diff --git a/staging/src/k8s.io/kubectl/pkg/cmd/delete/delete.go b/staging/src/k8s.io/kubectl/pkg/cmd/delete/delete.go
index 09023420114..f59e99e8725 100644
--- a/staging/src/k8s.io/kubectl/pkg/cmd/delete/delete.go
+++ b/staging/src/k8s.io/kubectl/pkg/cmd/delete/delete.go
@@ -72,8 +72,8 @@ var (
 		rest of the resource.
 
 		After a CustomResourceDefinition is deleted, invalidation of discovery cache may take up
-		to 10 minutes. If you don't want to wait, you might want to run "kubectl api-resources"
-		to refresh the discovery cache.`))
+		to 6 hours. If you don't want to wait, you might want to run "kubectl api-resources" to refresh
+		the discovery cache.`))
 
 	deleteExample = templates.Examples(i18n.T(`
 		# Delete a pod using the type and name specified in pod.json
diff --git a/test/cmd/discovery.sh b/test/cmd/discovery.sh
index fdfd3c8a84d..0db5587e4e5 100755
--- a/test/cmd/discovery.sh
+++ b/test/cmd/discovery.sh
@@ -110,7 +110,7 @@ run_crd_deletion_recreation_tests() {
   output_message=$(kubectl delete -f hack/testdata/CRD/example-crd-1-cluster-scoped.yaml)
   kube::test::if_has_string "${output_message}" 'deleted'
   # Invalidate local cache because cluster scoped CRD in cache is stale.
-  # Invalidation of cache may take up to 10 minutes and we are manually
+  # Invalidation of cache may take up to 6 hours and we are manually
   # invalidate cache and expect that scope changed CRD should be created without problem.
   kubectl api-resources
   output_message=$(kubectl apply -f hack/testdata/CRD/example-crd-1-namespaced.yaml)