[BugFix] Fix profile run in pd-disaggregated deployment (#4584)

* [fix] fix pd+dp+ep bug

* [fix] fix again

* [ci] fix code style

Author: liyonghua0910

fastdeploy/engine/common_engine.py

@@ -760,6 +760,9 @@ def _fetch_request():
                 else:
                     time.sleep(0.005)
 
+            except RuntimeError as e:
+                if "cannot schedule new futures after shutdown" in str(e):
+                    break
             except Exception as e:
                 err_msg = "Error happend while insert task to engine: {}, {}.".format(e, str(traceback.format_exc()))
                 self.llm_logger.error(err_msg)

fastdeploy/engine/engine.py

@@ -101,6 +101,25 @@ def start(self, api_server_pid=None):
         Initializes the engine and starts its sub-services.
         If `api_server_pid` is defined, will launch a thread
         to keep getting request from zmq_server.
+
+        NOTE: To clarify the launch order of the components of the LLM engine:
+        1. First, launch splitwise scheduler (if necessary) and expert services (if necessary).
+        2. Then, launch common engine, which includes some background threads that inserts tasks and receives ouptuts.
+        3. Most importantly, launch workers and cache services. The launch order of them are listed as follows.
+
+            | Profile | Mixed | PrefixCache | Cache -> Worker | Worker -> Cache |
+            |---------|-------|-------------|-----------------|-----------------|
+            | 1       | 1     | 1           | 0               | 1               |
+            | 1       | 1     | 0           | 0               | 0               |
+            | 1       | 0     | 1           | 0               | 1               |
+            | 1       | 0     | 0           | 0               | 1               |
+            | 0       | 1     | 1           | 0               | 1               |
+            | 0       | 1     | 0           | 0               | 0               |
+            | 0       | 0     | 1           | 1               | 0               |
+            | 0       | 0     | 0           | 1               | 0               |
+
+        4. Finally, inform user the engine has successfully started.
+
         """
         assert not self.is_started, "The engine is already started."
         start_time = time.time()
@@ -109,7 +128,6 @@ def start(self, api_server_pid=None):
         self.ipc_signal_suffix = self.cfg.parallel_config.engine_worker_queue_port[0]
         self._init_worker_signals()
 
-        # Launch components: scheduler, cache_manager, expert_service et.al.
         self.launch_components()
 
         self.engine.start()
@@ -151,7 +169,7 @@ def check_worker_initialize_status_func(res: dict):
         # and then start the cache manager
         if self.do_profile:
             self._stop_profile()
-        elif self.cfg.cache_config.enable_prefix_caching:
+        elif self.cfg.scheduler_config.splitwise_role == "mixed" and self.cfg.cache_config.enable_prefix_caching:
             device_ids = self.cfg.parallel_config.device_ids.split(",")
             self.cache_manager_processes = self.engine.start_cache_service(device_ids, self.ipc_signal_suffix)
 

fastdeploy/inter_communicator/ipc_signal.py

@@ -80,13 +80,15 @@ def __init__(
             name = name + f".{suffix}"
 
         if create:
+            llm_logger.debug(f"creating ipc signal: {name}")
             if shared_memory_exists(name):
                 llm_logger.warning(f"ShareMemory: {name} already exists, delete it")
                 SharedMemory(name=name, create=False).unlink()
             self.shm = SharedMemory(create=True, size=array.nbytes, name=name)
             self.value: np.ndarray = np.ndarray(array.shape, dtype=array.dtype, buffer=self.shm.buf)
             self.value[:] = array  # Initialize with input array data
         else:
+            llm_logger.debug(f"attaching ipc signal: {name}")
             self.shm = SharedMemory(name=name)
             self.value: np.ndarray = np.ndarray(array.shape, dtype=array.dtype, buffer=self.shm.buf)
 

fastdeploy/worker/worker_process.py

@@ -41,7 +41,12 @@
     StructuredOutputsConfig,
 )
 from fastdeploy.inter_communicator import EngineWorkerQueue as TaskQueue
-from fastdeploy.inter_communicator import ExistTaskStatus, IPCSignal, ModelWeightsStatus
+from fastdeploy.inter_communicator import (
+    ExistTaskStatus,
+    IPCSignal,
+    ModelWeightsStatus,
+    shared_memory_exists,
+)
 from fastdeploy.model_executor.layers.quantization import parse_quant_config
 from fastdeploy.model_executor.utils import v1_loader_support
 from fastdeploy.platforms import current_platform
@@ -426,7 +431,7 @@ def graph_optimize_and_warm_up_model(self) -> None:
                 array=prefilled_step_idx_data,
                 dtype=np.int32,
                 suffix=gpu_id,
-                create=False,
+                create=not shared_memory_exists(prefilled_step_name),
             )
             step_shm_value.value[0] = -1