Wednesday, November 27, 2013

Take Backups Seriously

We do many things to support our clients. The most important thing we do is to test backups. Here's why:

All service requests have to be prioritized. For us the priorities are critical, high, medium, and low. Luckily, there are almost no critical issues. "Critical" means the entire company is affected. You're losing money and some or all systems are down company-wide.

The most important "critical" event is a disaster recovery. That means we need to completely rebuild the server - or sometimes all the technology in the office. Disaster recovery happens when a hard drive crashes, the office is flooded, the office is burned down, etc. Disaster recovery means restoring everything because you lost everything!

There's only one way that's possible: You need a perfectly working backup system. That's no exaggeration: If it's less than perfect, then you can't restore everything.

Luckily, there are many ways to achieve a perfectly working backup system. You might use tape, hard drives, cloud backups, "backup and disaster recovery" devices, fail-over systems, and more. In many cases, your perfect backup will include a combination of these.

Your perfect backup has to be designed by a qualified technician. That's not you, unless you're a technician and you are qualified. If you're very good at what you do, then you understand what it means to hire someone who knows what they're doing.

If you're in business long enough you will experience a catastrophic data loss. That's not a variable. It's just a matter of time. The real question is, how long can your company survive when you have no server or none of you systems are working?

You need to take this very seriously.

Here's a quick way to determine how much data you need to backup:

1) What do you want to recover if your office burns down?

2) Back that up.

Another way to look at this is based on the amount of data you are willing to recreate. Here's how it works. Let's say you just completed a full backup and your server crashes fifteen minutes later. You will need to re-create fifteen minutes worth of data (fifteen minutes worth of the history of your company).

If you backup every hour, you will need to recreate an hour's worth of data. If you backup every day, you will need to recreate a day's worth of data. If you backup once a week, you will need to recreate a week's worth of data.

You get the picture.

Everyone says they don't want to lose anything. Ever. That used to be nearly impossible. It was very difficult and very expensive. Today it is very possible and not extremely expensive. But it has to be designed right, implemented properly, and tested regularly.

The testing part is the most important. If you don't restore from you backup then you cannot guarantee that you actually have a working backup. Your technology consultant should be documenting your backup AND performing test restores at least once a month. You should receive a notice that the backup has been properly tested.

Here's a scary statistic for you: Exactly half - 50% - of all companies we take on as new clients have NO working backup. A few of them know this. Most believe that their backup is working, but it isn't.

Unfortunately, some people only discover this when they call us and their hard drive has crashed. They call us and we have to tell them that they have no backup. They we have to start an extremely expensive process for recovering data from a failed hard drive. Don't be that company!

Here's your to-do list:

- Hire a professional

- Test your backup system

- Update your backup system if necessary

- Test your backup monthly

:-)