[Bugfix] Improve engine ready timeout error message (#35616)

Signed-off-by: damaozi <1811866786@qq.com>

vllm/v1/engine/core_client.py

@@ -609,8 +609,13 @@ class MPClient(EngineCoreClient):
                     timeout=VLLM_ENGINE_READY_TIMEOUT_S * 1000  # convert to ms
                 ):
                     raise TimeoutError(
-                        "Timed out waiting for engines to send "
-                        "initial message on input socket."
+                        f"Timed out waiting for engine core processes to "
+                        f"start. This is often caused by slow weight loading "
+                        f"for large models. Waited "
+                        f"{VLLM_ENGINE_READY_TIMEOUT_S}s (configured by "
+                        f"VLLM_ENGINE_READY_TIMEOUT_S). To increase the "
+                        f"timeout, set the environment variable: "
+                        f"VLLM_ENGINE_READY_TIMEOUT_S=<seconds>"
                     )
                 identity, _ = sync_input_socket.recv_multipart()
                 identities.remove(identity)
@@ -1586,8 +1591,12 @@ class DPLBAsyncMPClient(DPAsyncMPClient):
                 timeout=VLLM_ENGINE_READY_TIMEOUT_S * 1000  # convert to ms
             ):
                 raise TimeoutError(
-                    "Timed out waiting for new engines to send initial "
-                    "message on input socket."
+                    f"Timed out waiting for new engine core processes to "
+                    f"start. Waited "
+                    f"{VLLM_ENGINE_READY_TIMEOUT_S}s (configured by "
+                    f"VLLM_ENGINE_READY_TIMEOUT_S). To increase the "
+                    f"timeout, set the environment variable: "
+                    f"VLLM_ENGINE_READY_TIMEOUT_S=<seconds>"
                 )
             identity, _ = sync_input_socket.recv_multipart()
             new_engine_identities.discard(identity)