[BugFix][V1] Fix parallel sampling finishing/aborts (#14512)

Signed-off-by: Nick Hill <nhill@redhat.com>
2025-03-12 13:29:48 -04:00
parent 916836bbfb
commit f5d3acd474
7 changed files with 137 additions and 113 deletions
--- a/vllm/v1/engine/async_llm.py
+++ b/vllm/v1/engine/async_llm.py
@@ -298,9 +298,8 @@ class AsyncLLM(EngineClient):
    async def abort(self, request_id: str) -> None:
        """Abort RequestId in OutputProcessor and EngineCore."""

-        request_ids = [request_id]
+        request_ids = self.output_processor.abort_requests((request_id, ))
        await self.engine_core.abort_requests_async(request_ids)
-        self.output_processor.abort_requests(request_ids)

        if self.log_requests:
            logger.info("Aborted request %s.", request_id)