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, 19:45 UTC
[Misc] Add log information for handle_process_request. (#14130)
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
Chauncey committed
Mar 10, 2025 at 16:40 UTC
460f553a6d51c1ee10f8e86ec5829b473167b991
1 file changed
+2
vllm/engine/multiprocessing/engine.py
+2
@@ -292,6 +292,8 @@ class MQLLMEngine:
292
# We do not set self._errored = True here, since the error
293
# is due to an issue adding this request to the engine,
294
# rather than an issue with the engine itself.
295
+ logger.debug("Failed to add request %s to engine. %s",
296
+ request.request_id, e)
297
is_errored = self._errored_with is not None
298
rpc_err = RPCError(request_id=request_id,
299
is_engine_errored=is_errored,