diff --git a/cpp/src/server/VecServiceTask.cpp b/cpp/src/server/VecServiceTask.cpp
index 9c996cb3a1..c2435ff51b 100644
--- a/cpp/src/server/VecServiceTask.cpp
+++ b/cpp/src/server/VecServiceTask.cpp
@@ -488,10 +488,14 @@ ServerError AddBatchVectorTask::OnExecute() {
                 std::list<std::future<void>> threads_list;
 
                 uint64_t begin_index = 0, end_index = USE_MT;
-                while(end_index < vec_count) {
+                while(true) {
                     threads_list.push_back(
                             GetThreadPool().enqueue(&AddBatchVectorTask::ProcessIdMapping,
                                                this, vector_ids, begin_index, end_index, tensor_ids_));
+                    if(end_index >= vec_count) {
+                        break;
+                    }
+
                     begin_index = end_index;
                     end_index += USE_MT;
                     if(end_index > vec_count) {
diff --git a/cpp/test_client/src/ClientTest.cpp b/cpp/test_client/src/ClientTest.cpp
index 3a11e8bcc3..188f30a02e 100644
--- a/cpp/test_client/src/ClientTest.cpp
+++ b/cpp/test_client/src/ClientTest.cpp
@@ -211,7 +211,10 @@ TEST(AddVector, CLIENT_TEST) {
                 server::TimeRecorder rc("Add " + std::to_string(count) + " binary vectors in one batch");
                 std::vector<std::string> ids;
                 session.interface()->add_binary_vector_batch(ids, group.id, bin_tensor_list_2);
-                ASSERT_TRUE(!ids.empty());
+                ASSERT_EQ(ids.size(), bin_tensor_list_2.tensor_list.size());
+                for(size_t i = 0; i < ids.size(); i++) {
+                    ASSERT_TRUE(!ids[i].empty());
+                }
                 rc.Elapse("done!");
             }
 #endif