Huchen is a tool to discover the upstream dependencies of your Opscode Chef cookbooks. Upstream dependencies may include:
- OS packages downloaded by your OS package manager
- Language specific packages installed by your recipes (Rubygems, PyPI package)
- Other resources downloaded by software you are installing
Infrastructure as code promises you can reproducibly create entire environments from source code.
Awesome... except:
- Your converge fails because you depend on OS package repositories being available and not breaking. You probably set up a local repo mirror.
- Then your converge fails because you aren't vendoring and RubyGems is having a bad day.
- Or it fails because you're using RVM head.
- Or it fails because the Chef recipe you are using goes around your package manager to install software.
- Or it fails because it starts a daemon which wants to retrieve a XML DTD on startup.
Adding huchen::default to your run_list doesn't fix any of this. In fact it's yet another dependency.
But it is a convenient way to see during development what stuff your build actually depends on.
Huchen is made up of three parts:
- A cookbook to capture network traffic that you include at the start of your run_list.
- A Go binary that uses the pcap library to write network traffic to JSON.
- A Chef handler that processes the JSON and outputs the list of upstream dependencies.
See the huchen-example repository for an example of using Huchen with Vagrant.