Hello!

https://technet.microsoft.com/en-us/library/dd351027%28v=exchg.150%29.aspx

Messages transmitted within a transport high availability boundary

Message routing is optimized in Exchange 2013 so that when the ultimate destination is in a DAG or Active Directory site, multiple hops between the Transport service on Mailbox servers in that DAG or Active Directory site aren't typically required. Once the message is accepted by the Transport service on a Mailbox server in the DAG or Active Directory site that holds the ultimate destination for the message, the next hop for the message is typically the ultimate destination itself. Shadow redundancy's goal of keeping two copies of a message in transit is fulfilled when one shadow copy of the message exists anywhere within the DAG or Active Directory site. Typically, only failover scenarios in a DAG that require the Redirect-Message cmdlet to drain the active queues on a Mailbox server would require multiple hops within the same transport high availability boundary.

I can't understand where exactly should the shadow copies be created, for example, within a DAG with two CAS/MB servers (Exch1 and Exch2). Does the procedure described in the "Messages received from outside a transport high availability boundary" apply for the servers in a DAG?

Thank you in advance,

Michael

Hi,

According to that document, the shadow redundancy requires multiple Exchange 2013 Mailbox servers. So your theoretical environment with two servers(one CAS, one MBX) doesnt be applied by that description.

We suppose one CAS server and three MBX server. Exch1 for CAS, Exch2, Exch3 and Exch4 for MBX. One message sent to user on Exch4. An SMTP server transmits this message to the Transport service on Exch2. The Exch2 should be the primary server. Because the primary server is a member of a DAG, so primary server connects to a different Mailbox server in the same DAG (we suppose its Exch3). Then Exch3 is the shadow server, it holds the shadow copy of the message. And one message copy will be transmitted to Exch3.

But the destination is Exch4, so the next hop should be Exch4. In this situation Exch4 will act as the primary server. Exch4 will connect to another MBX in the DAG (we suppose it Exch2). Because the one shadow copy of the message exists within the DAG, so Exch4 will transmit just the message copy to Exch2.

To sum up, in this situation, two copies of a message will be transmitted to Exch3 and Exch2. Hope my description will helpful for you to understand these words  Shadow redundancy's goal of keeping two copies of a message in transit is fulfilled when one shadow copy of the message exists anywhere within the DAG or Active Directory site.

At last, all the descriptions above are my understanding. If I am wrong in some points, thanks for pointing out.

Best Regards.

Hi Lynn-Li,

Thank you for the reply!

Frankly speaking, I didn't understand anything...

"One message sent to user on Exch4. An SMTP server transmits this message to the Transport service on Exch2." - if a user after connecting to Exch1 wants to send a message to a mailbox located on Exch4 why "An SMTP server" should connect to Exch2 and not to Exch4"?

My test configuration is two CAS AND MB servers (multirole servers) within a DAG.

The main question: does the procedure described in the "Messages received from outside a transport high availability boundary" apply for servers in a DAG or not?

In other words I can't understand how (if) smtp transactions  in "Messages received from outside a transport high availability boundary" differ from ones in  "Messages transmitted within a transport high availability boundary" .

Regards,

Michael