diff --git a/doc/COMPILE.md b/doc/COMPILE.md
index 3ebab6e11e7de6208de4637aefaab81b3be71e05..f4a6639bdb38fac97662084f7d927d24b6179717 100644
--- a/doc/COMPILE.md
+++ b/doc/COMPILE.md
@@ -36,6 +36,8 @@ cd Serving && git submodule update --init --recursive
 export PYTHONROOT=/usr/
 ```
 
+In the default centos7 image we provide, the Python path is `/usr/bin/python`. If you want to use our centos6 image, you need to set it to `export PYTHONROOT=/usr/local/python2.7/`.
+
 ## Compile Server
 
 ### Integrated CPU version paddle inference library
diff --git a/doc/COMPILE_CN.md b/doc/COMPILE_CN.md
index 19123a220e4c69fbbe56ef639fd59b1b15fa2249..d8fd277131d7d169c1a47689e15556e5d10a0fdb 100644
--- a/doc/COMPILE_CN.md
+++ b/doc/COMPILE_CN.md
@@ -36,6 +36,8 @@ cd Serving && git submodule update --init --recursive
 export PYTHONROOT=/usr/
 ```
 
+我们提供默认Centos7的Python路径为`/usr/bin/python`，如果您要使用我们的Centos6镜像，需要将其设置为`export PYTHONROOT=/usr/local/python2.7/`。
+
 ## 编译Server部分
 
 ### 集成CPU版本Paddle Inference Library
diff --git a/python/examples/fit_a_line/test_multi_process_client.py b/python/examples/fit_a_line/test_multi_process_client.py
index 46ba3b60b5ae09b568868531d32234ade50d8556..5272d095df5e74f25ce0e36ca22c8d6d1884f5f0 100644
--- a/python/examples/fit_a_line/test_multi_process_client.py
+++ b/python/examples/fit_a_line/test_multi_process_client.py
@@ -22,15 +22,19 @@ def single_func(idx, resource):
     client.load_client_config(
         "./uci_housing_client/serving_client_conf.prototxt")
     client.connect(["127.0.0.1:9293", "127.0.0.1:9292"])
-    test_reader = paddle.batch(
-        paddle.reader.shuffle(
-            paddle.dataset.uci_housing.test(), buf_size=500),
-        batch_size=1)
-    for data in test_reader():
-        fetch_map = client.predict(feed={"x": data[0][0]}, fetch=["price"])
+    x = [
+        0.0137, -0.1136, 0.2553, -0.0692, 0.0582, -0.0727, -0.1583, -0.0584,
+        0.6283, 0.4919, 0.1856, 0.0795, -0.0332
+    ]
+    for i in range(1000):
+        fetch_map = client.predict(feed={"x": x}, fetch=["price"])
+        if fetch_map is None:
+            return [[None]]
     return [[0]]
 
 
 multi_thread_runner = MultiThreadRunner()
 thread_num = 4
 result = multi_thread_runner.run(single_func, thread_num, {})
+if None in result[0]:
+    exit(1)
diff --git a/python/paddle_serving_server_gpu/__init__.py b/python/paddle_serving_server_gpu/__init__.py
index d4631141f8173b4ae0cb41d42c615566ac81ae7e..e40c0fa48763eaa66373e9f2149552c4f8693eb7 100644
--- a/python/paddle_serving_server_gpu/__init__.py
+++ b/python/paddle_serving_server_gpu/__init__.py
@@ -384,7 +384,7 @@ class Server(object):
                 finally:
                     os.remove(tar_name)
         #release lock
-        version_file.cloes()
+        version_file.close()
         os.chdir(self.cur_path)
         self.bin_path = self.server_path + "/serving"
 
diff --git a/tools/serving_build.sh b/tools/serving_build.sh
index 8e78e13ef8e86b55af6a90df1b9235611508c0ba..7ac1ed31c22b0271e06266d9f905d94b77d1897c 100644
--- a/tools/serving_build.sh
+++ b/tools/serving_build.sh
@@ -375,16 +375,17 @@ function python_test_multi_process(){
     sh get_data.sh
     case $TYPE in
         CPU)
-            check_cmd "python -m paddle_serving_server.serve --model uci_housing_model --port 9292 &"
-            check_cmd "python -m paddle_serving_server.serve --model uci_housing_model --port 9293 &"
+            check_cmd "python -m paddle_serving_server.serve --model uci_housing_model --port 9292 --workdir test9292 &"
+            check_cmd "python -m paddle_serving_server.serve --model uci_housing_model --port 9293 --workdir test9293 &"
             sleep 5
             check_cmd "python test_multi_process_client.py"
             kill_server_process
             echo "bert mutli rpc RPC inference pass"
             ;;
         GPU)
-            check_cmd "python -m paddle_serving_server_gpu.serve --model uci_housing_model --port 9292 --gpu_ids 0 &"
-            check_cmd "python -m paddle_serving_server_gpu.serve --model uci_housing_model --port 9293 --gpu_ids 0 &"
+            rm -rf ./image #TODO: The following code tried to create this folder, but no corresponding code was found
+            check_cmd "python -m paddle_serving_server_gpu.serve --model uci_housing_model --port 9292 --workdir test9292 --gpu_ids 0 &"
+            check_cmd "python -m paddle_serving_server_gpu.serve --model uci_housing_model --port 9293 --workdir test9293 --gpu_ids 0 &"
             sleep 5
             check_cmd "python test_multi_process_client.py"
             kill_server_process