locked
Troubleshoot Slow Mail Delivery - Outlook 2010

    Question

  • One of my clients has raised an issue which is affecting some but not all users.  All are on Win 7/ Outlook 2010 - Exchange 2010 SP1 RU3. All are in the same physical site, spread over different databases in different DAGs.

    The symptom is slow mail delivery to the full Outlook 2010 32bit and 64bit client.  This started on Monday.

    I can watch a test email hit the Exchange server mailbox database for a user, and also watch the item count change for the user mailbox through the EMS.  The user also receives the message in OWA at the same time and on his BB.

    However, the Outlook client does not receive the message; this can range from a few minutes to up to an hour.  A manual Send/Receive does not help and the Auto Connection test shows the client is being fully connected.  Sometimes the mail will come in a burst of 5 or 6 messages showing a time range of up to an hour.

    I figured the next thing to do would be to enable diagnostic logging on the Outlook client and read the contents of the OPMLOG.log.  By doing this, i should be able to view transport requests made by the Outlook client and what it sees in the mailbox database for the user.  If i know it's already in OWA and on his BlackBerry, then the outlook client should show that the message has been recognized on the server and then being downloaded.

    However, when i compare the time received in OWA etc., the Outlook OPMLOG.log shows no messages as having been found.

    Can someone point me in the direction of further troubleshooting here, or am i missing something in the way OPMLOG.log displays server requests.  If the item is in the database, why wouldn't the outlook client see it? 

    What's happening  on the MAPI connection with the RPC endpoint that i can look at, which should show any latency or communication issues?  The CAS servers are load balanced with the same VIP.



    • Edited by MbxMan Thursday, June 2, 2011 5:40 PM additional info
    Thursday, June 2, 2011 5:15 PM

Answers

  • Update Roll-up 3 v 3 fixed the issue immediately.  We had a support call opened and sent back 2 days worth of affected client RPC traces.  Once debugged this showed that the server couldn't contact the affected clients though the feedback from the MS Engineer was still a little hesitant.  HE did tell us that although unpublished, one of the notes on his briefing for Update Roll-up 3v3 was a fix for this type of issue.

     

     

    • Marked as answer by MbxMan Sunday, July 3, 2011 4:46 PM
    Sunday, July 3, 2011 4:46 PM

All replies

  • One of my clients has raised an issue which is affecting some but not all users.  All are on Win 7/ Outlook 2010 - Exchange 2010 SP1 RU3. All are in the same physical site, spread over different databases in different DAGs.

    The symptom is slow mail delivery to the full Outlook 2010 32bit and 64bit client.  This started on Monday.

    I can watch a test email hit the Exchange server mailbox database for a user, and also watch the item count change for the user mailbox through the EMS.  The user also receives the message in OWA at the same time and on his BB.

    However, the Outlook client does not receive the message; this can range from a few minutes to up to an hour.  A manual Send/Receive does not help and the Auto Connection test shows the client is being fully connected.  Sometimes the mail will come in a burst of 5 or 6 messages showing a time range of up to an hour.

    I figured the next thing to do would be to enable diagnostic logging on the Outlook client and read the contents of the OPMLOG.log.  By doing this, i should be able to view transport requests made by the Outlook client and what it sees in the mailbox database for the user.  If i know it's already in OWA and on his BlackBerry, then the outlook client should show that the message has been recognized on the server and then being downloaded.

    However, when i compare the time received in OWA etc., the Outlook OPMLOG.log shows no messages as having been found.

    Can someone point me in the direction of further troubleshooting here, or am i missing something in the way OPMLOG.log displays server requests.  If the item is in the database, why wouldn't the outlook client see it? 

    What's happening  on the MAPI connection with the RPC endpoint that i can look at, which should show any latency or communication issues?  The CAS servers are load balanced with the same VIP.

    Thursday, June 2, 2011 5:12 PM
  • Update:  We were told to update to Rollup Update 3 (v3) fir Exchange 2010 SP1.  Apparently this fixes a bug inwhich Outlook 2010 clients are skow to receive mail.  Currently have a case open with MS to troubleshoot and have been uploading Advanced RPC logging trace results as described here:  <cite>support.microsoft.com/kb/2260527 </cite>

    Some more behaviour:

    • doing a manual send/receive does not work
    • pressing Reconnect forces mail to download
    • deleting or forwarding an existing email seems to trigger a reconnect and subsequent download of waiting messages

    I'll post the final results.

    Monday, June 6, 2011 5:36 PM
  • Did RU 3v3 fix your issues ? I'm having the exact same issues with delays just starting to happen in my environment. Single Exchange 2k10 server with OLK 2K10 X86 clients.
    Thursday, June 30, 2011 9:31 PM
  • Update Roll-up 3 v 3 fixed the issue immediately.  We had a support call opened and sent back 2 days worth of affected client RPC traces.  Once debugged this showed that the server couldn't contact the affected clients though the feedback from the MS Engineer was still a little hesitant.  HE did tell us that although unpublished, one of the notes on his briefing for Update Roll-up 3v3 was a fix for this type of issue.

     

     

    • Marked as answer by MbxMan Sunday, July 3, 2011 4:46 PM
    Sunday, July 3, 2011 4:46 PM
  • Have this issue with Exchange 2010 SP2 RU4, anyone else?

    d

    Thursday, October 11, 2012 8:04 PM
  • Hi,

    we are having the exact same issue, Running SBS2011. Recently applied the RU4 update but made no difference.

    Any ideas?

    Monday, November 5, 2012 3:58 PM
  • I have seen this issue if you are using a hardware load balancer.  You'll need to have the LB persistence and tcp timeouts increased.  For F5 LB's it's recommended to increase RPC static ports to 7200 seconds.  I believe the defaults were set to 500 seconds.

    d

    Monday, November 5, 2012 4:11 PM
  • Have this issue with Exchange 2010 SP2 RU6 with no load balancer.

    Ideas?


    ~bc

    Wednesday, April 24, 2013 8:19 PM