in 4.20, when TP is enabled, there's a CAPM3 deployment in openshift-cluster-api namespace.
They are using the same base manifest, hence a lot of cluster-wide resources have the same name and get overridden.
We need to:
- rename conflicting (or probably all) resources
- make sure we have a system to not reconcile resources in openshift-cluster-api
- this should be done in a similar way to CAPI. (TBD rh-ee-rpiccoli)