none
DAG clients not working after failover

    Question

  • Hello

    I got a strange issue at a cusomer location. Still did not get the chance to have a look myself, but I was thinking
    maybe someone may have an idea what is the issue.

    Configuration:
    A single site with two mailbox servers in a DAG, Two CAS/Hub servers configured in a CAS Array using NLB.
    There is a single mailbox database for now.
    Clients are Outlook 2007 and Outlook 2010

    When Moving the active database from mailbox server A to mailbox server B clients are tansfered correctly
    and users do not notice that the database has been moved (administrator initiated activation).
    However, then the same database is moved back from mailbox server B to mailbox server A Outlook 2007 clients
    stay connected but no mail is going in or out of their mailbox until Outlook is restarted, and Outlook 2010 clients
    are getting prompted to restart outlook due to administrator initiated change.
    So this only happens when moving the database from one specific mailbox server to the other.

    I guess that the cause may be due to some sort of issue with the activation manager not updating the CAS servers
    of the change.
    Where should I be looking for ?
    Any ideas ?

    Thanks
    Liran Zamir 

    Saturday, August 21, 2010 10:09 AM

Answers

  • Thank you everyone.

    It seems that a simple restart to the problem mailbox server resolved the problem.
    The two servers have the same rollup.
    I just hope this will not return.

    Thanks again

    Sunday, August 22, 2010 11:33 AM

All replies

  • If you stop a node in the NLB and leave one node active as a test , does it work correctly then?

     

    Run ExBpa as well. Anything in the event logs?
    Saturday, August 21, 2010 1:20 PM
    Moderator
  • I asked the cusotmer to create a static entry in the hosts file to point the a client PC to a specific CAS server
    instead of the NLB resolved DNS IP address of the CAS array, and the result of the test was exactly the same.

    I will ask the customer to run the ExBpa and review the event logs.

    Thanks

    Saturday, August 21, 2010 5:59 PM
  • I have heard of this before. are they on the latest role up. I seem to remember that this would happen at RTM but has been fixed with a Roll up
    Mitch Roberson |MCITP:Enterprise Server Admin, Messaging 2007, 2010 |MCTS:OCS with Voice Achievement |MCT |MCSE 2000\2003 |MCSE Messaging 2000\2003
    Saturday, August 21, 2010 9:33 PM
  • What if they actually stop one of the nodes?

     

    Saturday, August 21, 2010 11:03 PM
    Moderator
  • Thank you everyone.

    It seems that a simple restart to the problem mailbox server resolved the problem.
    The two servers have the same rollup.
    I just hope this will not return.

    Thanks again

    Sunday, August 22, 2010 11:33 AM