Current environment is one Exchange 2003 SP2 server and no front-end server and we are currently planning the transition to Exchange 2010 SP2. The corporate building, Site A, is known to have power outages in severe weather. We now have a DR site, Site B, and the company wants email to failover in the case of a power outage. We have 2 DCs in Site A and one in Site B, and in AD, they are in different sites with different subnets. We get educational pricing so we have plenty of Exchange and Server 2008 licenses and a robust Vmware environment connected to a fibre-based SAN. I have about 1200 users with a 150 MB mailbox size limit, about 20 messages sent/received per day, at 86KB message size. I've calculated 266GB database size as per http://technet.microsoft.com/en-us/library/ee832796(v=exchg.140).aspx. I'm getting the sizing together, so firstly, would I need to utilize database copies, and if so, do I need more than two copies? I don't want to over-engineer what is a really small environment but I do want to provide HA and failover as required. I planned to have at Site A, one CAS/HUB server, one MB, and one Edge on the DMZ. At Site B, there will be one CAS/HUB server and one MB. I thought I could configure a DAG between the MB servers and I would have an active/passive environment and I could manually failover when needed and the CAS/HUBs would be load balanced. As I have done more reading, I don't think this is adequate to have HA at Site A and to have failover to Site B. I think I need another MB role at Site A in order to configure DAC for the fail over, but I'm a little confused.

You should be very careful trying to do HA between sites. In your case, it just won't work. You are suggesting one server in the primary site and one in the secondary site. Presumably the file share witness will be in the primary site. If the primary site's power fails, the secondary site will not have a quorum and therefore will not take control. Sometimes people suggest putting the file share witness in a third site, but that can cause other problems, including false failovers. And even if it does fail over, it will still require you to manually make changes unless you have elegant monitoring and action systems, like System Center Operations Manager backed up by custom scripts. I believe you are on the right track to try to provide HA in one site, and consider the second site a DR site. You can switch over manually to the DR site without a lot of effort, but I wouldn't plan on having your Exchange services fail over automatically.Ed Crowley MVP "There are seldom good technological solutions to behavioral problems."

You should be very careful trying to do HA between sites. In your case, it just won't work. You are suggesting one server in the primary site and one in the secondary site. Presumably the file share witness will be in the primary site. If the primary site's power fails, the secondary site will not have a quorum and therefore will not take control. Sometimes people suggest putting the file share witness in a third site, but that can cause other problems, including false failovers. And even if it does fail over, it will still require you to manually make changes unless you have elegant monitoring and action systems, like System Center Operations Manager backed up by custom scripts. I believe you are on the right track to try to provide HA in one site, and consider the second site a DR site. You can switch over manually to the DR site without a lot of effort, but I wouldn't plan on having your Exchange services fail over automatically.Ed Crowley MVP "There are seldom good technological solutions to behavioral problems."

Also not that with one CAS server in a site you cannot load balance it. You cannot create a CAS array across multiple AD sites.

Thank you for the feedback. I can failover manually. Now I am not sure how to configure a DAG that will replicate to my DR. How can I confiure the FSW to have a proper DAG? Do I need more MB servers? Can you point me to a resource?