diff --git a/hack/local-up-cluster.sh b/hack/local-up-cluster.sh
index 1920670e5a4..f0243194aef 100755
--- a/hack/local-up-cluster.sh
+++ b/hack/local-up-cluster.sh
@@ -72,6 +72,8 @@ CLOUD_PROVIDER=${CLOUD_PROVIDER:-""}
 CLOUD_CONFIG=${CLOUD_CONFIG:-""}
 FEATURE_GATES=${FEATURE_GATES:-"AllAlpha=false"}
 STORAGE_BACKEND=${STORAGE_BACKEND:-"etcd3"}
+# preserve etcd data. you also need to set ETCD_DIR.
+PRESERVE_ETCD="${PRESERVE_ETCD:-false}"
 # enable swagger ui
 ENABLE_SWAGGER_UI=${ENABLE_SWAGGER_UI:-false}
 # enable Pod priority and preemption
@@ -378,8 +380,9 @@ cleanup()
 
   # Check if the etcd is still running
   [[ -n "${ETCD_PID-}" ]] && kube::etcd::stop
-  [[ -n "${ETCD_DIR-}" ]] && kube::etcd::clean_etcd_dir
-
+  if [[ "${PRESERVE_ETCD}" == "false" ]]; then
+    [[ -n "${ETCD_DIR-}" ]] && kube::etcd::clean_etcd_dir
+  fi
   exit 0
 }