[producer] Do not hard stop periodic task during exception

Use add_timer_args with stop_on_exception=False to avoid hard stop of
periodic task during exception.

Closes-Bug: #2064574

Change-Id: If5cb6ddcac114addc99f0e025d517b010a3ec739

designate/coordination.py

@@ -120,9 +120,10 @@ class Coordination:
 
     def _enable_grouping(self):
         self._create_group()
-        self.tg.add_timer(
+        self.tg.add_timer_args(
             CONF.coordination.run_watchers_interval,
-            self._coordinator_run_watchers
+            self._coordinator_run_watchers,
+            stop_on_exception=False,
         )
 
 

designate/producer/service.py

@@ -88,7 +88,7 @@ class Service(service.RPCService):
             self._partitioner.watch_partition_change(task.on_partition_change)
 
             interval = CONF[task.get_canonical_name()].interval
-            self.tg.add_timer(interval, task)
+            self.tg.add_timer_args(interval, task, stop_on_exception=False)
 
     def stop(self, graceful=True):
         super().stop(graceful)

designate/tests/unit/producer/test_service.py

@@ -73,8 +73,8 @@ class ProducerServiceTest(oslotest.base.BaseTestCase):
         mock_partition.start.assert_called()
 
         # Make sure that tasks were added to the tg timer.
-        self.tg.add_timer.assert_called()
-        self.assertEqual(6, self.tg.add_timer.call_count)
+        self.tg.add_timer_args.assert_called()
+        self.assertEqual(6, self.tg.add_timer_args.call_count)
 
     @mock.patch.object(service.coordination, 'Partitioner')
     @mock.patch.object(designate.service.RPCService, 'start')