stone-burner

Wrapper over terraform which adds a configuration layer and state splitting between projects and/or environments, among other features.

Requirements

• Terraform v0.11 or higher
• Python 3.6

Install

pip install --upgrade stone-burner

Usage

stone-burner --help
stone-burner <command> --help

*Note: The first time using terraform you will have to run stone-burner install to initialize your projects.

Example

stone-burner plan project_1               # plan all components from project_1
stone-burner plan project_1 -c c1         # plan only c1
stone-burner plan project_1 -c c1 -c c2   # plan only c1 and c2
stone-burner plan project_1 -xc c3 -xc c4 # plan all except c3 and c4

Note: You can plan one project only if the projects it depends on have been applied.

Passing extra parameters to terraform

If you want to send extra parameters to terraform (like for example, the -target option), make sure to use -- to avoid stone-burner trying to parse those options. For example:

stone-burner apply -e production project_1 -c c1 -- -target=some_resource.address

Install provider plugins

If you are using provider plugins in your configuration files, you will need to first install them in order to start working with terraform.

stone-burner installs all the plugins under ~/.stoneburner/plugins in order to save space disk and avoid re-downloading same plugins that are used in your different project/components.

There are 2 ways of installing terraform provider plugins:

• Discover and install from your configuration files:

  stone-burner install -p project_1 -c c1 -c c2

• Install plugins and versions manually:

  stone-burner install template@1.0.0 aws@1.5.0 kubernetes@1.0.1

Configuration

The way projects are configured is via the --config flag (config.yml by default). In this file, you can define projects by combining individual components and variables. For example:

projects:
  project_1:
    - database            # projects/project_1/database + variables/<environment>/project_1/database
    - app                 # projects/project_1/app + variables/<environment>/project_1/app
  project_2:
    - database:
      - database_1        # projects/project_2/database + variables/<environment>/project_2/database_1
      - database_2        # projects/project_2/database + variables/<environment>/project_2/database_2
    - app:
      - app_1             # projects/project_2/app + variables/<environment>/project_2/app_1
      - app_2             # projects/project_2/app + variables/<environment>/project_2/app_2

So, there are 3 different key/values here:

• Top level keys refer to projects.
• List elements under the project can be defined in 2 ways:
  • As a string, a component and variables with the same project and component name will be used.
  • As a dictionary, the key will refer to a generic component or component type. The list items under this key will all use the same component (terraform configuration) using different variables.

Environments

You must define your environment details in your configuration file before you can use your -e or --environment flag. For example:

environments:
  - name: production
    aws_profile: my-production-profile
    aws_region: us-east-1
    states_bucket: terraform-states-production

  - name: staging
    aws_profile: my-staging-profile
    aws_region: ap-northeast-1
    states_bucket: terraform-states-staging
    default: true