Sep<-- Oct 2017 -->Nov
S M T W T F S
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31 1 2 3 4

Disclaimer: The entries you find in these pages are based on my individual opinions and thoughts. Some of the entries may be just plain wrong, and others harmful. Should you choose to act on, or try, anything you find on this site, you assume any and all risks associated with your actions. So there.



 


More on Backups

September 4, 2006

A well thought-out contingency plan (for disaster, emergency, fire, theft, employee sabotage) can be summed up in one word: backup. If you have a good backup system in place, one which includes rotating your backup media off-site, you are pretty much covered for disaster. This includes hurricanes, fire, and theft. Here are a few of my thoughts on this:

1) Software. I prefer Retrospect. I know there are plenty of detractors of this software, but I have been using it for over 15 years (yep) and it has never failed me. Oh, I have had plenty of backups that failed-- but it has never been the software that caused it. Failures are caused by inattention, bad media (which, if you are attentive, you catch), and procrastination.

2) Automation. One of the reasons I love Retrospect is that, once set up, it does everything but switch out the backup media. My experience is that fewer than 1 in 20 clients, if left to determine when the backups should be run, will actually do it. This is why I design backup scripts to backup every day. For Mac OS X users, I have Retrospect backup "User folder and Prefs." With Windows, it's not quite as easy, but documents, desktop and the email files are easy enough to define for the backup. Plus I use "Backup Server" mode. This is a feature of Retrospect that allows it to backup 24/7, use any available backup media, and track what files need to be backed up to which backup set. Leave it to the user, and you'll never have as good of a backup.

3) Redundancy. A single backup is not, in and of itself, sufficient. Firstly, what happens if you have a fire? But, more to the point, many times I have to go to a backup not to restore a failed hard drive, but because a user screwed something up. With only one backup, often enough, the "good" file is gone, replaced by the screw-up (even with incremental backup, when you backup to external hard drives, you have to start over periodically). Having 2 or even 3 drives in some kind of regular rotation gives you the grace to choose from many different versions of files when you restore.

4) Off-site storage by a principal of the company. If you want to be safe from disaster, then you simply must have one of your backups stored off site. Don't make the mistake of giving it to some employee to do, because s/he won't have the investment in safety that you, a principal, will have. Yes, I have had a client get burned because an employee he had to terminate just so happened to be the keeper of backups. Due to stupidity, he gave her a chance to "copy" some of her files before making her leave. She trashed their entire project folder, then went home and cut the ribbon cables on the backup drive. We were blown away-- no one expected this person to respond this way. But people are seldom what they appear to be.

5) Don't be too ambitious. Yep, you heard it. I am paranoid when it comes to backup, plus a cynic. Paranoid in that I think about disaster. I've seen theft, sabotage, failure, lightening.... pretty much everything. But I'm also a student of human nature, and understand business dynamics, thus I'm a cynic. People are the least attentive when they are under pressure, and what causes pressure in businesses is being busy. When you are the busiest, you generate the most work, you push your equipment the most, and you have the least amount of time. If backups require too much of your precious time, you won't do them. Although it may cost more to setup in terms of hardware, I always recommend a dedicated backup system-- an entire computer devoted to this. If a user's workstation is the backup computer, backup will get blown off when the company is the busiest; and if you backup too much (systems and apps for instance), the backups require too much storage, and therefore more maintenance, and eventually become a nuisance. So balance is achieved by not being too ambitious.

6) Backup to hard disk. Tape and CD/DVD media are a nuisance. In the case of a server with over 1Tb (terabyte) of storage, you have no choice. But most small companies won't face that kind of storage requirement. External hard drives, from 250Gb up to 1Tb are ideal. They are portable, and easily swappable. Buy two, and swap them out once a week (not daily, see #5). Assuming you keep one off-site (see #4), even in the worst disaster, you can't lose more than 1 week of data-- acceptable by most standards. If you can't stand to loose that much, you can do other things (daily swapping, etc.), but then, you're not the target for this entry.

So, the contingency plan for your computer systems becomes this: when the storm comes, leave town with the backup drive in your possession. That's it. You may think that you need to plan/prepare to protect computers, etc., but in reality, this is folly. If you have plenty of time, and won't risk life or limb, go get the other backup drive too.

In the event of a truly devastating event, such as a Katrina (or a Hugo, more dear to us in the Lowcountry), there isn't much you can do to protect the hardware. And the idea of taking hardware with you is foolishness. Also, the idea that you should make employees do the same is even worse. Abandon the hardware (except the backup drive(s), of course). Take your family, your photo albums and heirlooms; let your employees do the same. Since you haven't burdened your employees with anything but their own worries, you'll have more loyal employees, because they know you have a plan to put it all back together-- a plan that doesn't encumber them with your problems.

The idea is that computers and software are covered by insurance (if not, you should look in to your insurance), and are easily replaced (sometimes preferably replaced, as in older hardware). If your backup is done right, and you lose everything at the office, you can be back up and running in the time it takes to get new hardware (usually days), and restore the data (another day or two). The assumption here is that if you are facing total devastation, this timeframe is more than sufficient because of all the other stuff you are facing (new location, no power, etc.).

Now, this is not to say that you should not take steps to secure your computers and building. Of course you should. You don't want minor damage needlessly taking out your systems. Shut things down, unplug them, move them to safer locations in your building. Do this under the advisement of your computer support provider (you don't want to be dismantling RAID drives, or other things that might literally be damaged by your acts of protection). Bear in mind that I am not addressing non-computer related files and paperwork-- you're on your own for those. (But it's all digital now-a-days, right? A real argument for digital document storage, in my opinion.)

I won't go in to all of the potential ins-and-outs here, because every company will be slightly different. But I'll leave you with this advice: think about the reality of total loss. Consider having no building, no computers, and no power for a few weeks, and people scattered around while their personal affairs are dealt with (because it is just as likely that their homes will be similarly affected). How does a small business cope with this? If you have your data, and loyal employees, all else can be put back together.