Intialize io_thread_pool attribute in the beginning. (#18331)

Signed-off-by: rabi <[email protected]>

Author: rabi

vllm/v1/executor/multiproc_executor.py

@@ -50,6 +50,7 @@ def _init_executor(self) -> None:
         self.is_failed = False
         self.shutdown_event = threading.Event()
         self.failure_callback: Optional[FailureCallback] = None
+        self.io_thread_pool: Optional[ThreadPoolExecutor] = None
 
         self.world_size = self.parallel_config.world_size
         tensor_parallel_size = self.parallel_config.tensor_parallel_size
@@ -107,7 +108,6 @@ def _init_executor(self) -> None:
 
         # For pipeline parallel, we use a thread pool for asynchronous
         # execute_model.
-        self.io_thread_pool: Optional[ThreadPoolExecutor] = None
         if self.max_concurrent_batches > 1:
             # Note: must use only 1 IO thread to keep dequeue sequence
             # from the response queue