fix: Do not reset the root node by default. Fixes #13196 (#13198)

Signed-off-by: oninowang <oninowang@tencent.com>

test/e2e/cli_test.go

@@ -986,6 +986,47 @@ func (s *CLISuite) TestRetryWorkflowWithContinueOn() {
 		})
 }
 
+func (s *CLISuite) TestRetryWorkflowWithFailedExitHandler() {
+	var workflowName string
+	s.Given().
+		Workflow(`@testdata/retry-workflow-with-failed-exit-handler.yaml`).
+		When().
+		SubmitWorkflow().
+		WaitForWorkflow(fixtures.ToBeFailed).
+		Then().
+		ExpectWorkflow(func(t *testing.T, metadata *metav1.ObjectMeta, status *wfv1.WorkflowStatus) {
+			workflowName = metadata.Name
+			assert.Len(t, status.Nodes, 2)
+		}).
+		RunCli([]string{"retry", workflowName}, func(t *testing.T, output string, err error) {
+			if assert.NoError(t, err, output) {
+				assert.Contains(t, output, "Name:")
+				assert.Contains(t, output, "Namespace:")
+			}
+		})
+
+	s.Given().
+		When().
+		WaitForWorkflow(fixtures.ToBeCompleted).
+		Then().
+		ExpectWorkflow(func(t *testing.T, metadata *metav1.ObjectMeta, status *wfv1.WorkflowStatus) {
+			workflowName = metadata.Name
+			assert.Equal(t, wfv1.WorkflowFailed, status.Phase)
+			assert.Len(t, status.Nodes, 2)
+		}).
+		ExpectWorkflowNode(func(status wfv1.NodeStatus) bool {
+			return status.Name == workflowName
+		}, func(t *testing.T, status *wfv1.NodeStatus, pod *corev1.Pod) {
+			assert.Equal(t, wfv1.NodeSucceeded, status.Phase)
+		}).
+		ExpectWorkflowNode(func(status wfv1.NodeStatus) bool {
+			return strings.Contains(status.Name, ".onExit")
+		}, func(t *testing.T, status *wfv1.NodeStatus, pod *corev1.Pod) {
+			assert.Equal(t, wfv1.NodeFailed, status.Phase)
+			assert.Contains(t, status.Message, "exit code 1")
+		})
+}
+
 func (s *CLISuite) TestWorkflowStop() {
 	s.Given().
 		Workflow("@smoke/basic.yaml").

test/e2e/testdata/retry-workflow-with-failed-exit-handler.yaml

@@ -0,0 +1,18 @@
+apiVersion: argoproj.io/v1alpha1
+kind: Workflow
+metadata:
+  name: retry-workflow-with-failed-exit-handler
+spec:
+  entrypoint: hello
+  onExit: exit-handler
+  templates:
+  - name: hello
+    container:
+      image: alpine:3.18
+      command: [sh, -c]
+      args: ["echo hello"]
+  - name: exit-handler
+    container:
+      image: alpine:3.18
+      command: [sh, -c]
+      args: ["exit 1"]

workflow/util/util.go

@@ -989,13 +989,6 @@ func FormulateRetryWorkflow(ctx context.Context, wf *wfv1.Workflow, restartSucce
 			// Do not allow retry of workflows with pods in Running/Pending phase
 			return nil, nil, errors.InternalErrorf("Workflow cannot be retried with node %s in %s phase", node.Name, node.Phase)
 		}
-
-		if node.Name == wf.ObjectMeta.Name {
-			log.Debugf("Reset root node: %s", node.Name)
-			newNode := node.DeepCopy()
-			newWF.Status.Nodes.Set(newNode.ID, resetNode(*newNode))
-			continue
-		}
 	}
 
 	if len(deletedNodes) > 0 {