Merge pull request #31330 from ixdy/kubekins-docker

Automatic merge from submit-queue

Explicitly set CLOUDSDK_CONFIG override early in e2e-runner

I'm guessing this is why a number of gci-on-gke builds started failing after https://github.com/kubernetes/test-infra/pull/432 was merged - we're overriding the gcloud configuration location from the default, so the activated service account no longer works. Previously we did this service activation late, but after we'd overridden the value.

I've changed it to explicitly set to a known-good value every time at the very start.

cc @vishh @maisem @spxtr @fejta

hack/jenkins/e2e-runner.sh

@@ -48,6 +48,10 @@ STAGE_KUBEMARK="KUBEMARK"
 : ${KUBE_GCS_RELEASE_BUCKET:="kubernetes-release"}
 : ${KUBE_GCS_DEV_RELEASE_BUCKET:="kubernetes-release-dev"}
 
+# Explicitly set config path so staging gcloud (if installed_ uses the same path
+# as release
+export CLOUDSDK_CONFIG="${WORKSPACE}/.config/gcloud"
+
 # record_command runs the command and records its output/error messages in junit format
 # it expects the first argument to be the class and the second to be the name of the command
 # Example:
@@ -228,8 +232,6 @@ if [[ -n "${CLOUDSDK_BUCKET:-}" ]]; then
     mv ~/$(basename "${CLOUDSDK_BUCKET}") ~/repo
     export CLOUDSDK_COMPONENT_MANAGER_SNAPSHOT_URL=file://${HOME}/repo/components-2.json
     install_google_cloud_sdk_tarball ~/repo/google-cloud-sdk.tar.gz ~/cloudsdk
-    # TODO: is this necessary? this won't work inside Docker currently.
-    export CLOUDSDK_CONFIG=/var/lib/jenkins/.config/gcloud
 fi
 
 # GCI specific settings.