Tiny Puppet

Essential Application Management
with Puppet

Puppet Forge Version

View the Project on GitHub example42/puppet-tp


TP playground



Tiny Data

Tiny Puppet data

The data used by Tiny Puppet to manage different applications on different operating systems is defined in the separated tinydata module.

Here for each supported application there is a directory inside the data/ dir which contains:

A sample hiera.yaml is like this:

   - "%{title}/osfamily/%{osfamily}"
   - "%{title}/default"
   - default

so the lookup is done, if $title == 'mariadb' and $::osfamily == 'RedHat' in these files:


The last file contains general defaults for every application.

Note that even if we use a file called hiera.yaml to configure the lookup hierarchy for each application, Tiny Puppet DOES NOT currently use Hiera for any of its lookups, it used a custom function called tp_lookup. The behavior is similar even if Hiera has other use cases and is much more complete, for example we can’t interpolate variables in tinydata.

Using custom data

We have two options at disposal to customize the data used to manage an application:

Both the arguments are available in all the tp functions.

Using a custom data module

The tinydata module is the default module where data is looked for, but we can provide a custom module for our own application settings:

tp::install { 'apache':
  data_module => 'my_data',

This implies that we need to have a directory like my_data/data/apache where we have an hiera.yaml where our hierarchy is described and the relevant yaml files where data for our apache is defined under the apache::settings key.

We can use a different data module for different tp defines.

Providing custom settings

If we don’t need to provide a complete separated module, we can override the default tinydata settings for a given application using the settings_hash parameter, which expects an hash having key names like the settings keys in the tinydata files.

tp::install { 'apache':
  settings_hash => {
    package_name     => 'my_apache',
    config_file_path => '/etc/my_apache/httpd.conf',

Whenever we use custom settings or a custom data module we will probably need to use the same settings for each define of a given application. Here follows an example where we use Hiera to store data both for settings and for options:

$nginx_settings = hiera_hash('nginx::settings')
$nginx_options = hiera_hash('nginx::options')

tp::install { 'nginx':
  settings_hash => $nginx_settings,
tp::conf { 'nginx':
  settings_hash => $nginx_settings,
  epp           => 'site/nginx/nginx.conf.epp',
  options_hash  => $nginx_options,

Note the different between the parameters:

Update policy

Our commitment is to keep Tiny Data as updated and correct as possible, also if this involves breaking backwards compatibility on existing setups.

Whenever new references to new versions of applications or operating systems (for example in additional repos url) are available, they will be updated.

If existing data for some Operating Systems is incorrect we will update it without caring about backwards incompatibilities, we won’t even follow SemVer rules for tinydata.

The driving principle is to have the correct data for each version of each supported operating system and application.

We recommend to make a local fork of this module and update it from the upstream version only after relevant testing.

Bug reporting or pull requests on GitHub are always welcomed.

For more info on cross OS compatibility testing and status, check the this page.