Chapter 3. Adding Puppet Modules to Red Hat Satellite 6

The first step to adding our Puppet module is to create a custom Product.

Creating a Custom Product

Satellite now has a new Product called MyProduct.

The next procedure creates a Puppet repository in our custom Product.

Creating a Custom Puppet Repository

The custom Product now contains a repository to store our Puppet modules.

Now we upload our mymodule module to the newly created repository, which adds it to our custom Product.

You can upload more modules to this repository. For our example, we only need to upload the mymodule module.

Our Puppet module is now a part of your Red Hat Satellite 6 environment. Next we publish the module as part of a content view.

If you aim to remove redundant modules from custom repositories in the future, use the Manage Puppet Modules feature.

Satellite removes the chosen modules from your repository.

As an alternative to manually uploading modules, Red Hat Satellite 6 includes a utility called pulp-puppet-module-builder. This tool checks out repositories containing a set of modules, builds the modules, and publishes them in a structure for Satellite 6 to synchronize. This provides an efficient way to manage module development in Git and include them in the Satellite 6 workflow.

The use of Modulefile is deprecated since Puppet version 3. If you build any Puppet module containing a Modulefile when using Puppet version 3.X, a deprecation warning is printed. Satellite 6.3 will include Puppet version 4 which will ignore the Modulefile data. The data in the Modulefile must be moved to a metadata.json file. You can convert a module to use a metadata.json file as follows:

One common method is to run the utility on the Satellite 6 Server itself and publish to a local directory.

Publishing a Git Repository to a Local Directory

The same procedure applies to publishing modules to an HTTP server.

Publishing Git Repository to a Web Server

In the Satellite 6 web UI, create a new repository with the URL set to the location of your published modules.

Creating a Repository for Puppet Modules from Git

The Puppet modules in the Git repository are now included in your Satellite 6 Server.

The final step to getting our Puppet module ready for consumption is to publish it as part of a content view. You can add this module to an existing view but for our example we will create a new view.

Publishing a Content View

Our content view is now published. As a part of the content view creation, Red Hat Satellite 6 creates a new Puppet environment for use in the provisioning process. This Puppet environment contains our module. You can view this new Puppet environment on the Configure > Environments page.

A Puppet environment is defined as an isolated set of Puppet agent nodes that can be associated with a specific set of Puppet modules. In Satellite 6 context we can think of a Puppet environment as a set of Hosts running the Puppet agent that are associated with a specific set of Puppet modules. For example, nodes that are associated with environment Production only have access to the modules that are in environment Production.

Puppet environments are used to separate Puppet modules from different types of Host. A typical use is to enable changes to a module to be tested in one environment before being pushed to another. A Puppet module can contain facts and functions as well as one or more Puppet Classes that you can assign to Hosts. Modules are part of an environment, Puppet Classes are only part of an environment because they are part of a module.

In Red Hat Satellite, a Puppet environment will be created automatically for a Content View if the CV has a Puppet module. The automatically created Puppet environment name includes the organization label, the life-cycle environment, the Content View name, and the content view ID. For example, KT_Example_Org_Library_RHEL6Server_3.

Creating Puppet Environments within Red Hat Satellite

Navigate to Configure > Environments and click New Puppet Environment. Give the new environment a name and click Submit to save the changes. Note that if the environment does not exist on the Puppet master and you subsequently run an import, Satellite will prompt for the environment to be deleted.

Importing Puppet Environments to Red Hat Satellite

Satellite can detect all the environments and Puppet modules contained on a Puppet master, and import them automatically. To do this, go to Configure > Environments and click on the "Import from" button. The button name will be include the FQDN of the internal or external Capsule. Satellite will scan the Puppet master via the Capsule, and display a list of the detected changes. Select the changes you want to apply and select Update to apply the changes.

Note that the Capsule will only detect environments that contain one or more Puppet Classes, so ensure that at least one Puppet module containing a class has been deployed to the Puppet master.

Assigning a Puppet Environment to a Host

Navigate to Hosts > All Hosts, select a host by name from the Hosts list and then select Edit. On the Host tab you can select a Puppet Environment for the host. Selecting an environment will filter the classes on the Puppet Classes tab to help distinguish the classes in the selected environment.

You can mass-assign an environment to a group of hosts. Select the check boxes of the required hosts in the Hosts list, and then select Change Environment from the Select Action drop-down menu at the top of the page.

Assigning Puppet Environments to Host Groups

When creating a Host Group, the environment field is pre-populated by the environment that was automatically created by the associated Content View, if any.

By default, a user creating a new host and selecting the Host Group will automatically have the environment pre-selected. The user can change the environment of the new host.

Some module classes contain complex parameters, called Smart Class Parameters, that define a value for a key just as a simple parameter does, but allow conditional arguments, validation, and overrides for specific object types. Satellite 6 has the ability to import these parameterized classes and allow modification of such smart parameters. For more information on the types of parameters in Puppet, see Parameters in the Red Hat Satellite Host Configuration Guide.

For example, mymodule contains a parameter for the HTTP port of our web server. This parameter, httpd_port, is set to a default of 8120. However, a situation might occur where we need to use a different port for a provisioned system. Satellite 6 can override the httpd_port parameter during configuration. This provides an easy way to change the HTTP port on our web server.

This procedure requires the mymodule module to be uploaded to a Product and added to a Content View. This is because we need to edit the classes in the resulting Puppet environment. Alternatively, you can download and install another module as described in Chapter 3, Adding Puppet Modules to Red Hat Satellite 6.

We now have an override value for the Smart Class Parameter httpd_port.

Smart Variables are a tool for providing global parameters to the Puppet master for use with Puppet Classes that do not contain Smart Class Parameters. The same Smart Matcher rules are used for both Smart Variables and Smart Class Parameters.

Before Smart Class Parameters were introduced, users who wanted to override a parameter where asked to rewrite their Puppet code to use a global parameter. For example:

class example1 {
  file { '/tmp/foo': content => $global_var }
}

For the above example, $global_var is set in the Smart Variables section of the web UI and the value is associated with the "example1" class. Although it is recommend to precede global variables with :: to restrict Puppet to search the global scope, their absence does not mean a variable is not a global variable.

With the introduction of Smart Class Parameters, the following form could be used:

class example2($var="default") {
  file { '/tmp/foo': content => $var }
}

For the above example, $var is set in the Smart Class Parameters section of the web UI and the value is associated with the "example2" class. If you see a variable defined in the class header, as in the above class example2($var="default"), then you can be sure that $var is a class parameter and you should use the Smart Class Parameter function to override the variable.

As demonstrated, Smart Variables require custom-designed modules using global-namespace parameters, rather than standard modules from the Puppet community, and the result is the same: text placed in '/tmp/foo' in the examples above. Consequently, there is no longer a reason to use Smart Variables except to support legacy modules.

Although Smart Variables are global variables, they are associated with a Puppet class and will only be sent to a host that has that specific Puppet Class assigned in Satellite. You can create a Smart Variable with any name, no validation is done in Satellite, but unless the associated Puppet module applied has a matching variable in its code, the Smart Variable will not be used.

Satellite adds the variable you create in Satellite to the Host YAML file. This file can be viewed in the web UI by navigating to Hosts > All Hosts, selecting the name of the host, and then clicking on the YAML button. Satellite sends the Host YAML file to the external node classifier (ENC), a function included with the Puppet master. When the Puppet master queries the ENC about a host, the ENC returns a YAML document describing the state of the host. This YAML document is based on data taken from Puppet manifests, but is subject to Smart Class Parameter overrides and any Smart Variables.

class anothermodule {
   file { '/tmp/motd':
     ensure  => file,
     content => $::content_for_motd,
  }
}

This example will supply a value for the $::content_for_motd parameter.