FIX test_no_crash_max_workers

CHANGES.md

@@ -10,6 +10,12 @@
   attempt to use more than 61 workers on that platform (or 60 depending on the
   Python version). (#390).
 
+- Fix loky compat with python 3.11 for nested calls. (#394).
+
+- Adapt the cooldown strategy when shutingdown an executor with full
+  ``call_queue``. This should accelerate the time taken to shutdown
+  in general, in particular on overloaded machines. (#399).
+
 ### 3.3.0 - 2022-09-15
 
 - Fix worker management logic in `get_reusable_executor` to ensure

loky/process_executor.py

@@ -867,17 +867,19 @@ def shutdown_workers(self):
                     self.call_queue.put_nowait(None)
                     n_sentinels_sent += 1
                 except queue.Full as e:
-                    if cooldown_time > 10.0:
+                    if cooldown_time > 5.0:
                         mp.util.info(
                             "failed to send all sentinels and exit with error."
+                            f"\ncall_queue size={self.call_queue._maxsize}; "
+                            f" full is {self.call_queue.full()}; "
                         )
                         raise e
                     mp.util.info(
                         "full call_queue prevented to send all sentinels at "
                         "once, waiting..."
                     )
                     sleep(cooldown_time)
-                    cooldown_time *= 2
+                    cooldown_time *= 1.2
                     break
 
         mp.util.debug(f"sent {n_sentinels_sent} sentinels to the call queue")

tests/test_reusable_executor.py

@@ -1020,13 +1020,19 @@ def test_no_crash_max_workers_on_windows():
     # user asks for more workers than the maximum number of workers supported
     # by the platform.
 
+    # To make sure we have the proper size for the call_queue, we
+    # shutdown the existing executor.
+    # See https://github.com/joblib/loky#396
+    get_reusable_executor().shutdown()
+
     # Note: on overloaded CI hosts, spawning many processes can take a long
     # time. We need to increase the timeout to avoid spurious failures when
     # making assertions on `len(executor._processes)`.
-    idle_worker_timeout = 10 * 60
+    idle_worker_timeout = 5 * 60
     with warnings.catch_warnings(record=True) as record:
         executor = get_reusable_executor(
-            max_workers=_MAX_WINDOWS_WORKERS + 1, timeout=idle_worker_timeout
+            max_workers=_MAX_WINDOWS_WORKERS + 1,
+            timeout=idle_worker_timeout,
         )
         assert executor.submit(lambda: None).result() is None
     if sys.platform == "win32":