Tune _get_candidates for faster scheduling in dvr

Right now we have seen some performance issues when
dvr routers are scheduled on multiple compute nodes
with thousands of VMs on the routed subnets.

The _get_candidates call get_l3_agent_candidates with
a complete list of agents irrespective of the routers
already hosted on the agents or not.

So this fix will reduce the amount of iterations that
get_l3_agent_candidates need to process for all the
agents and would increase the control plane performance.

Closes-Bug: #1513678
Change-Id: I8f781d4cbc996ce13441303c9296e4f6ec822b94

diff --git a/neutron/scheduler/l3_agent_scheduler.py b/neutron/scheduler/l3_agent_scheduler.py
index 7edc4a306d7873368dfcb66822e592fa3ba20e6b..eb6f4a1bd87da7bc7cfe701316c43d06f2818125 100644 (file)
--- a/neutron/scheduler/l3_agent_scheduler.py
+++ b/neutron/scheduler/l3_agent_scheduler.py
@@ -170,33 +170,30 @@ class L3Scheduler(object):
             # one enabled l3 agent hosting since active is just a
             # timing problem. Non-active l3 agent can return to
             # active any time
-            l3_agents = plugin.get_l3_agents_hosting_routers(
+            current_l3_agents = plugin.get_l3_agents_hosting_routers(
                 context, [sync_router['id']], admin_state_up=True)
-            if l3_agents and not sync_router.get('distributed', False):
+            is_router_distributed = sync_router.get('distributed', False)
+            if current_l3_agents and not is_router_distributed:
                 LOG.debug('Router %(router_id)s has already been hosted '
                           'by L3 agent %(agent_id)s',
                           {'router_id': sync_router['id'],
-                           'agent_id': l3_agents[0]['id']})
+                           'agent_id': current_l3_agents[0]['id']})
                 return []
 
             active_l3_agents = plugin.get_l3_agents(context, active=True)
             if not active_l3_agents:
                 LOG.warn(_LW('No active L3 agents'))
                 return []
-            new_l3agents = plugin.get_l3_agent_candidates(context,
-                                                          sync_router,
-                                                          active_l3_agents)
-            old_l3agentset = set(l3_agents)
-            if sync_router.get('distributed', False):
-                new_l3agentset = set(new_l3agents)
-                candidates = list(new_l3agentset - old_l3agentset)
-            else:
-                candidates = new_l3agents
-                if not candidates:
+            potential_candidates = list(
+                set(active_l3_agents) - set(current_l3_agents))
+            new_l3agents = []
+            if potential_candidates:
+                new_l3agents = plugin.get_l3_agent_candidates(
+                    context, sync_router, potential_candidates)
+                if not new_l3agents:
                     LOG.warn(_LW('No L3 agents can host the router %s'),
                              sync_router['id'])
-
-            return candidates
+            return new_l3agents
 
     def _bind_routers(self, context, plugin, routers, l3_agent):
         for router in routers:

diff --git a/neutron/tests/unit/scheduler/test_l3_agent_scheduler.py b/neutron/tests/unit/scheduler/test_l3_agent_scheduler.py
index aca6bc8b0da5bca192a12d7bfdba0ef64c9dada9..3cc9588d85e7a2a9d16d9e94976782d93453c9b9 100644 (file)
--- a/neutron/tests/unit/scheduler/test_l3_agent_scheduler.py
+++ b/neutron/tests/unit/scheduler/test_l3_agent_scheduler.py
@@ -282,6 +282,33 @@ class L3SchedulerBaseTestCase(base.BaseTestCase):
         router['distributed'] = True
         plugin.get_l3_agents.return_value = []
         iter(self.scheduler._get_candidates(plugin, mock.MagicMock(), router))
+        self.assertFalse(plugin.get_l3_agent_candidates.called)
+
+    def test__get_candidates_skips_get_l3_agent_candidates_if_dvr_scheduled(
+            self):
+        plugin = mock.MagicMock()
+        # distributed router already hosted
+        plugin.get_l3_agents_hosting_routers.return_value = ['a1']
+        router = {'distributed': True, 'id': str(uuid.uuid4())}
+        plugin.get_l3_agents.return_value = ['a1']
+        self.scheduler._get_candidates(plugin, mock.MagicMock(), router)
+        self.assertFalse(plugin.get_l3_agent_candidates.called)
+
+    def test__get_candidates_calls_get_l3_agent_candidates_if_agent_available(
+            self):
+        plugin = mock.MagicMock()
+        # distributed router already hosted in two agent 'a1' and 'a2'
+        plugin.get_l3_agents_hosting_routers.return_value = ['a1', 'a2']
+        router = {'distributed': True, 'id': str(uuid.uuid4())}
+        # Available distributed agents
+        plugin.get_l3_agents.return_value = ['a1', 'a2', 'a3', 'a4', 'a5']
+        unscheduled_agents = ['a3', 'a4', 'a5']
+        plugin.get_l3_agent_candidates.return_value = ['a3', 'a4']
+        agents_returned = self.scheduler._get_candidates(
+            plugin, mock.MagicMock(), router)
+        plugin.get_l3_agent_candidates.called_once_with(
+            mock.ANY, router, unscheduled_agents)
+        self.assertEqual(['a3', 'a4'], sorted(agents_returned))
 
 
 class L3SchedulerBaseMixin(object):