fix bind failed with Address already in use (#38174)

python/paddle/distributed/fleet/base/private_helper_function.py

@@ -43,6 +43,10 @@ def wait_server_ready(endpoints):
             with closing(socket.socket(socket.AF_INET,
                                        socket.SOCK_STREAM)) as sock:
                 sock.settimeout(2)
+                sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
+                if hasattr(socket, 'SO_REUSEPORT'):
+                    sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEPORT, 1)
+
                 result = sock.connect_ex((ip_port[0], int(ip_port[1])))
                 if result != 0:
                     all_ok = False

python/paddle/fluid/tests/unittests/npu/test_collective_base_npu.py

@@ -46,6 +46,11 @@ class TestCollectiveRunnerBase(object):
                         socket.socket(socket.AF_INET,
                                       socket.SOCK_STREAM)) as sock:
                     sock.settimeout(2)
+                    sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
+                    if hasattr(socket, 'SO_REUSEPORT'):
+                        sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEPORT,
+                                        1)
+
                     result = sock.connect_ex((ip_port[0], int(ip_port[1])))
                     if result != 0:
                         all_ok = False

python/paddle/fluid/tests/unittests/npu/test_sync_batch_norm_base_npu.py

@@ -65,6 +65,11 @@ class TestSyncBatchNormRunnerBase(object):
                         socket.socket(socket.AF_INET,
                                       socket.SOCK_STREAM)) as sock:
                     sock.settimeout(2)
+                    sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
+                    if hasattr(socket, 'SO_REUSEPORT'):
+                        sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEPORT,
+                                        1)
+
                     result = sock.connect_ex((ip_port[0], int(ip_port[1])))
                     if result != 0:
                         all_ok = False

python/paddle/fluid/tests/unittests/test_collective_base.py

@@ -44,6 +44,11 @@ class TestCollectiveRunnerBase(object):
                         socket.socket(socket.AF_INET,
                                       socket.SOCK_STREAM)) as sock:
                     sock.settimeout(2)
+                    sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
+                    if hasattr(socket, 'SO_REUSEPORT'):
+                        sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEPORT,
+                                        1)
+
                     result = sock.connect_ex((ip_port[0], int(ip_port[1])))
                     if result != 0:
                         all_ok = False

python/paddle/fluid/transpiler/details/checkport.py

@@ -42,6 +42,10 @@ def wait_server_ready(endpoints):
             with closing(socket.socket(socket.AF_INET,
                                        socket.SOCK_STREAM)) as sock:
                 sock.settimeout(2)
+                sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
+                if hasattr(socket, 'SO_REUSEPORT'):
+                    sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEPORT, 1)
+
                 result = sock.connect_ex((ip_port[0], int(ip_port[1])))
                 if result != 0:
                     all_ok = False