Ah, backups! Everywhere I turn, some system is needing, wanting or performing a backup. Some systems take them once an hour. Other systems are needier — they want a backup every 15 minutes. Others require less — once a day will do them just fine. But no matter how you slice it, my life is consumed with backups.
To be honest, why wouldn’t it be? Our data and intellectual property is the most valuable asset of any individual or organization. Our data is the tangible result of our continuing efforts to learn, grow and prosper. It’s the embodiment of everyone we know and love. Our data contains a lifetime of memories.
A commonly underestimated asset, the value of our data is not truly known until it is lost. Yes, it’s probably worth investing our time and our money to ensure that our data is safe and secure.
So what makes a good backup program? I’ve written about them before, but the 3-2-1s of backups are worth repeating.
3 — Different copies of backup data (not including the original)
2 — Different media types (e.g. disk and tape)
1 — Offsite location
While these rules may seem overkill to some, I attest that this is the minimum level of desired protection. The entire purpose of a backup is to provide protection in the unlikely event of a failure or attack.
Hold on… did I say “unlikely?” That just seems to give the wrong impression. In reality, something will happen. You just don’t know what. Or when. It is best to be prepared.
A few misconceptions exist regarding backups and the different techniques used to perform backups. A good backup system needs a solid architecture. A poor design or the misuse of different backup techniques can lead to a world of hurt.
For example, let’s talk about replication for a minute. Replication is simply the automatic copying of data from one storage system to another. Two copies of the data. That’s twice as good, right?
Not so fast. What happens if your data become corrupted, for example, by something like cryptolocker? The replication system will dutifully copy that corrupt data to secondary storage. Now, you have two copies of completely useless data.
Remember — replication is not a backup.
Folks also tend to abuse the use of snapshots. Snapshots are a feature of storage systems that create system checkpoints in case a quick recovery is needed. For example, you might want to take a snapshot before performing an upgrade just in case something goes wrong.
Some folks get snapshot happy. Snapshots are intended to be temporary checkpoints never lasting more than a day or so. Using snapshots to maintain six months of rollbacks is going eat up your storage. And it’s probably going to lead to data corruption issues.
Finally, we have to talk about tapes. Many believe that current technology has left tape behind. But don’t scoff at the almighty tape. The price per terabyte is still significantly less than disk. More importantly, while disk arrays must remain always online, tapes can be taken offline and transported to another location. When it comes to protecting data, air gaps provide an awesome defense.
So if you ever find yourself frustrated that your IT life is consumed with backups, don’t fret. It’s a significant and extremely important part of our profession. No, it’s not as sexy as software development, virtualization or any of the other buzz-worthy specialties. But if you’re responsible for somebody’s IT, it’s the one system that can never fail.