Retrying applying CRs in the e2e jobs.

The webhook services are not always available right after the
deployments are ready. This is causing a race condition between the
services being ready to get requests and the CRs applied to the cluster.

By retrying to apply the CRs we can give the services the time they need
to become ready.

Signed-off-by: Yoni Bettan <[email protected]>

Author: ybettan

ci/prow/e2e-hub-spoke-incluster-build

@@ -59,7 +59,7 @@ kubectl wait --for=condition=Available -n ${OPERATOR_NAMESPACE} \
 kubectl label managedcluster minikube name=minikube
 
 # Apply the ManagedClusterModule
-kubectl apply -f ci/managedclustermodule-kmm-ci-build-sign.yaml
+timeout 1m bash -c 'until kubectl apply -f ci/managedclustermodule-kmm-ci-build-sign.yaml; do sleep 3; done'
 
 # Waiting for the manifestwork to be created
 timeout 1m bash -c 'until kubectl -n ${MINIKUBE} get manifestwork/mod-example; do sleep 1; done'

ci/prow/e2e-incluster-build

@@ -36,7 +36,7 @@ kubectl apply -f ci/secret_kmm-kmod-signing.yaml
 kubectl wait --for=condition=Available deployment/kmm-operator-controller deployment/kmm-operator-webhook -n kmm-operator-system
 
 echo "Add an kmm-ci Module that contains a valid mapping..."
-kubectl apply -f ci/module-kmm-ci-build-sign.yaml
+timeout 1m bash -c 'until kubectl apply -f ci/module-kmm-ci-build-sign.yaml; do sleep 3; done'
 
 echo "Check that the module gets loaded on the node..."
 timeout 10m bash -c 'until minikube ssh -- lsmod | grep kmm_ci_a; do sleep 3; done'