1. Even my bookmarks are on a RAID with off-site backup
2. Yeah I keep meaning to get around to that
3. What’s disaster recovery?

Most people coming to freelance/self employment will be coming from an employed background. I.e. They’ve had a company surrounding them which provides more then a few creature comforts, one of which was most probably an IT department which looked after things such as network file stores, Internet access and backups.

However now you are on your own.

If you fall into camp 1 in the list above, then pat yourself on the back and go take a look at something else for a bit. If you don’t fall into camp 1, then read on, as you need to get seriously worried about your recovery plans.

One thing that all IT managers/Sysadmins etc will do before anything else is ensure that they have adequate backups. With a decent backup and disaster recovery plan, you can get yourself out of pretty much any IT hole with the minimum fuss and stress. Ever noticed how when something on the network goes critically wrong your sysadmin is completely laid back and un-stressed (sometimes almost annoyingly so)?

It’s because he/she has that safety net behind them.

Think about this scenario for a bit:

You’ve got work booked in that has to be delivered today, but your main work computer has [been stolen/died/had milkshake dropped on it/got a horrible virus]:

• How do you get back to being able to work for clients as quickly as possible?
• What steps do you need to take to get back to the status-quo?

What I do

Now – having been working in the IT industry for a long time and having been responsible for company IT departments for some of that time, I’m naturally quite obsessive about backups. This is what I currently do – hopefully this can give you a few ideas about how to secure your disaster recovery plans, and therefore secure your turnover.

Geographical separation

Quite simply don’t put all your eggs in one basket. Want to host client websites for preview etc? Don’t do it at the end of your ADSL line. Make sure that all key resources you use are split out over physically different systems. I currently have three resources that I use: My home computer, web server and Amazon S3 storage.

Version control

Store anything important or even mildly important in version control. I use SVN, but you can use whatever you like. All code, writing, notes, configs get stored in SVN. That way I can rollback easily in event of an error. If I am rebuilding my server as it’s failed, then it’s one checkout command to get back all my configs for apache etc.

The SVN server is offsite (not in my house), so is not at the whim of a break in/fire/flood etc. If I do work locally, once the data is checked in I can stop worrying about the security of it on my local machine.

Backups

Backup all this data that you store. As everything of importance that I use is stored in SVN, I just worry about backing up SVN on a regular basis rather then backing up 50 distinct locations. All my backups are stored in multiple locations and on separate physical devices/media.

The main backups are the SVN system and associated files, these upload themselves securely to Amazon S3 daily and, in addition, copy themselves onto an NFS backup mount. So I have each file in 3 secure locations: SVN server, Amazon S3, NFS mount.

Ad hoc backups (images, other binaries that need to be stored once and not version controlled), are stored in S3 as well. I dump these in a local folder on my machine, and this pushes up to S3 daily via a cron job.

I use the s3sync ruby script to manage all these transfers.

Once the data is on S3, I don’t really need to worry too much about it being lost:

Amazon S3 gives any developer access to the same highly scalable, reliable, fast, inexpensive data storage infrastructure that Amazon uses to run its own global network of web sites. The service was designed for 99.99% availability, and carries a service level agreement providing service credits if a customer’s availability falls below 99.9%.

However I do also grab down from S3 on a regular basis and backup the files to CD, just to ensure I have an offline copy of data in an emergency.

Local Machine

I, like any other tech worker, am very fussy about my machine. The mouse sensitivity has to be just so, the desktop environment has to be set up in a specific way and of course the applications I need to use have to be at my finger tips.

In order to get this back as quickly as possible, with minimum fuss, I’ve used mondo rescue to image my machine to DVD(s). It’s linux specific, and your mileage may vary for Windows/Mac partitions. Have a hunt on Google for other options for Windows/Mac – Symantec Ghost suite is a windows app I’ve used in the past.

Remember, all important docs go into SVN, so all I really need to back up here is the system – not personal files, these are already covered.

If my machine dies, and I can’t recover manually, I just pop the mondo rescue DVD into my drive, boot, and it sorts out the rest.

Summary

These are the main systems I use to ensure that I can recover from most problems. Beyond this I use various forms of encryption for data security, but again this is down to user choice and the type of data you store. I’m also about to get a laptop so I am not so reliant on my home machine and Internet connection.

One key thing to remember: when you have to use these systems to recover, try to make a note of what didn’t work too well in the recovery process. I image my local machine because the last time I did a rebuild it took too long.

If you can make a change now that will make your life easier in the event of a problem occurring, then do it.

Please ask any questions about specifics in the comments, and I’ll collate and respond in another post where appropriate.