Deployment with controllers in non-default nodegroup cause problems with floating IPs

[Maksym Strukov]

Given multirack environment on 8.0-429:

Rack01 (default):
 * slave01 = compute
 * slave02 = cinder

Rack02 (custom):
 * slave03..05 = controllers

Instances will not be able to reach internet and will not be reachable from public/private networks IF floating range configured as a part of `public` network of rack with no controllers. So in this case correct to configure float range as part of public net of rack2.

There few issues with this:
1. Issue are not described in our documentation: https://docs.mirantis.com/openstack/fuel/fuel-master/operations.html#configuring-multiple-cluster-networks
2. Fuel are not validate this case BEFORE deployment (both on UI or backend)

[Aleksandr Didenko (adidenko)]

I suppose the best way to fix it is to add validation to pre-deployment nailgun checks, it should check that provided floating IP range matches public network of node group with controller(s). If there's no controllers in environment, then no additional validation is needed. Assigned to fuel-python.

[Bug Checker Bot (bug-checker)]

(This check performed automatically)
Please, make sure that bug description contains the following sections filled in with the appropriate data related to the bug you are describing:

actual result

expected result

steps to reproduce

For more detailed information on the contents of each of the listed sections see https://wiki.openstack.org/wiki/Fuel/How_to_contribute#Here_is_how_you_file_a_bug

[Aleksandr Didenko (adidenko)]

Actually, we should not hardcode such limitations to 'controller' roles because it will make it impossible to detach network nodes from controllers. So the best way is to describe this in the documentation in the section provided by Maksym. Assigned to docs team.