ci-kubernetes-e2e-gce-device-plugin-gpu needs docker as default runtime

A recent change in kube-up switched us to using `containerd` as default, so we need to fix up CI jobs that depend on docker.
kubernetes · Jun 10, 2020 · 048ea1c · 048ea1c
1 parent c85932e
commit 048ea1c
Showing 1 changed file with 1 addition and 0 deletions.
diff --git a/config/jobs/kubernetes/sig-cloud-provider/gcp/gpu/gpu-gce.yaml b/config/jobs/kubernetes/sig-cloud-provider/gcp/gpu/gpu-gce.yaml
@@ -35,6 +35,7 @@ periodics:
       - --
       - --check-leaked-resources
       - --extract=ci/latest
+      - --env=KUBE_CONTAINER_RUNTIME=docker
       - --gcp-node-image=gci
       - --gcp-project-type=gpu-project
       - --gcp-zone=us-west1-b