Bug 1825315

Hi Luke,

It seems ansible failed to generate the inventory due to no ctlplane network mapping. 
As per the upstream guide [2], I guess below env parameters should help. Since these env values not included in our downstream guide[1], so we would like to know thought on before to proceed.

~~~
export OVERCLOUD_ROLES="ControllerDeployedServer ComputeDeployedServer"
export ControllerDeployedServer_hosts="192.168.25.1 192.168.25.2 192.168.25.3"
export ComputeDeployedServer_hosts="192.168.25.4"
~~~

Reproduced steps:
----------------
1. Copied the existing deployment script and updated with the config-download parameter to forcefully trigger ansible inventory.

	~~~
	$ diff test-deploy.sh test-deploy-config-download.sh
	18c18,19
	< --overcloud-ssh-key ~/.ssh/id_rsa --log-file overcloud_install.log &> /home/stack/overcloud_install.log
	---
	> --config-download-only --config-download-timeout 300 \
	> --overcloud-ssh-key ~/.ssh/id_rsa --log-file overcloud_install_config.log &> /home/stack/overcloud_install_config.log
	~~~

2. Overcloud failed to configure Ansible inventory.
	~~~
	$ tail -f overcloud_install_config.log
	Waiting for messages on queue 'tripleo' with no timeout.
	2020-04-20 14:06:49.748 372067 WARNING tripleoclient.plugin [  admin] Waiting for messages on queue 'tripleo' with no timeout.                                                                                                               
	2020-04-20 14:07:27.419 372067 ERROR openstack [  admin] Overcloud configuration failed.
	or Controller role on ctlplane network', action_cls='<class 'mistral.actions.action_factory.AnsibleGenerateInventoryAction'>', attributes='{}', params='{'ansible_ssh_user': 'tripleo-admin', 'ansible_python_interpreter': None, 'work_dir':
	'/var/lib/mistral/overcloud', 'plan_name': 'overcloud', 'ssh_network': 'ctlplane', 'undercloud_key_file': '/var/lib/mistral/.ssh/tripleo-admin-rsa'}']                                                                                       
	~~~

	$ openstack overcloud failures --stack overcloud
	Ansible errors file not found at /var/lib/mistral/overcloud/ansible-errors.json



	$ openstack action execution list --fit-width|grep -v SUCCESS
	+--------------------------------------+--------------------------------------------+-------------------------------------------------------+--------------------+------------------------------+--------------------------------------+---------+----------+---------------------+---------------------+
	| ID                                   | Name                                       | Workflow name                                         | Workflow namespace | Task name                    | Task ID                              | State   | Accepted | Created at          | Updated at          |
	+--------------------------------------+--------------------------------------------+-------------------------------------------------------+--------------------+------------------------------+--------------------------------------+---------+----------+---------------------+---------------------+
	| b23cd350-78ed-449c-8b2a-3769d90b8d1a | tripleo.ansible-generate-inventory         | tripleo.deployment.v1.config_download_deploy          |                    | generate_inventory           | baea976e-6be3-4496-a30b-3d5133afaaaa | ERROR   | True     | 2020-04-20 14:07:25 | 2020-04-20 14:07:26 |
	+--------------------------------------+--------------------------------------------+-------------------------------------------------------+--------------------+------------------------------+--------------------------------------+---------+----------+---------------------+---------------------+

	$ openstack action execution show b23cd350-78ed-449c-8b2a-3769d90b8d1a --fit-width
	+--------------------+----------------------------------------------+
	| Field              | Value                                        |
	+--------------------+----------------------------------------------+
	| ID                 | b23cd350-78ed-449c-8b2a-3769d90b8d1a         |
	| Name               | tripleo.ansible-generate-inventory           |
	| Workflow name      | tripleo.deployment.v1.config_download_deploy |
	| Workflow namespace |                                              |
	| Task name          | generate_inventory                           |
	| Task ID            | baea976e-6be3-4496-a30b-3d5133afaaaa         |
	| State              | ERROR                                        |
	| State info         | None                                         |
	| Accepted           | True                                         |
	| Created at         | 2020-04-20 14:07:25                          |
	| Updated at         | 2020-04-20 14:07:26                          |
	+--------------------+----------------------------------------------+

	$ openstack action execution output show b23cd350-78ed-449c-8b2a-3769d90b8d1a 
	{
	    "result": "The action raised an exception [action_ex_id=b23cd350-78ed-449c-8b2a-3769d90b8d1a, msg='No IPs found for Controller role on ctlplane network', action_cls='<class 'mistral.actions.action_factory.AnsibleGenerateInventoryAction'>', attributes='{}', params='{'ansible_ssh_user': 'tripleo-admin', 'ansible_python_interpreter': None, 'work_dir': '/var/lib/mistral/overcloud', 'plan_name': 'overcloud', 'ssh_network': 'ctlplane', 'undercloud_key_file': '/var/lib/mistral/.ssh/tripleo-admin-rsa'}']"
	}


[1] https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/16.0/html-single/director_installation_and_usage/index#configuring-a-basic-overcloud-with-pre-provisioned-nodes
[2] https://docs.openstack.org/project-deploy-guide/tripleo-docs/latest/features/deployed_server.html#deployment-command

BR,
Pradipta

The problems outlined here have been addressed outside of the BugZilla. To recap everything so far:

1. The Undercloud does not provide DHCP / IP addressing to pre-provisioned servers. It is expected that those nodes are managed by an external source.
2. NIC configurations can be modified using a custom os-net-config configuration via a "net-config-*.yaml" template.
3. We were able to get the deployment to work with and without network isolation.

All the issues have come down to a misunderstanding of the expectations of pre-provisioned nodes and issues with incorrect custom THT settings.