Bug 1367843

Description of problem:

Deleting a deployment doesn't delete the associated compute resources, compute profiles, host groups, and activation keys. This can cause failed deployments if the same deployment name is used later.

Version-Release number of selected component (if applicable):

QCI-1.0-RHEL-7-20160815.t.0

How reproducible:

100%

Steps to Reproduce:
1.) Start a deployment for RHV + CFME.
2.) Delete the deployment and host records after it completes or fails.
3.) See under Configure > Host groups, Infrastructure > Compute resources, and Infrastructure > Compute profiles that the associated records still exist.
4.) Create and start a new RHV + CFME deployment with the same name.

Actual results:

The new deployment fails during the CFME installation, with error in /var/log/foreman/production.log:

----
2016-08-16 17:15:21 [foreman-tasks/action] [E] ERF56-1309 [Foreman::FingerprintException]: The remote system presented a public key signed by an unidentified certificate authority. If you are sure the remote system is authentic, go to the compute resource edit page, press the 'Test Connection' or 'Load Datacenters' button and submit (Foreman::FingerprintException)
 | /usr/share/foreman/app/models/compute_resources/foreman/model/ovirt.rb:332:in `rescue in client'
 | /usr/share/foreman/app/models/compute_resources/foreman/model/ovirt.rb:319:in `client'
 | /usr/share/foreman/app/models/compute_resources/foreman/model/ovirt.rb:14:in `clusters'
 | /opt/theforeman/tfm/root/usr/share/gems/gems/fusor_server-1.0.1/app/lib/actions/fusor/deployment/rhev/cfme_launch.rb:55:in `create_compute_profile'
 | /opt/theforeman/tfm/root/usr/share/gems/gems/fusor_server-1.0.1/app/lib/actions/fusor/deployment/rhev/cfme_launch.rb:33:in `run'
 | /opt/theforeman/tfm/root/usr/share/gems/gems/dynflow-0.8.11/lib/dynflow/action.rb:506:in `block (3 levels) in execute_run'
[...]
----

Expected results:

Host groups, Compute resources, Compute profiles, and activation keys associated with deployments don't get deleted when deleting the deployment. New deployment with the same label can fail.

Additional info:

The above error from a failed CFME deployment appears to be due to the fact that the CA certificate in the old RHEV compute resource is from the previous deployment, and doesn't work for the new deployment.