According to Don Jones, in his The Definitive Guide to Windows Application and Server Backup 2.0,the first step in any healing process is admitting your mistakes, and these may include your current backup practices. Don suggests six things an IT professional should be willing to reject to provide better backups and surer recoveries:
Point-in-time backups? "Horrible," he intones. "You're always going to lose some data if you have to rely on point-in-time backups, and who wants to lose data?" He suggests that continuous backups deliver far more flexibility, involve a lower loss risk, and, all in all, require less manual effort and overhead. But that's not all. Point-in-time backups also lack granularity, he says. "Even those quarter-hour SQL Server transaction log backups leave too much data at risk and don't let you roll back the server to a precise point in the past, if needed."
Backup windows are horrible too, he continues, pointing out that taking servers offline does nothing to enhance the value of an organization. Sure, from time to time, maintenance is unavoidable; but backups force you to choose what data to back up: "You can only grab what will fit within the window," he explains, and goes on to say that continuous backups don't require a window since snapshots don't require a server to be taken offline, "so you are free to make better decisions without all the time constraints."
Another thing to reject: Duplicated data, which by definition means you're backing up lots of things more than once. Bloating backups, consuming more computer and administrative time, and taking a whole lot more storage space than they, a real problem in a time when data growth is exploding. Things get expensive when you consider the cost of offsite storage), and since there's more data, backups can get longer than your backup window can support. Jones suggests, "Look for solutions that can automatically integrate date deduplication software when making backups."
Incomplete backups. "Backing up anything less than the entire server is horrible," he continues. "Sure, your worst nightmare might be losing a single email from the CEO, but that's far from the only nightmare you need to prepare for. Don't make decisions about what's important—back it all up. With the right solution, that one all-inclusive backup will enable everything from message level restores to bare-metal server recovery—and everything in between, including restoring an entire application to an earlier point in time.
Corrupted backups—also horrible. Backup solutions should tell you when something goes wrong—not wait for you to find out for yourself.
Old-school backup software. Finally, Jones concludes, "Old-school backup solutions make it incredibly difficult to actually test backups." He recommends finding a solution to make it easier to test backups, and suggests considering an ‘app-like' approach that restorations to either a virtual machine or dissimilar hardware, which can make it easier to test restorability.
To learn more about a unified, cross-platform backup, replication & recovery software from AppAssure, click on www.appasure.com/solutions/backup/.
For a free trial download please click here.
