Configuring Exchange 2010 HT for fault tolerance/HA when using smarthosts
Recently I was trying to find a way to provide high availability across two Exchange sites using two smarthosts. The key issue was finding a way to make the local site's smarthost primary while still having the failover site's smarthost available to each HT server.
First off, we need to define some records. Hopefully you already have your smarthosts defined in DNS and for this article, we will assume they are primary-sh.test.local and failover-sh.test.local. These will point the IP address of the smarthosts in the DMZ.Next, we create the MX records. For the primary site, we want to create two records with the same name, but they should resolve to the two smarthosts with different costs. For example:mx1.test.local should resolve to primary-sh.test.local with a cost of 10
and
mx1.test.local should resolve to failover-sh.test.local with a cost of 20Now in your primary outgoing send connector, you should use mx1.test.local as the smarthost fqdn and allow your primary site's HT server access to the connector. If the primary (lowest cost) smarthost is down, DNS will automatically direct the send connector to use the failover smarthost. It is worth noting that if you set the costs the same, the HT server will use the smarthosts in a round-robin fashion.
and
mx2.test.local should resolve to primary-sh.test.local with a cost of Similarly, when you configure the send connector for the failover HT server to use, you would use mx2.test.local as the smarthost. Again, it will failback to use to primary smarthost if there is an issue with the failover device.Once you apply the send connector settings, check your queues. Everything should be good to go. Be sure to take your primary smarthost offline to test the failover. Testing is key!

