[PD] Fix unclosed prefill connection warning of mini_lb (#5155)
Signed-off-by: Shangming Cai <caishangming@linux.alibaba.com>
This commit is contained in:
@@ -232,6 +232,9 @@ async def handle_generate_request(request_data: dict):
|
|||||||
yield b"data: " + orjson.dumps(
|
yield b"data: " + orjson.dumps(
|
||||||
error_msg, option=orjson.OPT_NON_STR_KEYS
|
error_msg, option=orjson.OPT_NON_STR_KEYS
|
||||||
) + b"\n\n"
|
) + b"\n\n"
|
||||||
|
finally:
|
||||||
|
if prefill_response is not None:
|
||||||
|
await prefill_response.release()
|
||||||
|
|
||||||
return StreamingResponse(
|
return StreamingResponse(
|
||||||
stream_results(),
|
stream_results(),
|
||||||
|
|||||||
Reference in New Issue
Block a user