We focus so much on the fine-grained details of exactly when and where we’re backing up our servers.  Sometimes it helps to take a quick step back and ask what we’re trying to protect ourselves, because it’ll help us do a better job of designing a recovery strategy.

From small problem to big disaster, we need to protect ourselves from:

• Someone deleting a few rows of data
• Someone deleting a database
• A hard drive (or an entire array) failing
• A server failing
• The entire datacenter turning into a smoking crater

In all of these events, we need to bring up the most recent copy of the data, as fast as we can.  But that’s not the only time we need protection, because sometimes we need to step back in time:

• Someone deleted some data, but didn’t tell us right away
• Hardware (particularly storage) started going bad and gradually corrupted more and more data over time
• Financial auditors need to see the data the way it looked a year ago

Some backup vendors sell their solution as everything you need in one single package.  Before you sign on the dotted line, step through each of these scenarios with your vendor and ask, “With your solution, how would I protect myself from this particular problem?”