From d636c32ff94a76ae44ab36dacafcb2da32aa2ecc Mon Sep 17 00:00:00 2001 From: Canlin Guo Date: Wed, 18 Mar 2026 17:56:13 +0800 Subject: [PATCH] [Bugfix] Set PREEMPTED status when moving requests from running to waiting queue (#1893) Signed-off-by: gcanlin --- .../omni_connectors/transfer_adapter/chunk_transfer_adapter.py | 1 + 1 file changed, 1 insertion(+) diff --git a/vllm_omni/distributed/omni_connectors/transfer_adapter/chunk_transfer_adapter.py b/vllm_omni/distributed/omni_connectors/transfer_adapter/chunk_transfer_adapter.py index 8e9939d38..3139a0a0d 100644 --- a/vllm_omni/distributed/omni_connectors/transfer_adapter/chunk_transfer_adapter.py +++ b/vllm_omni/distributed/omni_connectors/transfer_adapter/chunk_transfer_adapter.py @@ -302,6 +302,7 @@ class OmniChunkTransferAdapter(OmniTransferAdapterBase): ) while len(running_queue) > self.scheduler_max_num_seqs: request = running_queue.pop() + request.status = RequestStatus.PREEMPTED waiting_queue.prepend_requests([request]) def restore_queues(self, waiting_queue: Any, running_queue: list[Request]) -> None: