none
Load Balance Cross Forest Mailbox Moves (Exchange 2007 to 2010)

    Question

  • Hi All,

    Just curious as to what the best method of load balancing Exchange 2007 to 2010 mailbox moves cross forest will be. When Irun the New-MoveRequest command from a CAS server does that mean that the moves will all take place through that CAS server or will Exchange load balance accrosse the available servers?

    Thanks,

    Michael.

    Friday, July 11, 2014 6:39 AM

Answers

  • I don't' remember it does load balance the migration batch. I do remember I did between two Exchange 2010 mailboxes and i was limited to 3 concurrent connection and the more I increased the session the worst it become.


    Where Technology Meets Talent

    Friday, July 11, 2014 8:59 PM
  • Hi,

    Yes, move requests can handle transient errors. MRS conducts checkpoints every 5 minutes to make sure that the database to which the mailbox being moved is still operational. If MRS finds that the target database isn't operational, MRS will pause for 30 seconds, and then retry the move. If you experience a failover, the move won't fail. Instead, MRS will detect a database failover, determine the new location of the database, and then restart the move process.

    Another error that could occur is if the Client Access server on which MRS is running stops responding. If this happens, the move stops, and one of the other MRS instances will continue the process and complete the move.

    For more details, I recommend you refer to the following articles:

    Understanding Move Requests

    Hope it helps!

    Thanks.


    Niko Cheng
    TechNet Community Support

    Tuesday, July 15, 2014 2:25 AM

All replies

  • I don't' remember it does load balance the migration batch. I do remember I did between two Exchange 2010 mailboxes and i was limited to 3 concurrent connection and the more I increased the session the worst it become.


    Where Technology Meets Talent

    Friday, July 11, 2014 8:59 PM
  • Hi,

    Thanks for the reply, however I am not too concerned about throughput it is really the load balancing component I am concerned about. I am assuming that I will need to create batches to be run on each 2010 CAS server to maximise throughput? 

    Regards, 

    Michael. 

    Friday, July 11, 2014 10:12 PM
  • Hi,

    No, you just need to edit the following file on every cas server to maximum throughtput:

    MSExchangeMailboxReplication.exe.config

    Under the MRSConfiguration section change these fields into a desired value:

    MaxActiveMovesPerSourceMDB = "25"
    MaxActiveMovesPerTargetMDB = "25"
    MaxActiveMovesPerSourceServer = "50"
    MaxActiveMovesPerTargetServer = "25"
    MaxTotalMovesPerMRS = "100"
    

    After changing, we must restart the Microsoft Exchange Mailbox Replication Service.
    From Powershell:

    restart-service MSExchangeMailboxReplication

    Hope this helps!

    Thanks.


    Niko Cheng
    TechNet Community Support

    Monday, July 14, 2014 10:02 AM
  • Hi Niko, 

    I have already edited the MRS proxy settings to maximise throughput. 

    So you are saying that I can run the new-moverequest command from a single CAS server and the load will be distributed across all available CAS servers? 

    I may be moving up to 250 mailboxes in a night so will need to load balance the moves as much as possible. 

    Regards, 

    Michael. 

    Monday, July 14, 2014 11:07 AM
  • Let me know how does it go.. curiously waiting to know.

    Where Technology Meets Talent

    Monday, July 14, 2014 3:55 PM
  • Hi,

    Yes, move requests can handle transient errors. MRS conducts checkpoints every 5 minutes to make sure that the database to which the mailbox being moved is still operational. If MRS finds that the target database isn't operational, MRS will pause for 30 seconds, and then retry the move. If you experience a failover, the move won't fail. Instead, MRS will detect a database failover, determine the new location of the database, and then restart the move process.

    Another error that could occur is if the Client Access server on which MRS is running stops responding. If this happens, the move stops, and one of the other MRS instances will continue the process and complete the move.

    For more details, I recommend you refer to the following articles:

    Understanding Move Requests

    Hope it helps!

    Thanks.


    Niko Cheng
    TechNet Community Support

    Tuesday, July 15, 2014 2:25 AM