openstack
/
compute-hyperv
Code Issues Proposed changes

Switch to new instance failover listener 

We had a bunch of issues with the WMI based event listener, which
appear to be bugs in the WMI provider.

We've added a new listener that uses the underlying C library. The
only issue is that this new listener cannot tell us the source of
the migrated VMs, so we'll have to rely on the information from
Nova.

Closes-Bug: #1798069
Depends-On: I63d1aa3a6f9fb12d08478eb41fe973b1582b540c

Change-Id: I9e777351eca84274cafc3f04b6bc27c3f6b572ca
This commit is contained in:
Lucian Petrut 2018-10-04 12:14:13 +03:00
parent 097ef07422
commit feb4bae60f
2 changed files with 18 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
compute_hyperv/nova/cluster/clusterops.py
View File

@ -82,7 +82,7 @@ class ClusterOps(object):
def start_failover_listener_daemon(self):
"""Start the daemon failover listener."""
listener = self._clustutils.get_vm_owner_change_listener()
listener = self._clustutils.get_vm_owner_change_listener_v2()
cbk = functools.partial(utils.spawn_n, self._failover_migrate)
utils.spawn_n(listener, cbk)
@ -105,18 +105,12 @@ class ClusterOps(object):
instance.name, instance.host,
self._this_node)
def _failover_migrate(self, instance_name, old_host, new_host):
def _failover_migrate(self, instance_name, new_host):
"""This method will check if the generated event is a legitimate
failover to this node. If it is, it will proceed to prepare the
failovered VM if necessary and update the owner of the compute vm in
nova and ports in neutron.
"""
LOG.info('Checking instance failover %(instance)s to %(new_host)s '
'from host %(old_host)s.',
{'instance': instance_name,
'new_host': new_host,
'old_host': old_host})
instance = self._get_instance_by_name(instance_name)
if not instance:
# Some instances on the hypervisor may not be tracked by nova
@ -124,6 +118,13 @@ class ClusterOps(object):
instance_name)
return
old_host = instance.host
LOG.info('Checking instance failover %(instance)s to %(new_host)s '
'from host %(old_host)s.',
{'instance': instance_name,
'new_host': new_host,
'old_host': old_host})
if instance.task_state == task_states.MIGRATING:
# In case of live migration triggered by the user, we get the
# event that the instance changed host but we do not want
@ -142,9 +143,8 @@ class ClusterOps(object):
instance_name)
return
LOG.info('Instance %(instance)s failover to %(host)s.',
{'instance': instance_name,
'host': new_host})
LOG.info('Handling instance %(instance)s failover to this host.',
{'instance': instance_name})
self._nova_failover_server(instance, new_host)
self._failover_migrate_networks(instance, old_host)

15
compute_hyperv/tests/unit/cluster/test_clusterops.py
View File

@ -112,7 +112,7 @@ class ClusterOpsTestCase(test_base.HyperVBaseTestCase):
spawn_args = mock_spawn.call_args_list[0][0]
self.assertEqual(
self._clustutils.get_vm_owner_change_listener.return_value,
self._clustutils.get_vm_owner_change_listener_v2.return_value,
spawn_args[0])
cbk = spawn_args[1]
@ -149,7 +149,6 @@ class ClusterOpsTestCase(test_base.HyperVBaseTestCase):
mock_get_instance_by_name.return_value = None
self.clusterops._failover_migrate(mock.sentinel.instance_name,
mock.sentinel.old_host,
mock.sentinel.new_host)
mock_LOG.debug.assert_called_once_with(
@ -166,7 +165,6 @@ class ClusterOpsTestCase(test_base.HyperVBaseTestCase):
instance.task_state = task_states.MIGRATING
self.clusterops._failover_migrate(mock.sentinel.instance_name,
mock.sentinel.old_host,
mock.sentinel.new_host)
mock_LOG.debug.assert_called_once_with(
@ -175,11 +173,11 @@ class ClusterOpsTestCase(test_base.HyperVBaseTestCase):
@mock.patch.object(clusterops.ClusterOps, '_get_instance_by_name')
def test_failover_migrate_at_source_node(self, mock_get_instance_by_name):
instance = mock_get_instance_by_name.return_value
old_host = 'old_host'
self.clusterops._this_node = old_host
instance.host = 'old_host'
self.clusterops._this_node = instance.host
self.clusterops._failover_migrate(mock.sentinel.instance_name,
old_host, mock.sentinel.new_host)
mock.sentinel.new_host)
self.clusterops._vmops.unplug_vifs.assert_called_once_with(instance,
self.clusterops._network_api.get_instance_nw_info.return_value)
@ -191,7 +189,7 @@ class ClusterOpsTestCase(test_base.HyperVBaseTestCase):
self.clusterops._this_node = 'new_host'
self.clusterops._failover_migrate(mock.sentinel.instance_name,
None, 'host')
'host')
mock_LOG.debug.assert_called_once_with(
'Instance %s did not failover to this node.',
@ -206,10 +204,11 @@ class ClusterOpsTestCase(test_base.HyperVBaseTestCase):
instance = mock_get_instance_by_name.return_value
old_host = 'old_host'
new_host = 'new_host'
instance.host = old_host
self.clusterops._this_node = new_host
self.clusterops._failover_migrate(mock.sentinel.instance_name,
old_host, new_host)
new_host)
mock_get_instance_by_name.assert_called_once_with(
mock.sentinel.instance_name)