Orchestrate with Ansible a complex deployment driven by Puppet OpenStack modules.
Prepare your environment by editing deployment.yaml file. An example is provided with openstack-example.yaml. This is a simple example that will deploy and validate OpenStack Keystone.
---
- hosts: controllers
become: yes
roles:
- common
- mysql
- keystone
- puppet_run
vars:
puppet_role: controller
Soon, you'll be able to deploy complex deployments such as this example:
---
- hosts: controllers
become: yes
roles:
- common
- mysql
- keystone
- nova-api
- nova-scheduler
(...)
- puppet_run
vars:
puppet_role: controller
- hosts: computes
become: yes
roles:
- common
- nova-compute
(...)
- puppet_run
vars:
puppet_role: compute
You'll also need to build your inventory by creating "inventory" file. An example is given in inventory_example.
[controllers]
controller.openstack.org ansible_host=controller.openstack.org
[computes]
compute1.openstack.org ansible_host=compute1.openstack.org
compute2.openstack.org ansible_host=compute2.openstack.org
The architecture of this module makes easy to compose roles. Any deployment can be different, by just editing deployment.yaml and the inventory.
Run the deployment and validation:
./run_tests.sh
- Ansible & dependencies will be installed where Ansible is run.
- Puppet will be installed, configured and tested on remote nodes.
- Each node will have correct Hieradata to deploy the services we want.
- Puppet will be run remotely, in masterless.
- After the deployment, services will be validated.
A demo of a controller deployment + validation: https://asciinema.org/a/eecud0zny7gd3pny1ht27cc6w
- only work on centos7, fedora and rhel.