Pro Puppet (38 page)

Next, you need to configure an Apache virtual host for our Dashboard implementation, which will include enabling the required Passenger module. The Puppet Dashboard provides an example of this virtual host, which you can see in
Listing 7-2
. We will put the file in our Apache configuration directory,
for example
/etc/httpd/conf.d
or
/etc/apache2/conf.d
, on Red Hat and Debian/Ubuntu flavored hosts, respectively.

The virtual host file first loads the mod_passenger module, in this case customized for a Debian or Ubuntu environment. If we'd installed Passenger via a gem those lines might look more like this:

   LoadModule passenger_module /usr/lib/ruby/gems/1.8/gems
/passenger-2.2.9/ext/apache2/mod_passenger.so
   PassengerRoot /usr/lib/ruby/gems/1.8/gems/passenger-2.2.9
   PassengerRuby /usr/bin/ruby


The next options control Passenger-specific options. You can read about each in more detail at
http://www.modrails.com/documentation/Users%20guide%20Apache.html#_configuring_phusion_passenger
.

Last is a very simple virtual host definition that specifies the location of the Dashboard application. The important options to note are
DocumentRoot
and
Directory
. For Passenger to serve out a Ruby on Rails application, these need to be set to the
public
directory underneath the root of the Rails application, in our case
/usr/share/puppet-dashboard/public
.

We can now reload Apache and browse to the URL and see the Dashboard, for example on Red Hat:

You can see the home page of the Dashboard in
Figure 7-1
.

Figure 7-1.
The Dashboard home page

With Puppet Dashboard installed we can now use it, starting with looking a how to integrate our Puppet masters and agents into the Dashboard.

DASHBOARD SECURITY

Now that you've installed the Dashboard, you can integrate Puppet and start to make use of some of its capabilities. We're going to demonstrate:

• Importing Puppet reports
  
• Live aggregation of Puppet reports
  
• Displaying reports
  
• Using the Dashboard for external node classification
  

Finally, we'll look at logging, database and data management, including backing up your Dashboard data.

But the first capability we're going to use is the Dashboard's ability to display Puppet reports and reporting statistics. We can do this in two ways: by importing existing reports using a Rake task, usually scheduled via a cron job, or by configuring Puppet to send its reports directly to the Dashboard.

Let's first see how we can incorporate existing reports from Puppet. The Dashboard comes with a Rake task to perform this action. Change into the root directory,
/usr/share/puppet-dashboard
, of the Dashboard application. Then, run the
reports:import
task.

By default, the task will look in
/var/puppet/lib/reports
for any reports to be imported. If your reports aren't located here then you can specify a location on the command line:

You can run this command multiple times, for example via a cron job. Any reports that have already been imported will be skipped. You should see lists of nodes appear in your Dashboard as reports for each are added. You can see some initial nodes in
Figure 7-2
.

Figure 7-2.
Initial nodes in the Dashboard

In addition to manual report importation, you can also configure Puppet to send your reports to the Dashboard. There are two methods to do this, depending on what version of Puppet you have running. For both methods you need to make changes on both the Puppet master and clients.

For versions 2.6.0 and later, you need to enable reporting on our clients (if you haven't already) by setting the
report
option in the
[agent]
section of the
puppet.conf
file, like so:

On the master you need to enable a new type of report called
http
, which sends report data over an HTTP connection, and specify a location to send our HTTP report,
reporturl
. You don't have to specify a URL, in which case it will default to the local host on port 3000. To do this you need to update the
[master]
section of the Puppet's master's
puppet.conf
file.