Setting shutdown asyncio event in a thread-safe manner (#78)

triton-inference-server · Dec 20, 2024 · 7348662 · 7348662
1 parent cd4cf06
commit 7348662
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 3 deletions.
diff --git 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
@@ -800,7 +800,7 @@ def _check_health(self, requests):
 
     def finalize(self):
         self.logger.log_info("[vllm] Issuing finalize to vllm backend")
-        self._llm_engine_shutdown_event.set()
+        self._event_loop.call_soon_threadsafe(self._llm_engine_shutdown_event.set)
 
         # Shutdown the event thread.
         if self._event_thread is not None:
-Original file line number
+Diff line change
@@ Expand Up / @@ -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
@@ Expand Down @@