1

At the moment I have a single Exchange 2010 server with all roles (MB, HT & CAS).

I am looking to improve our email recovery time in the event of a disaster. I understand that true HA with site resiliency requires CAS arrays and hardware network balancing. Unfortunately this isn't something my company is willing to pay for. At the moment we just backup the mailbox databases daily with DPM2012 and would need rebuild an Exchange server at our secondary site if we were to lose out primary site (head office). This gives us an email recovery time of about 4-5 hours.

I've been looking at DAGs. Am I able to simply build a second Exchange 2010 server offsite (with MN/HT/CAS roles) and group them together in a single, two node DAG?

If so, in the event of a disaster where by I lost my primary server / site (EX1), would the second server (EX2) automatically mount the mailbox DB copies and it just be a matter of redirecting our MX records to this new server?

Edit: My current Exchange 2010 server in installed on Server 2008 R2 DataCentre

  • What host Operating System is Exchange installed on? – joeqwerty May 10 '16 at 14:38
  • @joeqwerty - Server 2008 R2 DataCentre – OPGailey May 10 '16 at 14:44
1

CAS Array doesn't cost anything and you should set one up now. It is basically a DNS entry. With a load balancer the DNS entry would point there, but it can point to the live Exchange server instead. In the event of a failure you would update the DNS entry to point to the second server. If you are a small site, look at the free load balancer from Kemp - it might be enough for you.

Otherwise, a DAG is exactly what you have said - a second server with Exchange installed on it. It needs to be identical to the first - same OS, patch level etc, same storage layout.

The biggest headache with a retro-fit DAG is the initial seed of the database. If possible the easiest way is to have the second server on the same LAN as the first, do the seed, then move it. If not possible, then new databases, added to the DAG and the mailboxes moved in to the new databases over time works well.

Just to be clear though - you said DC above - I presume, based on your answer to Joe, that you have meant DataCenter edition and not a domain controller. A DAG cannot be on a DC.

Finally, on the subject of MX records - you can have both servers listed for MX records. Exchange copes with that fine. Email can be delivered to any server in the org and Exchange deals with delivery.

  • You mention a 'retro-fit' DAG being a bit of a pain. Is that because of the time it takes to seed the original DBs to the second DAG server? We have 5 databases totaling 600GB. Our head office LAN is 192.168.100.0/24 and our offsite LAN is 192.168.1.0/24. We have a site to site VPN established between the two LANs. Would I have any trouble building a new exchange server on the offsite LAN and seeding our EX1 DBs over the VPN tunnel? – OPGailey May 11 '16 at 9:34
  • Correct - the problem is that seeding time. 600gb over any kind of WAN connection is going to take some time. That is why I like to do the initial seed on the same LAN, then move the servers. The move mailbox method can work as well. If the seeding fails, then you have to start again! – Sembee May 12 '16 at 14:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.