Managing Infrastructure with Puppet

Introducing Puppet

Puppet helps you organize and execute configuration plans on servers. This is enabled through a resource abstraction layer that allows you to address the different configurable components of your system as generic objects. In the Puppet view, a server is a collection of resource objects that have a set of particular attributes that describe how that object looks.

It is your job to build a catalog of resource declarations that will tell Puppet how those resources should look when properly configured. When Puppet implements a catalog, it compares the existing resources on the server to the ones that you have defined in your descriptions. It then decides on a set of changes that need to occur to bring the catalog state into agreement with your descriptions. The execution is idempotent, meaning that only the changes needed to bring the state into agreement with the description will be made. The entire catalog can be run over and over again without causing deviation from the described state.

These resource descriptions are made in a Domain Specific Language implemented in Ruby. This means that the syntax is often similar to Ruby, but you cannot simply write Ruby code in a Puppet manifest and have it executed. In fact, the language is declarative, rather than imperative like Ruby. With Puppet, you say how you want things to look, as opposed to describing what should be done to make them look that way. It’s Puppet’s job to know how to make that description reality.

Putting ...