none
How to get a backup Send Connector to take over

    Question

  • We have 2 send connectors configured for outbound internet mail, both with smarthosts, and they both work but we can't get one to take over if the other fails.  We have one that is considered the primary, and the other a secondary.  But when the primary is unavailable or we have exceeded our quota for the day, the queue goes into a Retry state, and mail just sits there but we need it to go out the other send connector.  We have to disable the failed connector to get the mail to go out the other connector.

    The "primary" Send Connector is configured with * and a cost of 1:

    {smtp:*;1}

    The "secondary" Send Connector is configured with * and a cost of 10, along with comcast.net since that is the ISP:

    {SMTP:*;10, SMTP:*.comcast.net;1}

    I looked through TechNet like this link and didn't find much about the costs.  Any ideas to get the messages to automatically go out the other connector without disabling the one with the retry queue?  Thanks.


    Saturday, April 7, 2012 3:42 AM

Answers

  • The automatic failover between connector is not applicable instead you can configure the load balancer between the transport that make to split the queue to multiple servers. Since once message is delivered to the queue it cannot be moved automatically to other send connector.

    Find the url for detail..

    http://technet.microsoft.com/en-US/library/aa998825(v=exchg.140).aspx#Queue

    http://technet.microsoft.com/en-US/library/bb232161(v=exchg.140).aspx


    Exchange Queries

    • Marked as answer by Steve Lynch 2 Saturday, April 7, 2012 3:40 PM
    Saturday, April 7, 2012 4:09 AM
  • I don't think you are going to achieve what you want to achieve with Exchange.

    When it comes to costs, that is used to route email when there are two equal paths.

    The first thing that is used is closest match - so if you have a connector with address space of * and one of .com and send to example.com, then the .com wins.

    Your only option for routing email is going to be to have both smart hosts on the same connector. They will then be load balanced. Once the routing engine has decided where the email is going, it will not change unless you do something to stop it - such as disable the connector.

    Ultimately though the best long term solution will be to get in to a position where you do not need to use a smart host at all, or use a smart host without limits.

    Simon.


    Simon Butler, Exchange MVP
    Blog | Exchange Resources | In the UK? Hire Me.

    • Marked as answer by Steve Lynch 2 Saturday, April 7, 2012 3:40 PM
    Saturday, April 7, 2012 11:57 AM

All replies

  • The automatic failover between connector is not applicable instead you can configure the load balancer between the transport that make to split the queue to multiple servers. Since once message is delivered to the queue it cannot be moved automatically to other send connector.

    Find the url for detail..

    http://technet.microsoft.com/en-US/library/aa998825(v=exchg.140).aspx#Queue

    http://technet.microsoft.com/en-US/library/bb232161(v=exchg.140).aspx


    Exchange Queries

    • Marked as answer by Steve Lynch 2 Saturday, April 7, 2012 3:40 PM
    Saturday, April 7, 2012 4:09 AM
  • I don't think you are going to achieve what you want to achieve with Exchange.

    When it comes to costs, that is used to route email when there are two equal paths.

    The first thing that is used is closest match - so if you have a connector with address space of * and one of .com and send to example.com, then the .com wins.

    Your only option for routing email is going to be to have both smart hosts on the same connector. They will then be load balanced. Once the routing engine has decided where the email is going, it will not change unless you do something to stop it - such as disable the connector.

    Ultimately though the best long term solution will be to get in to a position where you do not need to use a smart host at all, or use a smart host without limits.

    Simon.


    Simon Butler, Exchange MVP
    Blog | Exchange Resources | In the UK? Hire Me.

    • Marked as answer by Steve Lynch 2 Saturday, April 7, 2012 3:40 PM
    Saturday, April 7, 2012 11:57 AM
  • Good info, thank you both for replying.

    I don't think we can add both smart hosts to the same connector, because the authentication for each one is different and the EMC says "all smart hosts must accept the same user name and password".

    We use DynDNS.com as a smart host because in the past we had problems with some servers scoring our messages as junk because of the ISP's address pool, I don't know if that is still a problem or not. They've been really reliable so I guess we could pay for a higher limit and that would eliminate most of our problems unless the smart host was unavailable.

    Next we need to figure out why we didn't get an alert from OpsMgr when the queue was in the retry state.  Thanks for your help.

    Saturday, April 7, 2012 3:39 PM