We help IT Professionals succeed at work.

DAG design question, and Powershell cmdlet required to perform DR. Exchange 2013 version

Jerry Seinfield
Jerry Seinfield asked
Hello MS team,

I am about to propose the following design to my client. Client has 2 physical sites. One in US, and the other one in Canada.

Site 1


2 MBx servers

DB1, DB2, DB3, DB4 DB5.. DB8, are active databases residing on the 2 MBX servers in site 1

Site 2


2 MBx servers

A passive copy of DB1..DB8 resides on this servers

A single DAG has been created to host all DBs, . All the CAS servers must be placed behind a load balancer

Here is my question:

if we lost 1 MBx on each site, all DBs must be failover automatically to other site, assuming there is WAN connectivity. Here i need to know, what are the powershell cmdlets to make sure all dbs are fully failovered to second site. Any extra task to be performed? Here I assume, that we have quorum to failover all DBs

Second question:

If we lost WAN connectivity, tell me how can I validate all emails will be restored on site 1, and all users will be using site 2. please provide additional steps and powershell cmdlets to failover, and advise if I need to create a second share witness to something has to be done on outlook or dbs to restore email flow on second site. once the WAN access is restored, what are the failback tasks, cmdlets?

Consider also DNS, autodiscover, and mobile devices, any task that must be performed to restore services
Watch Question

You have asked for a tall order, one which I can't hope to fulfill, but let me ask for a few clarifications to help out the next person who's willing to answer...

In your DAG arrangement, are there 4 copies of each mailbox database, one on each of the 4 MBX servers? Each of the mailbox databases may consider a different MBX server to be "primary" when things are running normally. If one MBX server at a site goes down, I'd think you'd want to the other MBX server at the same site to take over the primary role temporarily. In this scenario, you could lose one, two, or three MBX servers and the poor overburdened fourth MBX server would still be there to keep mail available.

By "lost WAN connectivity", are you speaking of just the link between the two sites (which is supposed to be redundant)? I'd expect to consider a scenario where one or the other sites loses all WAN connectivity, cutting it off from both the other site and from users, before a freakish accident that severed just both lines of the site interconnect. The witness server would need to be located outside either site for you to get the full benefit of automatic failover.

Properly set up, this system could weather either the simultaneous loss of any two MBX servers or the complete loss of one of the sites through automatic failover and should resume normal operation automatically as well when normalcy is restored.


On the design there are 10 databases.

Site 1 will retain 10 active databases

Site 2 will have passive copies DBs of the 10 DBs above

One DAG will keep all the databases