fix: instruct the controller to stop retrying under specific circumstances

Author: zijiren233

controllers/apps/configuration/sync_upgrade_policy.go

@@ -98,6 +98,9 @@ func sync(params reconfigureParams, updatedParameters map[string]string, pods []
 		return makeReturnedStatus(ESFailedAndRetry), err
 	}
 	if len(pods) == 0 {
+		if replicas == 0 {
+			return makeReturnedStatus(ESNone), nil
+		}
 		params.Ctx.Log.Info(fmt.Sprintf("no pods to update, and retry, selector: %v", selector))
 		return makeReturnedStatus(ESRetry), nil
 	}

controllers/apps/transformer_component_validation.go

@@ -91,7 +91,9 @@ func validateCompReplicas(comp *appsv1alpha1.Component, compDef *appsv1alpha1.Co
 		replicasLimit = compDef.Spec.ReplicasLimit
 	}
 	replicas := comp.Spec.Replicas
-	if replicas >= replicasLimit.MinReplicas && replicas <= replicasLimit.MaxReplicas {
+	// Allow replicas=0 (stopped state) to pass validation, as stopping a cluster is a valid operation.
+	// This prevents infinite requeue loops for stopped clusters.
+	if replicas == 0 || (replicas >= replicasLimit.MinReplicas && replicas <= replicasLimit.MaxReplicas) {
 		return nil
 	}
 	return replicasOutOfLimitError(replicas, *replicasLimit)