This module provides a download_file function to download files on the puppet master to then serve them to the puppet agents.

This is (almost) equivalent to having the following, but running on the puppet master rather than the puppet agent:

exec { '/usr/bin/wget -q -O /tmp/test http://server/test':
  creates => '/tmp/test',
}

The (almost) equivalent to the above would be:

file { '/tmp/test':
  ensure => present,
  source => download_file('files', 'server/test',
                'http://server/test'),
}

The puppet master would download the file into the files file server mount at the server/test relative path, unless that file already exists, and return puppet:///files/server/test so the puppet agent can download it from the file server.

The modules mount point is also supported.

Use case

• The puppet master's /etc/puppet is managed as a Git repository.
• I don't want to store large binaries in Git when they can easily be downloaded.
• I don't want puppet agents to download files themselves using wget or curl. I prefer they load files from the puppet master where all files are centralized.

Note this is not about reducing network load (a caching HTTP proxy would do that more easily), it's about not storing large binaries in Git (where they'd be kept for posterity in the Git history).