For a storage or Windows administrator, the only thing worse than losing data is discovering that your backups don't work in the first place. Here are four tips that will help you ensure that your backups are indeed in working order.
1. Test regularly. This is the bottom line. Can you recover the data? Never mind that the backup program "verified" the backup. If you don't actually restore data, you'll never be certain your backup is working. How often you should test depends in part on the nature of your operation. The general recommendation is to run your test protocol once a quarter.
2. Test realistically. You can't wipe the server just to do a complete test restore, but there's a lot you can do to make sure the entire recovery process works. A backup test protocol should be designed to make sure all aspects of your recovery operation works with minimal disruption to your regular operations. Often you can automate large parts of the test, but don't skimp on something just because it has to be done manually.
Remember: You need to test the entire backup chain. It's not enough to know that the tape is readable. Can you read it on the system it will be restored on? This is particularly important when performing a restore over a network, SAN or whatever, because a misconfigured network component can keep a restore from working just as much (although perhaps not as permanently) as a corrupted tape.
Don't forget that your people and the documentation are part of the backup chain as well. Backup tests are an opportunity to train your staff in backup and recovery procedures as well as to find holes in your documentation.
A worthwhile procedure is to select a tape or a volume at random, find it, and then try to back up from the data on it. (Note the part about finding the tape; a misfiled tape is as unavailable as an unreadable one.)
3. Test all your backups regularly. Backup is more than what goes into the vault. It includes all the tiers of your backup system, including the incremental backups, any disk-based backups, recovery disks and anything else you need to get up and running again. All these things need to be tested to make sure they're working properly – and that they're working together.
4. Test compliance-related recovery regularly. Technically, record recovery for compliance isn't a part of backup. But it's an increasingly important function fulfilled by backup systems. Test to make sure you can recover files, logs, chains of emails and anything else your company is likely to need for compliance.
None of this has to be traumatic. In fact, if it is traumatic, then you've already discovered you have problems. A procedure that encompasses all it needs to for your enterprise shouldn't take an excessive amount of time or effort. Once you have one, do it regularly. If nothing else, you'll sleep better at night.
About the author: Rick Cook specializes in writing about issues related to storage and storage management.
More information on this topic:
