Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm tasked with providing fail over of our Exchange 2003 Server. Currently we're using Microsoft backup to backup the mail store nightly over the network to another machine. We currently use double take to mirror a database server and it seems like a reasonable solution for the exchange server but it is quite expensive and I'd need to dedicate a standby server for it. I've heard double take can sync to a VM, but I'm not on support with them anymore and I can't find much info on the subject.

My goals are (of course) to minimize cost and downtime. I'd love to use my double take backup server as a standby for multiple machines (it's got the space), but I don't know if it's possible. The maximum downtime I can afford is less then a day (my servers are on hardware support contracts, but hardware (drives, raid controllers, motherboards etc) still takes time to ship), so I'm not looking for immediate fail over, just the ability to smoothly transition to a backup machine in an emergency.

What have people used? What works well?

share|improve this question
up vote 2 down vote accepted

You need to plan this rather than shoot from the hip, but this is a case where virtualisation can work as a kind of "poor man's clustering" from the point of view of protection against hardware issues.

Assuming you back up the virtual hard disks themselves rather than the files inside them, which takes some thought but is perfectly "do-able", you can take these to any server running your virtual server software (whether VMWare, Microsoft, Xen or whatever) and just create a new virtual server on that machine using the "current" virtual hard disks and you're away - you don't need all the fancy stuff like vmotion and its equivalents.

share|improve this answer
Nice idea, especially for situations that don't require "quick" failover. Reimporting a VM can take several hours still, so you'll need that kind of time frame - and you'll need to be prepared to lose all the data since the last backup of course. – Daniel Lawson Jun 4 '09 at 22:12
In order to make it work really well in Virtualization environment, you will need a SAN to store all your VMs. – kentchen Jun 4 '09 at 22:39
@Kentchen, yes that's ideal but I was suggesting a more simple arrangement as the full on HyperV Server/Xen Enterprise/VMware ESX with a SAN and autofailover seems like overkill for the OP's needs and budget, from the vibe I get out of their post. You can simply back up the virtual hard disks daily on a simpler product and restore them elsewhere if you need to with something like VMware server. As dlawson says, you'll lose backup since your last virtual hard disk backup that way, but then that's true with "real" backups too. – RobM Jun 4 '09 at 22:43
@Robert, yes, I understand. If SAN isn't an option, which isn't really expensive nowadays anyway, make sure store VMs in the external detached storage so it can be easily attached to the new server when the old one fails. And that storage of course has to be RAID'd. – kentchen Jun 4 '09 at 23:05
I've been pushing for some sort of SAN for a while now. We're at a critical mass of servers where if we want to grow any bigger and still have our backs covered we're going to need to break into "enterprisey" hardware. Maybe figuring out a way to help my boss explain the costs to his boss should be another question... – reconbot Jun 5 '09 at 13:32

If you are planing for the hardware failure, the simplest solution is virtualizing your Exchange server.

If further more you also need to be prepared for system self, you then just make a good backup or snapshot of the VM.

share|improve this answer

I've used CA XOSoft Replication and HA for this exact scenario. From my experience it performs pretty well. It'll handle both replication of specified files, and HA migration of either an IP address or modifying a DNS name, and then starting/stopping services on the appropriate nodes.

It's somewhat cheaper than Doubletake, and last I looked it was cheaper to license for running in a VM than on physical hardware, which lends itself well to a DR VM situation.

share|improve this answer

Your Answer


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.