.github/workflows: Fix the release quickstart.yaml GH action

Update the .github/workflows/quickstart.yml github action and replace
the usage of `kubectl wait ...` with a function that waits until the
various OLM component deployment resources are present and reporting an
available status.

Using `kubectl wait ...` is potentially problematic as it doesn't
support waiting until the creation of that resource, so in the case the
PackageServer deployment doesn't exist yet as the catalog/olm operators
are still being setup, this action will fail as `kubectl wait ...` will
return a non-zero exit code.

Signed-off-by: timflannagan <timflannagan@gmail.com>

.github/workflows/quickstart.yml

@@ -20,4 +20,24 @@ jobs:
  kubectl apply -f deploy/upstream/quickstart/crds.yaml
  kubectl wait --timeout=5m --for=condition=Established crd $(kubectl get crd --output=jsonpath='{.items[*].metadata.name}')
  kubectl apply -f deploy/upstream/quickstart/olm.yaml
- kubectl wait --timeout=5m --for=condition=Available -n olm deploy olm-operator catalog-operator packageserver
+
+ # Note(tflannag): `kubectl wait` does not support waiting for resource creation: https://github.com/kubernetes/kubernetes/pull/87399.
+ wait_for_deployment() {
+ local deployment_name=$1
+ timeout=60
+ i=1
+ echo "Checking if the ${deployment_name} deployment is ready"
+ until kubectl -n olm get deployment ${deployment_name} -o jsonpath='{.status.conditions[?(@.status=="True")].type}' | grep "Available" 2>/dev/null; do
+ ((i++))
+ if [[ ${i} -gt ${timeout} ]]; then
+ echo "the ${deployment_name} deployment has not become ready before the timeout period"
+ exit 1
+ fi
+ echo "waiting for ${deployment_name} deployment to report a ready status"
+ sleep 5
+ done
+ echo "The ${deployment_name} deployment is ready"
+ }
+ wait_for_deployment catalog-operator
+ wait_for_deployment olm-operator
+ wait_for_deployment packageserver