From 0e6f5ef938c7556238b4d62de4b08befc82a2e5a Mon Sep 17 00:00:00 2001
From: oandreeva-nv <oandreeva@nvidia.com>
Date: Thu, 19 Dec 2024 18:00:42 -0800
Subject: [PATCH] replasing asyncio event with threading event

---
 ci/L0_backend_vllm/metrics_test/test.sh | 4 ++--
 src/model.py                            | 5 +++--
 2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/ci/L0_backend_vllm/metrics_test/test.sh b/ci/L0_backend_vllm/metrics_test/test.sh
index 884be62..a9a4db9 100755
--- a/ci/L0_backend_vllm/metrics_test/test.sh
+++ b/ci/L0_backend_vllm/metrics_test/test.sh
@@ -75,11 +75,11 @@ run_test() {
         fi
     fi
 
+    set -e
+
     # TODO: Non-graceful shutdown when metrics are enabled.
     kill $SERVER_PID
     wait $SERVER_PID
-
-    set -e
 }
 
 RET=0
diff --git a/src/model.py b/src/model.py
index 46c35a2..ad2a5c8 100644
--- a/src/model.py
+++ b/src/model.py
@@ -228,7 +228,7 @@ def _init_engine(self):
         # Run the engine in a separate thread running the AsyncIO event loop.
         self._llm_engine = None
         self._llm_engine_start_cv = threading.Condition()
-        self._llm_engine_shutdown_event = asyncio.Event()
+        self._llm_engine_shutdown_event = threading.Event()
         self._event_thread = threading.Thread(
             target=asyncio.run, args=(self._run_llm_engine(),)
         )
@@ -268,7 +268,8 @@ async def _run_llm_engine(self):
                     self._llm_engine_start_cv.notify_all()
 
                 # Wait for the engine shutdown signal.
-                await self._llm_engine_shutdown_event.wait()
+                while not self._llm_engine_shutdown_event.is_set():
+                    await asyncio.sleep(0.1)  # Prevent busy-waiting
 
                 # Wait for the ongoing requests to complete.
                 while self._ongoing_request_count > 0: