From f668bd0ef1c82c60e59752b50fcec3be6490e5ca Mon Sep 17 00:00:00 2001
From: zhengd-nv <200704041+zhengd-nv@users.noreply.github.com>
Date: Fri, 4 Jul 2025 05:53:15 +0000
Subject: [PATCH] log stack trace on error

Signed-off-by: zhengd-nv <200704041+zhengd-nv@users.noreply.github.com>
---
 tensorrt_llm/serve/openai_disagg_server.py | 5 +++--
 tensorrt_llm/serve/openai_server.py        | 5 ++++-
 2 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/tensorrt_llm/serve/openai_disagg_server.py b/tensorrt_llm/serve/openai_disagg_server.py
index 62a64a1cd73..6eb05971b60 100644
--- a/tensorrt_llm/serve/openai_disagg_server.py
+++ b/tensorrt_llm/serve/openai_disagg_server.py
@@ -3,6 +3,7 @@
 import copy
 import os
 import signal
+import traceback
 from contextlib import asynccontextmanager
 from http import HTTPStatus
 from typing import List, Optional, Type, Union
@@ -168,12 +169,12 @@ async def openai_chat_completion(self, req: ChatCompletionRequest) -> Response:
 
     async def _handle_exception(self, exception):
         if isinstance(exception, CppExecutorError):
-            logger.error(exception)
+            logger.error(traceback.format_exc())
             signal.raise_signal(signal.SIGINT)
         elif isinstance(exception, HTTPException):
             raise exception  # Re-raise HTTP exceptions properly
         else:
-            logger.error(exception)
+            logger.error(traceback.format_exc())
             raise HTTPException(status_code=500, detail=f"Internal server error {str(exception)}")
 
     async def _send_context_request(self, ctx_server: str, ctx_req: Union[CompletionRequest, ChatCompletionRequest]):
diff --git a/tensorrt_llm/serve/openai_server.py b/tensorrt_llm/serve/openai_server.py
index 63f7e82c73d..8a5c55a1090 100644
--- a/tensorrt_llm/serve/openai_server.py
+++ b/tensorrt_llm/serve/openai_server.py
@@ -310,9 +310,11 @@ async def create_chat_response(
                 response = await create_chat_response(promise, postproc_params, disaggregated_params)
                 return JSONResponse(content=response.model_dump())
         except CppExecutorError:
+            logger.error(traceback.format_exc())
             # If internal executor error is raised, shutdown the server
             signal.raise_signal(signal.SIGINT)
         except Exception as e:
+            logger.error(traceback.format_exc())
             return self.create_error_response(str(e))
 
     async def openai_completion(self, request: CompletionRequest, raw_request: Request) -> Response:
@@ -436,10 +438,11 @@ async def create_completion_response(
                     generator, disaggregated_params)
                 return JSONResponse(content=response.model_dump())
         except CppExecutorError:
+            logger.error(traceback.format_exc())
             # If internal executor error is raised, shutdown the server
             signal.raise_signal(signal.SIGINT)
         except Exception as e:
-            traceback.print_exc()
+            logger.error(traceback.format_exc())
             return self.create_error_response(str(e))
 
     async def __call__(self, host, port):