A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 11:00 UTC
Improve exception reporting in MP engine (#17800)
Signed-off-by: Vadim Markovtsev <vadim@poolside.ai>
Vadim Markovtsev committed
May 8, 2025 at 07:32 UTC
b2da14a05a33d9ec9317d4589ee09f4d57ee9672
1 file changed
+2
-2
vllm/engine/multiprocessing/engine.py
+2
-2
@@ -284,7 +284,7 @@ class MQLLMEngine:
284
except Exception as e:
285
self._set_errored(e)
286
self._send_unhealthy(e)
287
- raise e
287
+ raise e from None
288
289
def _handle_process_request(self, request: RPCProcessRequest):
290
"""Handle RPCProcessRequest by adding it to the LLMEngine."""
@@ -447,4 +447,4 @@ def run_mp_engine(vllm_config: VllmConfig, usage_context: UsageContext,
447
except BaseException as e:
448
logger.exception(e)
449
engine_alive.value = False
450
- raise e
450
+ raise e from None