Make OVS controller inactivity_probe configurable

This parameter applies to the OVSDB Controller table when the
native openflow driver is used. There are reports that increasing
it can reduce errors on busy systems. This patch also sets the
default value to 10s which is more than the OVS default of 5s.
See the ovs-vswitchd.conf.db man page for full description.

Change-Id: If0d42919412dac75deb4d7f484c42cea630fbc59
Partial-Bug: #1817022

neutron/agent/common/ovs_lib.py

@@ -687,6 +687,13 @@ class OVSBridge(BaseOVS):
         """
         self.set_controller_field('connection_mode', connection_mode)
 
+    def set_controllers_inactivity_probe(self, interval):
+        """Set bridge controllers inactivity probe interval.
+
+        :param interval: inactivity_probe value in seconds.
+        """
+        self.set_controller_field('inactivity_probe', interval * 1000)
+
     def _set_egress_bw_limit_for_port(self, port_name, max_kbps,
                                       max_burst_kbps):
         with self.ovsdb.transaction(check_error=True) as txn:

neutron/conf/plugins/ml2/drivers/ovs_conf.py

@@ -121,6 +121,11 @@ ovs_opts = [
                help=_("Timeout in seconds to wait for a single "
                       "OpenFlow request. "
                       "Used only for 'native' driver.")),
+    cfg.IntOpt('of_inactivity_probe', default=10,
+               help=_("The inactivity_probe interval in seconds for the local "
+                      "switch connection to the controller. "
+                      "A value of 0 disables inactivity probes. "
+                      "Used only for 'native' driver.")),
 ]
 
 agent_opts = [

neutron/plugins/ml2/drivers/openvswitch/agent/openflow/native/ovs_bridge.py

@@ -94,6 +94,7 @@ class OVSAgentBridge(ofswitch.OpenFlowSwitchMixin,
         #
         # [1] https://github.com/openvswitch/ovs/blob/master/DESIGN.md
         self.set_controllers_connection_mode("out-of-band")
+        self.set_controllers_inactivity_probe(conf.OVS.of_inactivity_probe)
 
     def drop_port(self, in_port):
         self.install_drop(priority=2, in_port=in_port)

neutron/tests/functional/agent/common/test_ovs_lib.py

@@ -429,3 +429,11 @@ class BaseOVSTestCase(base.BaseSudoTestCase):
         self.ovs.update_minimum_bandwidth_queue(self.port_id, [], 1, 2800)
         self._check_value(2800, self.ovs.get_egress_min_bw_for_port,
                           port_id=self.port_id)
+
+    def test_set_controllers_inactivity_probe(self):
+        self._create_bridge()
+        self.ovs.set_controller(['tcp:127.0.0.1:6633'])
+        self.ovs.set_controllers_inactivity_probe(8)
+        self.assertEqual(8000,
+                         self.ovs.db_get_val('Controller', self.br_name,
+                                             'inactivity_probe'))

neutron/tests/unit/plugins/ml2/drivers/openvswitch/agent/openflow/native/ovs_bridge_test_base.py

@@ -142,12 +142,15 @@ class OVSBridgeTestBase(ovs_test_base.OVSOSKenTestBase):
 
         m_add_protocols = mock.patch.object(self.br, 'add_protocols')
         m_set_controller = mock.patch.object(self.br, 'set_controller')
+        m_set_probe = mock.patch.object(self.br,
+                                        'set_controllers_inactivity_probe')
         m_set_ccm = mock.patch.object(self.br,
                                       'set_controllers_connection_mode')
 
-        with m_set_ccm as set_ccm, m_set_controller, m_add_protocols:
-            self.br.setup_controllers(cfg)
-            set_ccm.assert_called_once_with("out-of-band")
+        with m_set_ccm as set_ccm:
+            with m_set_controller, m_add_protocols, m_set_probe:
+                self.br.setup_controllers(cfg)
+                set_ccm.assert_called_once_with("out-of-band")
 
 
 class OVSDVRProcessTestMixin(object):

releasenotes/notes/ovs-make-inactivity-probe-configurable-39d669014d961c5c.yaml

@@ -0,0 +1,7 @@
+---
+other:
+  - |
+    A new option ``[ovs] of_inactivity_probe`` has been added to allow
+    changing the inactivity probe interval when using the OVS ML2 agent
+    with the native OpenFlow driver. Operators can increase this if they
+    are experiencing OpenFlow timeouts. The default value is 10 seconds.