A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 8, 10:00 UTC
[BUG] Reset running requests when clearing cache for pause/resume (#34382)
Signed-off-by: hao-aaron <ahao@anyscale.com>
Aaron Hao committed
Feb 12, 2026 at 08:19 UTC
7b5a8b4a9dd6eb26057e3c8e0fa07db0d89f6d54
1 file changed
+1
-1
vllm/v1/engine/async_llm.py
+1
-1
@@ -793,7 +793,7 @@ class AsyncLLM(EngineClient):
793
794
# Clear cache
795
if clear_cache:
796
- await self.reset_prefix_cache()
796
+ await self.reset_prefix_cache(reset_running_requests=True)
797
await self.reset_mm_cache()
798
await self.reset_encoder_cache()
799