[None][fix] Fix port conflict in disagg tests (#9474)

Signed-off-by: Junyi Xu <[email protected]>

Author: JunyiXu-nv

tests/integration/defs/accuracy/test_disaggregated_serving.py

@@ -67,6 +67,41 @@ def __exit__(self, exc_type, exc_val, exc_tb):
         return False
 
 
+def check_port_available(port: int) -> int:
+    import socket
+    try:
+        with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
+            s.bind(('localhost', port))
+            return port
+    except socket.error:
+        # find a free port
+        sock = socket.socket()
+        sock.bind(('', 0))
+        return sock.getsockname()[1]
+
+
+def revise_disaggregated_server_config_urls_with_free_ports(
+        disaggregated_server_config: Dict[str, Any]) -> Dict[str, Any]:
+    disaggregated_server_config['port'] = check_port_available(
+        disaggregated_server_config['port'])
+    ctx_urls = disaggregated_server_config["context_servers"]["urls"]
+    gen_urls = disaggregated_server_config["generation_servers"]["urls"]
+
+    new_ctx_urls = []
+    new_gen_urls = []
+    for url in ctx_urls:
+        port = check_port_available(int(url.split(":")[1]))
+        new_ctx_urls.append(f"localhost:{port}")
+    for url in gen_urls:
+        port = check_port_available(int(url.split(":")[1]))
+        new_gen_urls.append(f"localhost:{port}")
+
+    disaggregated_server_config["context_servers"]["urls"] = new_ctx_urls
+    disaggregated_server_config["generation_servers"]["urls"] = new_gen_urls
+
+    return disaggregated_server_config
+
+
 @contextlib.contextmanager
 def launch_disaggregated_llm(
         disaggregated_server_config: Dict[str, Any],
@@ -87,6 +122,9 @@ def launch_disaggregated_llm(
             f"Using unified tp parameter for testing is not recommended. Please use server configs instead."
         )
 
+    disaggregated_server_config = revise_disaggregated_server_config_urls_with_free_ports(
+        disaggregated_server_config)
+
     with open(disaggregated_serving_config_path, "w") as f:
         yaml.dump(disaggregated_server_config, f)
     ctx_server_config_path = os.path.join(temp_dir.name,
@@ -138,6 +176,7 @@ def launch_disaggregated_llm(
     ctx_urls = disaggregated_server_config["context_servers"]["urls"]
     gen_urls = disaggregated_server_config["generation_servers"]["urls"]
 
+    serve_port = disaggregated_server_config["port"]
     ctx_ports = [int(url.split(":")[1]) for url in ctx_urls]
     gen_ports = [int(url.split(":")[1]) for url in gen_urls]
 
@@ -236,14 +275,14 @@ def multi_popen(server_configs, server_name="", enable_redirect_log=False):
                     )
             try:
                 print("Checking health endpoint")
-                response = requests.get("http://localhost:8000/health")
+                response = requests.get(f"http://localhost:{serve_port}/health")
                 if response.status_code == 200:
                     break
             except requests.exceptions.ConnectionError:
                 continue
 
         client = openai.OpenAI(api_key="1234567890",
-                               base_url=f"http://localhost:8000/v1",
+                               base_url=f"http://localhost:{serve_port}/v1",
                                timeout=1800000)
 
         def send_request(prompt: str, sampling_params: SamplingParams,