By Mike Talon
With disaster recovery planning, we often feel bombarded with brand-new and improved technologies that make the process faster and better. Replication, high availability, remote failover, and a host of other technologies allow us to do more than ever. But because of these improvements, we often overlook what we need to do as a baseline.
Organizations have used tape backup systems in one form or another for decades, and there's just no better way to save data for point-in-time copies. I've often discussed the necessity of using tape backups, but I'd like to look at some best practices you can use to do it properly.
Following the correct procedures and establishing off-site storage policies can mean the difference between recovery and total failure. For example, a study a few years ago found that the average distance between backup hardware and tape storage was only three feet—which isn't going to help you if the data center catches on fire.
Almost any backup software supports the three common types of backup jobs: full backups, differential backups, and incremental backups. Full backups back up all of the data in the selected directories and volumes. This is the most complete backup available, but keep in mind that it takes up a lot of tape, probably much more than you're willing to devote to each individual backup job.
Differential backups capture all data that has changed since the last full backup. They generally use an archive bit on each file, which turns on each time the file changes; performing a full backup turns it off. Differential backups don't turn off the archive bit themselves, so each time you run them, it copies all changed files to tape.
Incremental backups capture all data from the last backup, be it full or incremental. These backups turn the archive bit off each time they run, so they copy only the files that change between backup jobs.
How do you combine these backups to best protect your data? Here's a normal routine: Run a full backup at the beginning of each week, and run either a differential or incremental backup each day.
Performing a restore requires the last full backup and either the last differential backup or all incremental backups between the last full backup and the current day. This allows you to completely restore all data up to the last completed backup.
Off-site storage is also necessary to properly protect your systems. Keeping tape systems in the same room as the live servers won't help you if there's a major disaster. Sending tapes to an off-site storage provider—or, if all else fails, just to someone's house—is a good way to ensure that you'll have them when you need them.
It's typical to store the weekly full backup off-site each time you make a new one and to recycle them once per year. So each week, create a new full backup, and send the one from the previous week off-site for storage. This practice lets you restore immediately in minor emergencies, but you risk losing only one week of data if you suffer a major incident.
For larger companies, tape backups can be an integral part of your overall DR strategy; for smaller shops, it may be your entire DR policy. In any case, knowing what to tape and how to store it will help you make sure you don't miss anything when it comes time to restore.
Mike Talon is an IT consultant and freelance journalist who has worked for both traditional businesses and dot-com startups.