controller-1 does not allow for openstack commands

[ruediger stock]

Brief Description
-----------------
controller-1 does not have the capabilities to run the openstack command

Severity
--------
Major

Steps to Reproduce
------------------
controller-0: source /etc/platform/openrc
controller-0: system host-swact controller-0
Then on
controller-1: OS_CLOUD=openstack_helm openstack service list
Cloud openstack_helm was not found.

On controller-1 there are only:
/etc/openldap and /etc/openstack-dashboard
/etc/openstack is missing on controller-1!

Expected Behavior
------------------
controller-1 should be a full replacement of controller-0 without any restriction.

Actual Behavior
----------------
controller-1 only allows for parts of the commands that are possible on controller-0

Reproducibility
---------------
Reproducible

System Configuration
--------------------
Multi-node system, Dedicated storage in virtual environment

Branch/Pull Time/Commit
-----------------------
cengn load http://mirror.starlingx.cengn.ca/mirror/starlingx/rc/4.0/centos/monolithic/20200715T174043Z/

Last Pass
---------
Sorry, unknown for now.

Timestamp/Logs
--------------
https://files.starlingx.kube.cengn.ca/download_file/289

Test Activity
-------------
Regression Testing

Workaround
----------
No workaround known

[Ghada Khalil (gkhalil)]

Assigning to the openstack team for review

[Ran An (an.ran)]

the config file "clouds.yaml" under "/etc/openstack" should be created manually on either controller before it could access openstack loacl CLI.

see https://docs.starlingx.io/deploy_install_guides/r4_release/openstack/access.html#local-cli to get more.

[ruediger stock (rstock)]

After creating /etc/openstack/clouds.yaml as described in the documentation openstack commands still don't work as expected:

controller-1:~$ openstack service list
Internal Server Error (HTTP 500)

controller-1 is the active controller - obviously.

[Ran An (an.ran)]

the openstack CLI was supported on both active and standby controllers once it was config well.

So pls make sure your openstack are running after host-swact.
1. check pods under "openstack" namespace were all running
2. make sure cmd "openstack service list" still worked as expected on controller-0 when the active controller is controller-1.

[ruediger stock (rstock)]

I can't reproduce the issue I saw 07-24th on the load from rc/4.0/centos/flock/20200724T195511Z.
The identical number and similar name of pods is running on both controllers.
"openstack service list" works on both controllers, regardless which one is the active one.