Monthly Archives: November 2009

Planning and Testing Your Backup Strategy – ComputorEdge Magazine

0
Filed under External Articles
Tagged as , , ,

From the ComputorEdge Issue from 11/13/09, Backup Strategies—Don’t Wait for the Bomb!

I wrote another article:

Planning and Testing Your Backup Strategy

Dealing with backups is important, and having a good plan in place, maintaining that plan and periodically testing that plan will help ensure that your backup is solid.

This IS a Drill

Few computer users have to be convinced of the importance of backing up their data, even if vast numbers of people still admit that they don’t. But having a solid backup isn’t as simple as selecting your files and assuming they’ll be there if you need them. A big part of backing up that is not so obvious is testing your backup. To truly test it, you must have a good plan in place in advance.

When I was in second grade, I really looked forward to fire drills, for a number of reasons. I had a crush on a girl who sat on the far side of the room, and if I was very clever about getting in the line out the door, I ended up standing right next to her (smooth!). Another benefit was that if we students played our cards right, we could delay classes by almost a half hour. But while I was busy furthering my own selfish interests, I was missing the real importance of the drills. The fire drills gave the administration the chance to test the routes, gauge the capacity of the hallways, and measure the speed of the evacuation. For the students, the repetition prepared us to automatically know what to do in the event of a real emergency. The testing and the learning were both based on one thing: a plan. Backing up (and successfully restoring from a backup) has a similar requirement. You must have a good plan in place, and you must test the effectiveness of that plan.

Most people do have at least a general plan in place, although it may not be documented. And for most people, that may be enough. But, depending on the type of data you are backing up, it can get more complicated. Some programs save settings in a different location, others have log files, others create libraries to save data, and these files could be in a particular folder or even a different drive. Part of the importance and value of the plan lies in making sure you are familiar with your data, and then documenting it, so you do not have to piece together this info later, after a problem.

After documenting the settings and file locations of your important programs, and their recovery requirements, you can use this as a checklist when setting up your backup, to make sure you have everything important included. After you have configured your backup job, it is valuable to print out the list of files and folders that are backed up, the details of the backup, and include this as an appendix in your plan. When you have the plan nailed down, I’d suggest storing a hard copy someplace safe, and also e-mailing a copy to yourself at an online e-mail account, where you can save it separate from all your other data.

Once the plan is established, most people will happily forget about their backup, assuming it is running and taking care of all their very important information. But how do you know? The good answer is that you don’t, so now you go about testing your backup. The “drill” will cover several important aspects: It will test your backup method (reliability and integrity), it will test your familiarity with the data, and it will test the applications you use. Lastly, it will reinforce the need to keep the plan current.

Method

Pick a random file, any file. Try to restore in whatever software you use to backup. So many folks never do this until they are in a tough spot and are really trying to retrieve a file. That is already a pretty stressful situation, so being familiar with the process, and confident that it works, can really serve to diffuse some of that tension. In my experience, this is often when someone learns some very important limitations or even failings of their backup procedure. For example, one client learned that their inexpensive online backup solution requires 24 hours to generate a restore file and make it available via download. There is nothing wrong with that, if you can wait, but if you were up against a deadline and only discovered that while trying to restore, you can imagine it would be pretty disappointing.

So, knowing that this online provider follows this procedure is important, and there won’t be expectations of something else during crunch time. Another customer thought they were doing solid backups, but in fact, they were never swapping their tapes. So day one got a great backup of the files. And day two overwrote that file backup with the database. And day three zapped that and stored the e-mail. The client didn’t know it, but on any day she would only have been able to restore whatever had been backed up the previous evening. A test would expose this problem, and give you time to resolve it before you need to rely on those backups.

Familiarity

This time, think of a single, important file to restore. Now go to the backup software and try to specifically restore that file. This is good practice; you are learning the software and testing your familiarity with where your files are stored. Again, with many clients, this is when they sheepishly realize that they’ve never actually set the backup to grab a particular folder, instead assuming it was part of the backup job. Better now than later!

Applications

Now, try restoring a file that you may not use directly (a picture or document is pretty normal and would be stored where you might expect it to be, but Outlook uses database files, for example, or your music library might have some hidden files you don’t directly engage). Restore the files, and follow through by opening or using them. Too many backup applications put so much effort on making the backup side super easy, with one click to grab all of Outlook. But not all of them have a one-click restore.

Many will restore the files required, but it is up to you to configure, for example, an Outlook profile so you can even open the file that was restored. A real-life data-loss emergency is not when you want to learn that restoring your e-mail database also requires log files, or that your financial software requires a license file in order to open the restored company file, or that you are not sure where to put the restored files. Becoming more familiar with these scenarios in advance will save you a lot of stress, and possibly even some money (that you might spend on a consultant or recovery specialist), in the event of catastrophic data loss.

Maintenance

The last step is maintenance. Your backups need recurring attention. Contrary to what is sometimes advertised, “set it and forget it” is not a good idea when it comes to disaster recovery or computer security in general. Set a task in your calendar to remind you to do monthly backup drills. Included in the drill should be a task to update your plan. When you visit this plan each month, make a note of new folders or drives you’ve added, or new applications you’ve installed, and add them to the backup if that is appropriate.

Just like any skill, the more frequently you practice, the better you will be at it. Dealing with backups is very important the few times you’ll ever need to do it. Having a good plan in place, maintaining that plan and periodically testing that plan will help ensure that your backup is solid and that you have the skills to make good use of it if the need arises.


Tim McGuire founded The Backup Plan, Inc. in 2005 and writes on a variety of blogs as tmcguire47.