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]>
rh-ecosystem-edge · Jan 12, 2025 · 0835c06 · 0835c06
1 parent 36447e9
commit 0835c06
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 5 deletions.
diff --git a/ci/prow/e2e-hub-spoke-incluster-build b/ci/prow/e2e-hub-spoke-incluster-build
@@ -1,7 +1,3 @@
-#!/bin/bash
-#
-set -euxo pipefail
-
 readonly SKIP_MAKE_DEPLOY="${SKIP_MAKE_DEPLOY:-false}"
 readonly SKIP_MAKE_DEPLOY_HUB="${SKIP_MAKE_DEPLOY_HUB:-false}"
 readonly SKIP_ACM_INSTALLATION="${SKIP_ACM_INSTALLATION:-false}"
@@ -51,7 +47,7 @@ else
 fi
 
 # Apply resources
-oc apply -k ci/e2e-hub
+timeout 1m bash -c 'until oc apply -k ci/e2e-hub; do sleep 3; done'
 
 # Wait for the build build to be created and completed
 timeout 1m bash -c 'until oc -n ${HUB_OPERATOR_NAMESPACE} get builds -o json | jq -er ".items[].metadata.name | select(.? | match(\"build\"))"; do sleep 1; done'

diff --git a/ci/prow/e2e-incluster-build b/ci/prow/e2e-incluster-build
@@ -42,6 +42,7 @@ oc wait --for=condition=Available deployment/kmm-operator-controller deployment/
 
 echo "Add resources"
 oc apply -k ci/e2e
+timeout 1m bash -c 'until oc apply -k ci/e2e; do sleep 3; done'
 
 # Wait for the build pod to be created. `kubectl wait` doesn't support such option,
 # see https://github.com/kubernetes/kubernetes/issues/83242.