ansible-openstack-inventory

ansible-openstack-inventory

An opinionated dynamic inventory for openstack clouds that integrates environment filtering.

Installation

go get -u github.com/xlucas/ansible-openstack-inventory

Usage

Create a .clouds.hcl file in your home directory. This file follows the HCL format to describe your cloud providers and their regions.

Below is a sample content for OVH Cloud:

provider "ovh" {
  options {
    meta {
      environment     = "ansible_environment" // server metadata used to define environments
      groups          = "ansible_groups"      // server metadata used to define groups
      hostvars_prefix = "ansible_hostvar_"    // server metadata to be added to hostvars
      user            = "image_original_user" // image metadata used to guess ssh user
    }

    fallback_user = "debian" // ssh user to fall back to if image metadata is missing
  }

  identity {
    version   = 2
    endpoint  = "https://auth.cloud.ovh.net" // the identity version is added by the inventory
    username  = "xxxxxxxxxxxx"
    password  = "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
    tenant_id = "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
  }

  regions "eu-west" {
    region "Gravelines" {
      name = "GRA1"
    }

    region "Strasbourg" {
      name = "SBG1"
    }

    region "London" {
      name = "UK1"
    }
  }

  regions "eu-central" {
    region "Frankfurt" {
      name = "DE1"
    }
  }

  regions "eu-east" {
    region "Warsaw" {
      name = "WAW1"
    }
  }
}

Usage with environments

• In the configuration file, make sure you have specified the env key in the meta section. This value represents the key that the dynamic inventory will look for when building the environment host list.
• Make sure your instances are created with the previous metadata key and the environment name as a value. It should be the same name than the folder containing environment-specific configuration in your ansible setup (environments/<env>).
• Make sure the inventory for each environment is invoked with --env=<env>.

Example

Ansible layout:

.
├── ansible.cfg
├── hosts
├── environments
│   ├── dev
│   │   ├── group_vars
│   │   │   └── ...
│   │   └── hosts -> ../../hosts
│   ├── preprod
│   │   ├── group_vars
│   │   │   └── ...
│   │   └── hosts -> ../../hosts
│   └── prod
│       ├── group_vars
│       │   └── ...
│       └── hosts -> ../../hosts
├── roles
│   └── ...
└── playbook.yml

The hosts file at the root of the directory is a simple script that is used to invoke the dynamic inventory with the desired environment. Other inventories are simple symlinks to this script. Its content is:

#!/bin/sh
dir=${0%/*}
env=${dir##*/}

ansible-openstack-inventory --env=$env ${@}

In ansible.cfg, specify the default inventory that should be used when not specified on the command line:

[defaults]
inventory = ./environments/dev

Check that only instances from the development environment are targeted when no inventory has been specified:

ansible -m ping all
4b7ae578-bf37-4179-b324-9fad1b1d78ad | SUCCESS => {
    "changed": false,
    "ping": "pong"
}

Run the same command for a specific environment, for instance preproduction:

ansible -i environments/preprod -m ping all
b7bd2783-0082-4a4e-997d-00472c6dc940 | SUCCESS => {
    "changed": false,
    "ping": "pong"
}

Notice how these hosts are different since they pertain to different environments.

Usage without environments

Simply configure ansible to use it as the default inventory in ansible.cfg and omit the env parameter in the metadata section.

[defaults]
inventory = /path/to/ansible-openstack-inventory

License

This project is distributed under the Apache License, Version 2.0.