Description of problem:
The current examples upstream only show an example of an IP pool for the controllers. It shows that the pools are defined under "ControllerIPs" and it shows which networks this role is connected to. I had to guess the other roles and it lead to hard-to-detect failures in the deployment (I didn't guess correctly that the compute IPs go under "NovaComputeIPs).

Here is the example that currently works for me, but it too is still a bit partial because it doesn't include the cinder and swift roles. Please document an example with all roles and their networks:

parameter_defaults:
  ControllerIPs:
    external:
      - 192.168.200.210
      - 192.168.200.211
      - 192.168.200.212
    internal_api:
      - 192.168.100.30
      - 192.168.100.40
      - 192.168.100.50
    storage:
      - 192.168.110.20
      - 192.168.110.30
      - 192.168.110.40
      - 192.168.110.50
      - 192.168.110.60
      - 192.168.110.70
    storage_mgmt:
      - 192.168.120.20
      - 192.168.120.30
      - 192.168.120.40
      - 192.168.120.50
      - 192.168.120.60
      - 192.168.120.70
    tenant:
      - 192.168.150.50
      - 192.168.150.60
      - 192.168.150.70
  NovaComputeIPs:
    internal_api:
      - 192.168.100.60
      - 192.168.100.70
      - 192.168.100.80
    storage:
      - 192.168.110.80
      - 192.168.110.90
      - 192.168.110.100
      - 192.168.110.110
      - 192.168.110.120
      - 192.168.110.130
    tenant:
      - 192.168.150.20
      - 192.168.150.30
      - 192.168.150.40
  CephStorageIPs:
    storage:
      - 192.168.110.115
      - 192.168.110.125
      - 192.168.110.135
    storage_mgmt:
      - 192.168.120.80
      - 192.168.120.90
      - 192.168.120.99