complete setup.User's are also experiencing some random slow file access: Using DFS namespace so it's a good start for me to look at the same problem caused by slow or unavailable links between DCsThanks so much!

ping back any updates?check out my blog at http://cbfive.com/blogChris cbfive.com

Problem fixed! You were right. It had something to do with the list of DCs that Outlook can talk too. By default, each of them are in «auto config». I simply removed each DC from the list and put them back again after unchecking «automatically discover servers» These are now manually configured. So here what i have done: In the Exchange System Manager, under first administrative groups, go the properties page of the mailserver. Click on the directory access tab. Take a look at the «automatically discover servers» option. By default, this checkbox is greyed out and you need to remove this option. There is also the word «Auto» under «Type» column in each DC's row. Click on the drop down menu and select «Configuration domain controller» Uncheck «automatically discover servers» Following this and the warning message, The list will be left empty so you need to manually re-enter the DC (or DC's) that was previously appearing in the list Using the same drop down menu, Repeat the same procedure for «Domain controllers» and «Global catalog servers» At the end, under «All domain controllers» drop down menu, you will be left with the same list of DC's but the «Auto» config will no longer show. Repeat the above procedure for each mailservers of the organization. Thanks so much for all those precious tips!