"One of the most overused and abused statements used to scare small and medium-size businesses into buying a disaster recovery plan goes something like this:
"According to SCORE, Counselors to America's Small Businesses, 50% of businesses who suffer a data disaster lasting longer than ten days will go out of business within five years."
These mythical, Armageddon-like statements are usually written by backup and recovery software makers do a Google search on "disaster recovery statistics" and find the statements on other disaster recovery software sites. Following these statements to their source almost always proves futile, and if you can track the statement to its source, as I have, it's common to find the statement was based on a published figure that may be seven to as much as 10 years old.
Still there's a kernel of truth to it all
While these statements perpetuate a myth, there is, as with any myth, a kernel of truth: you have a right to be afraid of you, your customers, and your partners lose access to your application data for an extended period of time. How long depends on an evaluation of each application and how important it is to your internal staff, your customers and the partners who are tied into it. Whether you're looking for Windows Server Recovery, Exchange server recovery, VMware Server recovery or Hyper-V server recovery, here are four steps you can do to get your own DR plan off the ground, and reduce the fear and loathing you may have for the process.
1) Prioritize on recovering first those applications everyone depends on first, like Exchange Server email and SQL applications, along with mission-critical infrastructural components like your domain controller. Move down the list to more static back office functional areas that you can put on a secondary recovery tier, like your Human Resources applications.
2) Concentrate on protecting all these resources from a single management point. If you've recently virtualized some of your physical servers, it's likely you used a different backup and recovery product to protect them, compared to either your physical machines or other virtual platform you may also be running. This can come back to bite you in a disaster recovery situation, where pressure is high and speeds is of the essence. It's always better to try to consolidate all of your disaster recovery under a single solution. After all, a single solution is much easier to master than two or three, and it's cheaper to pay for and easier administer one product.
3) Keep your recovery plan up to date. Using the same example as above, let's say you've gone through a significant virtualization effort in 2011. Are your new machines as well protected as your existing stable of machines? Review at least annually and be safe.
4) Practice, Practice, Practice!
It works for every profession, whether you're a soldier going out on combat, a baseball power hitter taking cuts in the batting cage or an IT administrator encountering a fire or flood in the server room. If you've practiced, you're more likely to be able to recover successfully. In the end, that matters the most, for your company and for your job. Quarterly "events" and annual DR simulations will keep you sharp, and you'll learn in advance what needs to be fixed before you really need to carry out a disaster recovery.
