Puppet Best Practices

Chapter 8. Node Classification

The term node classification refers to both the tools and process used to supply a role, individual classes, and node parameters used to customize the catalog for a node.

Classifiers take many forms. Large sites typically employ a database or console for node classification. Smaller sites will often use Hiera or the Puppet language to classify nodes. The best solution is always site specific.

In this chapter, we look at a number of classification strategies. The goal of this chapter is not to redocument all of the available classifiers, but to look at the available options and specific considerations and best practices behind each one.

What Data Should the Node Classifier Provide?

As we discussed in Chapter 1, node classification provides data used to customize the model to produce a catalog specific to a node. Let’s take a quick look at the kind of data that is commonly set with a node classifier.

Roles and Profiles

Node classification’s original and most common purpose is to declare classes to be applied on the node. This selection can be direct and explicit or indirect through the assignment of roles and profiles.

Tip

A role doesn’t need to be a class. The role could be used to alter the Hiera hierarchy appropriately for inclusion of classes from a Hiera array.

The recommended approach is to assign a single role class, as discussed in Chapter 7; however, a classifier can also assign individual classes or profiles a-la-carte if necessary. Direct ...