[Bugfix][Pooling] Handle multimodal cache misses without crashing (#58975)

Signed-off-by: linchaoying6 <linchaoying6@jd.com>
Co-authored-by: linchaoying6 <linchaoying6@jd.com>
Co-authored-by: OpenAI Codex <codex@openai.com>
This commit is contained in:
Chaoying
2026-10-02 10:14:33 -04:00
committed by GitHub
co-authored by linchaoying6 OpenAI Codex
parent 443dccd3c9
commit 592c6f3fbc
21 changed files with 525 additions and 17 deletions
@@ -2,7 +2,7 @@
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
from http import HTTPStatus
from vllm.exceptions import GenerationError
from vllm.exceptions import GenerationError, RetryableRequestError
from vllm.logger import init_logger
from ..engine.protocol import ErrorInfo, ErrorResponse
@@ -56,7 +56,11 @@ def create_error_response(
status_code = HTTPStatus.BAD_REQUEST
param = None
elif isinstance(exc, GenerationError):
err_type = "InternalServerError"
err_type = (
"ServiceUnavailableError"
if isinstance(exc, RetryableRequestError)
else "InternalServerError"
)
status_code = exc.status_code
param = None
elif isinstance(exc, VLLMServerError):