Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

The scenario is:

I got 2 identical 2U rack mount servers with Operating System Windows Server 2008 R2 Enterprise. They are :

  1. in same domain.
  2. both installed Exchange 2010 in Hyper-V
  3. both are domain controller

If I want to enable failover clustering for both Exchange 2010, ensure zero downtime. What are the steps? Thanks for advice.

share|improve this question

2 Answers 2

up vote 2 down vote accepted

If I want to enable failover clustering for both Exchange 2010, ensure zero downtime. What are the steps?

RTFM.

First, failover cluster DOES NOT GIVE ZERO DOWNTIME. Bad news, isn't it - happens because you did not read the documentation. Failover cluster will start exchange / a VM on another server if the first fails. It takes some time (seconds) to realize the original system is down, plus the time it takes for the system to start (larger for a VM), so while the time is small, IT IS NOT ZERO DOWNTIME.

Second, Exchange can do that a lot better WITHOUT a failover cluster. Exchange has it's own cluster mechanisms that can work with multiple data copies... which means one less thing that can g o wrong (corrupt files in a node crash), so it is (a) superior to failover cluster and (b) faster in switching.

So, I sgugest getting your requirements straight, and then using Exchange integrated mechanisms to ensure uptime.

share|improve this answer
    
Thanks for the suggestion! –  Raptor Aug 16 '10 at 5:23

If I understand you correctly, you have a pretty weird setup. A single Enterprise VM with Exchange and DC? Why not install separate VMs? You're licensed to do that (An enterprise license allows you to install 4 VMs provided your host machine is doing nothing more than HyperV and backup). Here's the setup I would recommend to you. It has the advantage of being highly available and shared nothing (meaning no single point of failure provided your network is redundant. Using Hyper V failover requires shared storage which becomes a single point of failure. A Hyper V cluster can only be as reliable as the SAN that holds the VHDs).

  • Install Windows Server 2008 Enterprise Server Core on each server
  • Install Hyper V role on each server
  • Setup a VM on each machine as a DC (unless you already have other DCs. A preferred setup would have at least 1 physical DC). Make sure you disable time syncing on these DCs.
  • Setup a VM on each machine as Exchange servers.
  • Setup a Database Availability Group with these 2 Exchange servers
  • Enable database copies of each database on each server
  • Setup a CAS Array and assign it to each database
  • Get a hardware load balancer and create a DNS record pointing the CAS array name to the hardware load balancer

I don't think there is such a thing as zero downtime. There is such a thing as high availability. This will give you high availability and if done right can feel to your users as if there is no downtime.

share|improve this answer
    
You can get around the shared storage requirement by doing the CAS array and having 1 node on each server, then use the DAG across the two servers. The hardware load balancer isn't totally necessary when using NLB on the hosts, just watch for all the broadcast traffic on the vlan. –  Ryaner Apr 30 '11 at 8:43
    
@Ryaner - The solution I recommended was a solution without shared storage. I was saying Hyper V clustering would require shared storage. Also, the NLB role is not supported on the mailbox servers in a DAG so a hardware load balancer would be required or he would have to purchase additional licenses for Exchange in order to spin up a couple of additional VMs with the mailbox role separated. –  Jason Berg May 17 '11 at 17:58

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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