5 Steps To Make Sure Your Server Backups Work When You Need Them

April 28th, 2009 by Rich H.

I don’t care if you’re selling bulk widgets or providing financial information on emerging markets, if you rely on data and servers for any function of your business, backups should be the single most important item on your IT checklist.

An all-time favorite question of mine to ask business owners or IT managers when they are shopping for managed hosting, following “do you have backups”, is “how do you know your backups work”? The resounding silence and dumbfounded looks this question summons are almost a guarantee.

If you’re sporting that same look right now, here are 5 simple steps to follow that can help ensure when disaster strikes, you can strike back with a successful data recovery.

1. Perform a test recovery.

This sounds so obvious and simple, but hardly anyone actually does it. This is as simple as it sounds. Attempt to recover you data to an alternate location (don’t overwrite your production data), and test the dataset.

2. Backup to a separate physical storage from where your production data is stored.

Again, this sounds like a no-brainer, but people often overlook the fact they while they may have multiple partitions on their server that appear as different disks, it can still be the same physical drive. Ensure that your backups are written to storage that is separate and physically independent drive from your production data.

3. Plan your backup schedules to match how your data set changes.

This one isn’t quite as obvious, but you should ask yourself what makes up the majority of your data, and how often it changes. If you have a 10GB database and half the data changes on a daily basis, weekly backups aren’t going to be sufficient.

4. Consider whether or not the data you’re backing up will remain consistent during the entire backup process.

Large dynamic files (like databases) can often change thousands of times during the typical length of time it takes to complete a backup. Consider things like snapshots or integrating maintenance plans that generate static copies of large data files ideal for sweeping up as part of a full system backup.

5. Monitor your backups for completion, and re-run when failed.

Again, obvious and simple sounding, but commonly overlooked, and easily pushed to the bottom of a to-do list. I often talk to IT managers who only realized their backups had been failing for months when they logged in to perform a restore. Just because backups work when configured doesn’t mean they’ll work tomorrow. Monitor, monitor, monitor, and re-run when failed.