We first copy the etc/openstack_deploy folder out of the downloaded repository to /etc:

cd /opt/openstack-ansible
cp -R /etc/openstack_deploy /etc

We now have to tell Ansible which servers will do which OpenStack function, by editing the /etc/openstack_deploy/openstack_user_config.yml file, as shown here:

cp /etc/openstack_deploy/openstack_user_variables.yml.example /etc/openstack_deploy_openstack_user_variables.yml
vi /etc/openstack_deploy/openstack_user_variables.yml

The first section, cidr_networks, describes the subnets used by OpenStack in this installation. Here we describe the container network (each of the OpenStack services are run inside a container, and this has its own network so that each service can communicate with each other). We describe the tunnel network (when a user creates a tenant network in this installation of OpenStack, this will create a segregated VXLAN network over this physical network). Finally, we describe the storage network subnet. Edit this file so that it looks like the following:

  cidr_networks:
  container: 172.29.236.0/24
  tunnel: 172.29.240.0/24
  storage: 172.29.244.0/24

Continue editing the file to include any IP addresses that are already used by existing physical hosts in the environment where OpenStack will be deployed (and ensuring that you've included any reserved IP addresses for physical growth too). Include the addresses we have already configured leading up to this section. Single IP addresses or ranges (start and end placed either side of a ',') can be placed here. Edit this section to look like the following, adjust as per your environment and any reserved IPs:

used_ips:
  - "172.29.236.20"
  - "172.29.240.20"
  - "172.29.244.20"
  - "172.29.236.101,172.29.236.117"
  - "172.29.240.101,172.29.240.117"
  - "172.29.244.101,172.29.244.117"
  - "172.29.248.101,172.29.248.117"

The global_overrides section describes the bridges and other specific details of the interfaces used environment—particularly pertaining to how the container network attaches to the physical network interfaces. For the example architecture used in this book, the following output can be used. In most cases, the content in this section doesn't need to be edited apart from the load balancer information at the start, so edit to suit:

global_overrides:
  internal_lb_vip_address: 172.29.236.117
  external_lb_vip_address: 192.168.100.117
  lb_name: haproxy
  tunnel_bridge: "br-vxlan"
  management_bridge: "br-mgmt"
  provider_networks:
    - network:
        group_binds:
          - all_containers
          - hosts
        type: "raw"
        container_bridge: "br-mgmt"
        container_interface: "eth1"
        container_type: "veth"
        ip_from_q: "management"
        is_container_address: true
        is_ssh_address: true
    - network:
        group_binds:
          - neutron_linuxbridge_agent
        container_bridge: "br-vxlan"
        container_type: "veth"
        container_interface: "eth10"
        ip_from_q: "tunnel"
        type: "vxlan"
        range: "1:1000"
        net_name: "vxlan"
    - network:
        group_binds:
          - neutron_linuxbridge_agent
        container_bridge: "br-vlan"
        container_type: "veth"
        container_interface: "eth11"
        type: "vlan"
        range: "1:1"
        net_name: "vlan"

The remaining section of this file describes which server each service runs from. Most of the sections repeat, differing only in the name of the service. This is fine as the intention here is to tell OpenStack-Ansible which server (we give it a name so that Ansible can refer to it by name, and reference the IP associated with it) runs the Nova API, RabbitMQ, or the Glance service, for example. As these particular example services run on our controller nodes, and in a production setting there are at least three controllers, you can quickly see why this information repeats. Other sections refer specifically to other services, such as OpenStack compute. For brevity, a couple of sections are shown here, but continue editing the file to match your networking:

# Shared infrastructure parts
shared-infra_hosts:
  controller-01:
    ip: 172.29.236.110
  controller-02:
    ip: 172.29.236.111
  controller-03:
    ip: 172.29.236.112
# Compute Hosts
compute_hosts:
  compute-01:
    ip: 172.29.236.113
  compute-02:
    ip: 172.29.236.114

Save and exit the file. We will now need to generate some random passphrases for the various services that run in OpenStack. In OpenStack, each service—such as Nova, Glance, and Neutron (which are described through the book)—themselves have to authenticate with Keystone, and be authorized to act as a service. To do so, their own user accounts need to have passphrases generated. Carry out the following command to generate the required passphrases, which would be later used when the OpenStack playbooks are executed:

cd /opt/openstack-ansible/scripts
python pw-token-gen.py --file /etc/openstack_deploy/user_secrets.yml

Finally, there is another file that allows you to fine-tune the parameters of the OpenStack services, such as which backing store Glance (the OpenStack Image service) will be using, as well as configure proxy services ahead of the installation. This file is called /etc/openstack_deploy/user_variables.yml. Let's view and edit this file:

cd /etc/openstack_deploy
vi user_variables.yml

In a typical, highly available deployment—one in which we have three controller nodes—we need to configure Glance to use a shared storage service so that each of the three controllers have the same view of a filesystem, and therefore the images used to spin up instances. A number of shared storage backend systems that Glance can use range from NFS to Swift. We can even allow a private cloud environment to connect out over a public network and connect to a public service like Rackspace Cloud Files. If you have Swift available, add the following lines to user_variables.yml to configure Glance to use Swift:

glance_swift_store_auth_version: 3
glance_default_store: swift
glance_swift_store_auth_address: http://172.29.236.117:5000/v3
glance_swift_store_container: glance_images
glance_swift_store_endpoint_type: internalURL
glance_swift_store_key: '{{ glance_service_password }}'
glance_swift_store_region: RegionOne
glance_swift_store_user: 'service:glance'

View the other commented out details in the file to see if they need editing to suit your environment, then save and exit. You are now ready to start the installation of OpenStack!