Using Ansible

Ansible is a popular open source configuration management and application deployment tool. Ansible provides a set of core modules for interacting with OpenStack. This makes Ansible an ideal tool for providing both OpenStack orchestration and instance configuration, letting you use a single tool to setup the underlying infrastructure and configure instances. As such Ansible can substitute for Heat for OpenStack orchestration and Puppet for instance configuration.

Comprehensive documentation of the Ansible OpenStack modules is available at https://docs.ansible.com/ansible/list_of_cloud_modules.html#openstack

Install Ansible

We have written a script to install the required Ansible and OpenStack libraries within a Python virtual environment. This script is part of the catalystcloud-ansible git repository. Lets clone this repository and run the install script in order to install Ansible.

$ git clone https://github.com/catalyst/catalystcloud-ansible.git && CC_ANSIBLE_DIR="$(pwd)/catalystcloud-ansible" && echo $CC_ANSIBLE_DIR
$ cd catalystcloud-ansible
$ ./install-ansible.sh
Installing stable version of Ansible
...
Ansible installed successfully!

To activate run the following command:

source /home/yourname/src/catalystcloud-ansible/ansible-venv/bin/activate

$ source $CC_ANSIBLE_DIR/ansible-venv/bin/activate
$ ansible --version
ansible 2.1.1.0
  config file = /etc/ansible/ansible.cfg
  configured module search path = Default w/o overrides

Note

Catalyst recommends customers use Ansible >= 2.0 and Shade >= 1.4 with the Catalyst Cloud.

OpenStack credentials

Before we can run the playbooks we need to setup our OpenStack credentials, the easiest way to achieve this is to make use of environment variables. We will make use of the standard variables provided by an OpenStack RC file as described at Source an OpenStack RC file. These variables are read by the Ansible os_auth module to provide Ansible with permissions to access the Catalyst Cloud APIs.

Note

If you do not source an OpenStack RC file, you will need to set a few mandatory authentication attributes in the playbooks. See the vars section of the playbooks for details.

Now we have an Ansible installation that includes the required OpenStack modules and have setup our OpenStack credentials we can proceed to build our first instance. We have split the first instance playbooks into two playbook files, the first playbook create-network.yml creates the required network components and the second playbook launch-instance.yml launches the instance.

Run the create network playbook

Lets take a look at what tasks the create network playbook is going to complete:

$ ansible-playbook --list-tasks create-network.yml

playbook: create-network.yml

 play #1 (localhost): Create a network in the Catalyst Cloud   TAGS: []
   tasks:
     Connect to the Catalyst Cloud TAGS: []
     Create a network  TAGS: []
     Create a subnet   TAGS: []
     Create a router   TAGS: []
     Create a security group   TAGS: []
     Create a security group rule for SSH access   TAGS: []
     Import an SSH keypair TAGS: []

We are going to need to provide the path to a valid SSH key in order for this playbook to work. You can edit create-network.yml and update the ssh_public_key variable, or we can override the variable when we run the playbook as shown below:

$ ansible-playbook --extra-vars "ssh_public_key=$HOME/.ssh/id_rsa.pub" create-network.yml

PLAY [Deploy a cloud instance in OpenStack] ************************************

TASK [setup] *******************************************************************
ok: [localhost]

TASK [Connect to the Catalyst Cloud] *******************************************
ok: [localhost]

TASK [Create a network] ********************************************************
changed: [localhost]

TASK [Create a subnet] *********************************************************
changed: [localhost]

TASK [Create a router] *********************************************************
changed: [localhost]

TASK [Create a security group] *************************************************
changed: [localhost]

TASK [Create a security group rule for SSH access] *****************************
changed: [localhost]

TASK [Import an SSH keypair] ***************************************************
changed: [localhost]

PLAY RECAP *********************************************************************
localhost                  : ok=8    changed=6    unreachable=0    failed=0

Run the launch instance playbook

Now we have a network setup we can run the launch instance playbook:

$ ansible-playbook launch-instance.yml

PLAY [Deploy a cloud instance in OpenStack] ************************************

TASK [setup] *******************************************************************
ok: [localhost]

TASK [Connect to the Catalyst Cloud] *******************************************
ok: [localhost]

TASK [Create a compute instance on the Catalyst Cloud] *************************
changed: [localhost]

TASK [Assign a floating IP] ****************************************************
changed: [localhost]

TASK [Output floating IP] ******************************************************
ok: [localhost] => {
    "floating_ip_info.floating_ip.floating_ip_address": "150.242.41.75"
}

PLAY RECAP *********************************************************************
localhost                  : ok=4    changed=2    unreachable=0    failed=1

We can now connect to our new instance via SSH using the IP address output by the Output floating IP task:

$ ssh ubuntu@150.242.41.75

We can now write playbooks to configure the instance we have created as required.

Resource cleanup with an Ansible playbook

This playbook will remove all of resources created and allocated by running the previous playbooks.

You can download and run this playbook using the following commands:

$ wget -q https://raw.githubusercontent.com/catalyst/catalystcloud-ansible/master/remove-stack.yml
$ ansible-playbook remove-instance.yml --extra-vars "floating_ip=<ip-address>"

Where <ip-address> is the floating-ip that was assigned by the launch instance playbook

Note

This cleanup playbook assumes that you have created all resources using the default names defined in the original playbooks.