Linux Server Hacks, Volume Two

Hack #49. Clone Systems Quickly and Easily

Once you've customized and fine-tuned a sample machine, you can quickly and easily deploy other systems based on its configuration by simply cloning it.

Now that Linux is in widespread use, many businesses that don't want to roll their own Linux systems simply deploy out-of-the-box systems based on supported distributions from sources such as SUSE, Mandriva, Turbo Linux, and Red Hat. Businesses that need a wider array of system or application software than these distributions provide often spend significant effort adding this software to their server and desktop systems, fine-tuning system configuration files, setting up networking, disabling unnecessary services, and setting up their corporate distributed authentication mechanisms. All of this takes a fair amount of time to get "just right"—it also takes time to replicate on multiple systems and can be a pain to recreate if this becomes necessary. You do have backups, don't you?

To speed up deploying multiple essentially identical systems, the classic Unix approach that I used to take in the "bad old days" was to purchase large numbers of disks that were the same size, use the Unix dd utility to clone system disks containing my tricked out systems to new disks, and then deploy the cloned disks in each new system of the specified type. This still works, but the downside of this approach is that the dd utility copies every block on a disk, regardless of whether it's actually in use or not. ...