I’m not going to insult you and ask if you have a disaster recovery plan in place. When StorageCraft Channel Sales Manager Shawn Massey ran the data center of a large solutions provider, he had a great one in place, which included regular testing of file folders, databases, and Exchange-related data, as well as his backup images, and logging those test results.
Then the data center burned down—and Massey discovered too late that his backup images wouldn’t restore properly on his replacement environment.
His mistake? He failed to test his backup images on a “dissimilar system,” as he puts it. His backups worked flawlessly on his original system, but that system didn’t show the patch level of the hypervisor host and how that patch level would affect his ability to restore that disk image on a new machine.
It’s one of those things that seem so obvious once you’ve heard it that it’s embarrassing to have someone else point it out. We know that backup images used for disaster recovery need to be fast, flexible, and reliable, whether you’re doing near-line backups, offsite backups, or both. But what do you make sure you can get everything back up and running the way it was before without experiencing a communal meltdown? Massey has put forth a straightforward methodology:
1. Start from scratch.
You need to test your backup images in unrelated equipment, whether it’s different hardware or a testing environment, such as Oracle’s free VirtualBox. “You can’t simulate this type of strategy in Hyper-V, VMware, Xen Server, or whatever your production environment is on because you have no way of being able to restore to something dissimilar unless you build it yourself to do that,” says Massey.
Massey adds that VirtualBox is a great, inexpensive way to test your images, given that it’s a unique testing environment rather than a production environment.
2. Use a tool like StorageCraft VirtualBoot with VirtualBox to boot your backup image.
VirtualBoot, a tool in StorageCraft’s ShadowProtect backup and recovery solution, lets you mount a backup image to VirtualBox. When you use VirtualBoot in conjunction with VirtualBox, you can determine how functional that backup image is.
3. Ask yourself these questions when your test.
Before you test your backup image, Shawn says to make sure everything on that image conforms to its documented state with all its patches. Then during your test, figure out and log answers to the following:
- How long does it take to perform this recovery from start to finish?
- What types of issues (if any) have you encountered during this test restore?
- If the backup isn’t working properly, what steps do you need to take to fix these problems?
“If you’re able to ‘VirtualBoot’ it, you’re going to be able to restore that backup onto any hypervisor or server hardware you restore to,” says Massey. And if you can’t for whatever reason, you can be proactive in repairing those problems. It’s simple enough to integrate this testing into your current disaster recovery plan—and a much better alternative than performing this test in the midst of a catastrophe.