add support for OCP

[nddipiazza]

Make these scripts work for on premise Openshift - OCP 3.x and 4.x

Some changes needed to these scripts to make them work for OCP.

Changes

• we were hard coding the kubectl binary as kubectl. Make a parameter -k that allows openshift people to use oc instead of forcing them to use kubectl.

• When using oc instead of kubectl

• For oc, instead of --timeout we need to use --request-timeout.

• When using oc some of the checks on existing namespace are not valid so we should skip those. This allows users to run the k8s script, make adjustments, run the k8s script, make adjustments until they have it right.

• Dry runs are super important part of the process with OCP. So make some changes to to the scripts so that we can actually run --dry-run without it causing errors.

• It is not unlikely you find something you need to fix in the helm chart, and you don't want to wait for a new release. Add a new LOCAL_HELM_CHART variable that allows you to use a pulled down copy of the helm charts.

• OpenShift on prem users are typically not admins so

  • Create a "no-rbac" yaml file that they can use to avoid permissions issues.
  • Add a --skip-crds parameter so that when you need this feature, you don't need to update one of the shell scripts. We will need to send the CRDs to OpenShift admins to run for us.

• Add a example-values/repository.yaml to show an example of setting up an internal docker repository. This is important for many openshift users because it is very common to have openshift with no external internet access to public docker repositories.