Pro Puppet (32 page)

In this code example, Puppet will manage the group before the user if the
apache
account is present. However, if the
apache
account is absent, then the user is managed before the group to prevent the operating system from complaining that a group cannot be removed when a user exists with the same
gid
number.

The complete list of syntax arrows are ->, <-, ~> and <~. The tilde arrows add notifications to the relationship just like the subscribe and notify parameters.

The apache group is before the apache user.

The apache user requires the apache group.

The httpd.conf file notifies the httpd service.

The httpd service subscribes to the httpd.conf file.

Additional information about the new relationship-chaining syntax in Puppet 2.6.0 is available online at:
http://docs.puppetlabs.com/guides/language_tutorial.html
.

In the next section, we expand on the concept of virtual resources and make resources available across nodes and configuration catalogs. Resources available for collection across nodes are called exported resources, though it's important to think of them in terms of the virtual resources feature they are designed to resemble.

Now that you're ready to look at exported resources and stored configuration using the groundwork we've introduced with virtual resources, let's start with a database server.

The first step in using exported resources is to install and create the database your stored configuration will use. You can use a variety of database back-ends to store your configuration, including:

• MySQL
  
• PostgreSQL
  
• SQlite3, and
  
• Oracle
  

To allow Puppet to use these different database back ends, Puppet uses the Ruby Active Record object relational mapper (see the Ruby Active Record sidebar). Many people start with the SQLite3 database as a stored configuration back end because it's fast and easy to set up. Unfortunately, it relies on direct file access to write transactions, and this makes it difficult to scale for larger configurations. As a result, we recommend you use a more fully-featured database server. In this chapter, we demonstrate how to use MySQL as our stored configuration database server.

RUBY ACTIVE RECORD

Your database server needs to be installed on a host that is accessible through the network by your Puppet master or Puppet masters. You can install the database server locally on your Puppet master, but we don't recommend this for performance and scalability reasons.

In the following sections we show you how to install the MySQL server on Enterprise Linux- and Debian/Ubuntu-based systems.

MySQL server packages are available from the vendor-provided media on most Enterprise Linux-based systems without the need to enable third-party repositories. Either the yum package manager or Puppet may be used to install MySQL. Unfortunately, the MySQL Ruby library package,
mysql-ruby
, is not available from the vendor package repositories and should be obtained from the Enhanced Packages for Enterprise Linux third party repository.

To install MySQL on Red Hat Enterprise Linux using Puppet, add this line of code:

You also need to ensure taht the Ruby MySQL bindings are present on each Puppet master system:

With the MySQL server RPM packages and Ruby client libraries installed, the next step is to use RubyGems to install the Rails framework.

The first step to configure stored configurations is to install and configure a SQL server. On Debian and Ubuntu systems, this task is easily accomplished by installing the
mysql-server
package:

In addition to the MySQL server packages, the client libraries allowing Ruby programs to connect to a MySQL server need to be installed. On Debian and Ubuntu, these client libraries are contained in the
libmysql-ruby1.8
and
libmysql-ruby
packages.

Once the MySQL server packages and Ruby client libraries are present on the system, you can move on to installing the Ruby on Rails framework.

Exported resources and stored configurations in Puppet take advantage of the Ruby on Rails framework to model and store Puppet resources in a relational database supported by the Active Record library. Installing the Rails framework is straightforward if you are working with a recent version of Ruby and the
rubygems
package.

In this section, we will install Ruby on Rails using the
gem
system command, which is well supported on Enterprise Linux- and Debian-based systems. Indeed, any system with the
gem
command will support this installation process.

First, install Rails for Puppet versions 0.25.x, 2.6.x and later, as you can see in
Listing 6-5
.

There is a problem with Puppet and ActiveRecord versions prior to version 2.3.5, so you need to update the ActiveRecord library to at least this version:

Once Rails and ActiveRecord have been installed, you can verify that the proper versions are present using the
gem list
command.

Notice that
activerecord
and
activesupport
are both available at version 2.3.5. With these libraries installed, you're ready to proceed with the Puppet settings to enable stored configurations.

In the previous sections you installed Ruby on Rails, ActiveRecord, and the MySQL Ruby libraries for the platform the Puppet master is executing on. You're now ready to configure the Puppet master to connect to the database and store configuration information. This configuration is done in the
puppet.conf
file located in the configuration directory,
/etc/puppet
by default.

Before configuring the Puppet master we need to make sure a database has been created for use with Puppet. Any database name will suffice; in this example, the operator uses the default name of “puppet” accompanied by a MySQL account named “puppet” with a password of “teppup.”

First, connect to the MySQL command line interface:

Once connected, create a new database named “puppet”:

Finally, create a MySQL account named “puppet” to access this new database. Notice the password is set to “teppup.” The username and password should be changed to something more secure and reflected in
puppet.conf
.

With the database and account created in MySQL, you're ready to configure
/etc/puppet/puppet.conf
. The lines in
Listing 6-6
need to be inserted in the
[master]
section of the configuration file.

If you chose to change the name of the database, the account, or the account password, please make sure to reflect those changes in the
puppet.conf
settings.

The database tables will not be created until the Puppet master compiles a catalog. We can easily test the configuration of Stored Configs using a standalone Puppet master and agent. After the agent runs, we can expect the tables and configuration information to be visible in the
mysql
console.

This command starts the standalone Puppet master with the new Stored Configuration settings on an alternate port number, 8141, using the
masterport
option. Next, we connect a single Puppet agent to this server in order to trigger the table creation in the “puppet” database: