Puppet – An Introduction

Most common issue while building and maintaining large infrastructure has always been wastage of time. Amount of redundant work performed by each member within team is significant. The idea of automatically configuring and deploying infrastructures has evolved out of a wider need to address this particular problem.

Puppet and Chef are few among the many configuration management packages available. They offer a framework for describing your application/server configuration in a text-based format. Instead of manually installing IIS on each of your web servers, you can instead write a configuration file which says “all web servers must have IIS installed”.

What Is Puppet ?
Puppet is Ruby -based configuration management software, and it can run in either client-server or stand-alone mode. It can be used to manage configuration on UNIX (including OS X), Linux, and Microsoft Windows platforms. It is designed to interact with your hosts in continuous fashion,Unlike other provisioning tools that build your hosts and leave them on their own.

You define a “Desired State” for every node (agents) on puppet master. If agent node doesn’t resemble desired state, in puppet terms “drift” has occurred. Actual decision on how your machine is suppose to look is done by the master, whereas agents only provides data about itself and then responsible for actually applying those decisions. By default each agent will contact master every 30 min, which can be customized. The way this entire process work can be summed with this workflow.

1. Each nodes sends its current information (current state) in the form of facts.
2. Puppet master will use these facts and compile a catalog about desired state of that agent, and send it back to agent.
3. Agent will enforce the configuration as specified in catalog, and send the report back to master to indicate the success/failure.
4. Puppet Master will generate the detailed report which can be feed to any third party tool for monitoring.

Credits -Uthkarsh pandey