Monday, September 24, 2012

Backup Essentials Part Four

We have now finally arrived to the conclusion of the Backup Essentials series of posts. Part four may be the most important one to consider, as it is the factor that will have the single largest impact on your customers. For our conclusion we'll be covering testing and expectations management, and while these things don't sound as technically exciting as the previous parts they are critical to your business. In covering this topic we'll hit four key areas: Bench marking, availability, retention compliance, and configuration auditing.

Bench mark testing is critical in gathering data for you to use to set reasonable expectations for your customers. Periodic testing of exactly how long your backup solution takes to conduct operations is imperative if you wish to be able to speak credibly to a client about your capabilities. I am always pushing clients at my work to conduct test restores from random client machines at least a couple of times a month in order to keep track of restore time frames. Taking that data and coming up with an average of sorts will allow you to say with some confidence the amount of downtime a client could expect in a disaster. If your clients are attempting to form a service level agreement (SLA) it's especially prudent to have this data available to you.

Just as important is to check the integrity of your data. We gain hark back to conducting test restores, but in this case we have to take a step further than just testing I/O speeds. For example if you are backing up a SQL database the step has to be taken to attempt to do a RESTORE from the dump file (on a dev/test environment). The ideal test gets as close to a real world disaster scenario as possible, and even better if you train your associates to document practices and problems along the way. The data gathered from these tests can form the bedrock of a larger business continuity strategy as you figure out what resources need to go where, and the best placement for your experts. I generally suggest conducting some form of disaster test once every month or two, and at the minimum once every ninety days.

Let us also not forget about retention policies. It is important to double check your backup solution to make sure that any storage retention policies that you have set up are being enforced properly. The last thing you need is to either have your space taken up needlessly (driving up the cost of the solution and decreasing performance), or not have a set of data due to a retention policy that is too tight and unchecked.

Lastly we have configuration auditing. It's useful from time to time to go back through the rules that define what data is being backed up in order to make sure they are still valid. Over time companies move files around, migrate databases and email stores, and well ultimately things are always changing. It would be rather embarrassing to find that one was backing up an empty directory because the backup parameters haven't been checked in months. This is something that should be checked at a weekly or biweekly rate and requires close cooperation with leaders in other parts of the business in order to ensure accuracy.

The conclusion you should draw from this is that testing, auditing and compliance are terribly important. Conducting these properly will allow you to set reasonable expectations with your customers, give you confidence in your technical solutions, and will give your customer reason to have confidence in you.