Fix: Reject requests with a duplicate request ID which can cause server crash/hang (#19035)

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: hnyls2002 <lsyincs@gmail.com>
This commit is contained in:
Charles Chen
2026-03-03 00:33:25 -08:00
committed by GitHub
parent 6af0448cc9
commit af0d35b224
2 changed files with 97 additions and 0 deletions

View File

@@ -482,6 +482,7 @@ class TokenizerManager(TokenizerCommunicatorMixin, TokenizerManagerMultiItemMixi
# Normalize the request
obj.normalize_batch_and_arguments()
self._validate_rid(obj)
self._req_stats_init(obj, request)
if self.server_args.language_only:
@@ -737,6 +738,21 @@ class TokenizerManager(TokenizerCommunicatorMixin, TokenizerManagerMultiItemMixi
obj, input_text, input_ids, input_embeds, mm_inputs, token_type_ids
)
def _validate_rid(self, obj: Union[GenerateReqInput, EmbeddingReqInput]) -> None:
"""Validate the request ID (rid) uniqueness."""
rid = obj.rid
if rid is None:
return
ids = rid if isinstance(rid, list) else [rid]
if len(ids) != len(set(ids)):
raise ValueError(
f"Duplicate request IDs detected within the request: {ids}"
)
for i in ids:
if i in self.rid_to_state:
raise ValueError(f"Duplicate request ID detected: {i}")
def _validate_one_request(
self, obj: Union[GenerateReqInput, EmbeddingReqInput], input_ids: List[int]
) -> None: