Follow-up from "First commit"

The following discussion from !1 (merged) should be addressed:

• [DEVELOPMENT] Improve instructions on multiple reconcile loops !1 (comment 4394842)

• [TESTING] Improve section on exporting variables !1 (comment 4394629)

• [TESTING] This section should be improved to mention we have to disable ArgoCD and possibly update CRDs and only then should we worry about running the operator. Point raised by @alossent started on discussion: (+1 comment)

  Are you sure about this? It seems that with operator-sdk 1.4 the purpose of run seems to have changed completely, it only deploys via OLM. To run the operator locally I'm doing this:

  # set env vars...
  
  # disable argocd and in-cluster operator (to undo, run the same with --replicas=1)
  for d in argocd-server argocd-application-controller; do oc scale deploy/$d -n openshift-cern-argocd --replicas=0; done
  oc scale deploy/authz-operator -n openshift-cern-authz-operator --replicas=0
  
  # If changes to CRDs...
  # make manifests && oc replace -f config/crd/bases
  
  # build & run...
  make
  bin/manager --zap-log-level=3

• [TESTING] Improve sections on automated testing !1 (comment 4394630)