Audit your backups
February 21, 2017
Moving can feel like a disaster, or at least enough of one to make you reevaluate your company’s disaster recovery plan. Retrospect, Inc. recently relocated a number of servers offsite. We took this opportunity to audit our own internal backup and disaster recovery plan, and today’s blog post is going to go over how you can do that as well.
Plan for disaster
Planning for a disaster is the first step in being able to cope with one. A fire, a flood, or other calamity can be overcome, if you make a plan and follow through with it.
Knowing what you know
When we moved our servers, we overlooked that one of our Macs had four different Ethernet cables connected to it. Reconnecting this ended up wasting a lot of time. So while this is a backup blog, think beyond backups to include mundane details like these.
What machines do you or your customers have? Include both on and off site equipment. Include parts of your infrastructure that you may not think of, for example, modems, routers or switches that make up your network. This is also a good time to think about older machines - that XP box that runs that one piece of software that everyone forgot about. Will you be able to just buy new machines and restore or is there going to be some upgrade pain involved?
Update (or create) a map of your network, making sure to account for all of these boxes. Your company may have a very straightforward network, but Retrospect has multiple subnets and a wide variety of equipment (DHCP servers, web servers, email, file servers) for testing which makes restoring that more complex.
If your backups are on old tapes, are they compatible with the drives you can get in the near future? (We’ve heard from customers where that was a real issue.)
Full system backups vs "User data"
We always recommend full system backups - it is an easy way to make sure you aren’t overlooking something. That said, we understand not everyone has the time or the resources to back every file up on every machine. There are several alternatives however to the old "everyone should store their important stuff on the server". One alternative is to use Retrospect’s selectors to only back up "User Files and Settings". You can make this faster by defining favorite folders (on the Mac) or subvolumes (on Windows) of the Users folder and just back that up.
If you do take that approach, consider doing full system backups once a month or some other long interval.
And if you do only back up files from specific locations, make sure all your users know about this limitation.
History, Redundancy and Offsite
By default, Retrospect will keep adding changes to a backup set forever. To keep from running out of space, Retrospect has options for scheduling recycles of backup sets as well as grooming. The danger is that viruses, specifically ransomware, can make recent backups worthless. Having a longer history of backups can be a lifesaver so make sure you have a good balance - I would feel very uneasy if we did not have at least a month of backups and Retrospect’s backups go back years.
Beyond multiple points in time, make sure that you have multiple backup sets of your data. If the drive your backups are on has a failure, you no longer have a backup. Similarly, make sure you have at least one of those redundant sets offsite, to protect against fire, flood or other site-wide disaster. This is another area where Retrospect can help you out - Retrospect can do scheduled backups once, and then you can schedule backup set transfers on a regular basis - to different types of media for added protection. Retrospect supports local and remote network attached storage, multiple cloud storage providers, as well as hundreds of devices.
Test restoring your data
Retrospect has built in verification—turn it off at your peril!
Beyond verification however, you should periodically make sure that you have run through the restore process. Many folk may balk at this as it usually requires having redundant hardware. Many problems, however, can be identified however just by mentally walking through the process. Be sure to do one of these at least once a year.
Knowing what your environment is, having multiple backups to different locations, and testing your backups is a great start. Finish up by writing all of this down, posting it online, emailing it to yourself and your customers, and make sure all your users know about it. Now you can sleep more soundly knowing you are ready for any disaster—or the next major move.
JG Heithcock is CEO of Retrospect, Inc. and has eighteen years experience in the storage and backup industry.