Wait for network-vif-plugged before starting live migration
This adds a new config option which is read on the destination host during pre_live_migration and the value is returned back to the source host, which can be used to determine, from the source host, if it should wait for a "network-vif-plugged" event due to VIFs being plugged on the destination host. This helps us to avoid the guest transfer at all if vif plugging failed on the dest host, which we just wouldn't find out until post live migration and then we have to rollback. The option is disabled by default for backward compatibility and also because certain networking backends, like OpenDaylight, are known to not send network-vif-plugged events unless the port host binding information changes, which for live migration doesn't happen until after the guest is transferred to the destination host. We could arguably avoid the changes to the live migrate data versioned object and just assume the same networking backend is used within each cell, but this does allow the deployer to have the flexibility of live migrating between different network backends (eventually anyway). The ability to live migrate between different VIF types is being worked on as part of blueprint neutron-new-port-binding-api. Related to blueprint neutron-new-port-binding-api Change-Id: I0f3ab6604d8b79bdb75cf67571e359cfecc039d8
This commit is contained in:
Matt Riedemann
parent
ecaadf6d6d
commit
5aadff75c3
@ -6042,6 +6042,13 @@ class ComputeManager(manager.Manager):
|
||||
disk,
|
||||
migrate_data)
|
||||
LOG.debug('driver pre_live_migration data is %s', migrate_data)
|
||||
# driver.pre_live_migration is what plugs vifs on the destination host
|
||||
# so now we can set the wait_for_vif_plugged flag in the migrate_data
|
||||
# object which the source compute will use to determine if it should
|
||||
# wait for a 'network-vif-plugged' event from neutron before starting
|
||||
# the actual guest transfer in the hypervisor
|
||||
migrate_data.wait_for_vif_plugged = (
|
||||
CONF.compute.live_migration_wait_for_vif_plug)
|
||||
|
||||
# Volume connections are complete, tell cinder that all the
|
||||
# attachments have completed.
|
||||
@ -6074,6 +6081,51 @@ class ComputeManager(manager.Manager):
|
||||
LOG.debug('pre_live_migration result data is %s', migrate_data)
|
||||
return migrate_data
|
||||
|
||||
@staticmethod
|
||||
def _neutron_failed_live_migration_callback(event_name, instance):
|
||||
msg = ('Neutron reported failure during live migration '
|
||||
'with %(event)s for instance %(uuid)s')
|
||||
msg_args = {'event': event_name, 'uuid': instance.uuid}
|
||||
if CONF.vif_plugging_is_fatal:
|
||||
raise exception.VirtualInterfacePlugException(msg % msg_args)
|
||||
LOG.error(msg, msg_args)
|
||||
|
||||
@staticmethod
|
||||
def _get_neutron_events_for_live_migration(instance):
|
||||
# We don't generate events if CONF.vif_plugging_timeout=0
|
||||
# meaning that the operator disabled using them.
|
||||
if CONF.vif_plugging_timeout and utils.is_neutron():
|
||||
return [('network-vif-plugged', vif['id'])
|
||||
for vif in instance.get_network_info()]
|
||||
else:
|
||||
return []
|
||||
|
||||
def _cleanup_pre_live_migration(self, context, dest, instance,
|
||||
migration, migrate_data):
|
||||
"""Helper method for when pre_live_migration fails
|
||||
|
||||
Sets the migration status to "error" and rolls back the live migration
|
||||
setup on the destination host.
|
||||
|
||||
:param context: The user request context.
|
||||
:type context: nova.context.RequestContext
|
||||
:param dest: The live migration destination hostname.
|
||||
:type dest: str
|
||||
:param instance: The instance being live migrated.
|
||||
:type instance: nova.objects.Instance
|
||||
:param migration: The migration record tracking this live migration.
|
||||
:type migration: nova.objects.Migration
|
||||
:param migrate_data: Data about the live migration, populated from
|
||||
the destination host.
|
||||
:type migrate_data: Subclass of nova.objects.LiveMigrateData
|
||||
"""
|
||||
self._set_migration_status(migration, 'error')
|
||||
# Make sure we set this for _rollback_live_migration()
|
||||
# so it can find it, as expected if it was called later
|
||||
migrate_data.migration = migration
|
||||
self._rollback_live_migration(context, instance, dest,
|
||||
migrate_data)
|
||||
|
||||
def _do_live_migration(self, context, dest, instance, block_migration,
|
||||
migration, migrate_data):
|
||||
# NOTE(danms): We should enhance the RT to account for migrations
|
||||
@ -6082,6 +6134,15 @@ class ComputeManager(manager.Manager):
|
||||
# reporting
|
||||
self._set_migration_status(migration, 'preparing')
|
||||
|
||||
class _BreakWaitForInstanceEvent(Exception):
|
||||
"""Used as a signal to stop waiting for the network-vif-plugged
|
||||
event when we discover that
|
||||
[compute]/live_migration_wait_for_vif_plug is not set on the
|
||||
destination.
|
||||
"""
|
||||
pass
|
||||
|
||||
events = self._get_neutron_events_for_live_migration(instance)
|
||||
try:
|
||||
if ('block_migration' in migrate_data and
|
||||
migrate_data.block_migration):
|
||||
@ -6092,19 +6153,54 @@ class ComputeManager(manager.Manager):
|
||||
else:
|
||||
disk = None
|
||||
|
||||
migrate_data = self.compute_rpcapi.pre_live_migration(
|
||||
context, instance,
|
||||
block_migration, disk, dest, migrate_data)
|
||||
deadline = CONF.vif_plugging_timeout
|
||||
error_cb = self._neutron_failed_live_migration_callback
|
||||
# In order to avoid a race with the vif plugging that the virt
|
||||
# driver does on the destination host, we register our events
|
||||
# to wait for before calling pre_live_migration. Then if the
|
||||
# dest host reports back that we shouldn't wait, we can break
|
||||
# out of the context manager using _BreakWaitForInstanceEvent.
|
||||
with self.virtapi.wait_for_instance_event(
|
||||
instance, events, deadline=deadline,
|
||||
error_callback=error_cb):
|
||||
migrate_data = self.compute_rpcapi.pre_live_migration(
|
||||
context, instance,
|
||||
block_migration, disk, dest, migrate_data)
|
||||
wait_for_vif_plugged = (
|
||||
'wait_for_vif_plugged' in migrate_data and
|
||||
migrate_data.wait_for_vif_plugged)
|
||||
if events and not wait_for_vif_plugged:
|
||||
raise _BreakWaitForInstanceEvent
|
||||
except _BreakWaitForInstanceEvent:
|
||||
if events:
|
||||
LOG.debug('Not waiting for events after pre_live_migration: '
|
||||
'%s. ', events, instance=instance)
|
||||
# This is a bit weird, but we need to clear sys.exc_info() so that
|
||||
# oslo.log formatting does not inadvertently use it later if an
|
||||
# error message is logged without an explicit exc_info. This is
|
||||
# only a problem with python 2.
|
||||
if six.PY2:
|
||||
sys.exc_clear()
|
||||
except exception.VirtualInterfacePlugException:
|
||||
with excutils.save_and_reraise_exception():
|
||||
LOG.exception('Failed waiting for network virtual interfaces '
|
||||
'to be plugged on the destination host %s.',
|
||||
dest, instance=instance)
|
||||
self._cleanup_pre_live_migration(
|
||||
context, dest, instance, migration, migrate_data)
|
||||
except eventlet.timeout.Timeout:
|
||||
msg = 'Timed out waiting for events: %s'
|
||||
LOG.warning(msg, events, instance=instance)
|
||||
if CONF.vif_plugging_is_fatal:
|
||||
self._cleanup_pre_live_migration(
|
||||
context, dest, instance, migration, migrate_data)
|
||||
raise exception.MigrationError(reason=msg % events)
|
||||
except Exception:
|
||||
with excutils.save_and_reraise_exception():
|
||||
LOG.exception('Pre live migration failed at %s',
|
||||
dest, instance=instance)
|
||||
self._set_migration_status(migration, 'error')
|
||||
# Make sure we set this for _rollback_live_migration()
|
||||
# so it can find it, as expected if it was called later
|
||||
migrate_data.migration = migration
|
||||
self._rollback_live_migration(context, instance, dest,
|
||||
migrate_data)
|
||||
self._cleanup_pre_live_migration(
|
||||
context, dest, instance, migration, migrate_data)
|
||||
|
||||
self._set_migration_status(migration, 'running')
|
||||
|
||||
|
||||
@ -670,7 +670,57 @@ hw:emulator_threads_policy:share.
|
||||
|
||||
::
|
||||
cpu_shared_set = "4-12,^8,15"
|
||||
""")
|
||||
"""),
|
||||
cfg.BoolOpt('live_migration_wait_for_vif_plug',
|
||||
# TODO(mriedem): Change to default=True starting in Stein.
|
||||
default=False,
|
||||
help="""
|
||||
Determine if the source compute host should wait for a ``network-vif-plugged``
|
||||
event from the (neutron) networking service before starting the actual transfer
|
||||
of the guest to the destination compute host.
|
||||
|
||||
Note that this option is read on the destination host of a live migration.
|
||||
If you set this option the same on all of your compute hosts, which you should
|
||||
do if you use the same networking backend universally, you do not have to
|
||||
worry about this.
|
||||
|
||||
Before starting the transfer of the guest, some setup occurs on the destination
|
||||
compute host, including plugging virtual interfaces. Depending on the
|
||||
networking backend **on the destination host**, a ``network-vif-plugged``
|
||||
event may be triggered and then received on the source compute host and the
|
||||
source compute can wait for that event to ensure networking is set up on the
|
||||
destination host before starting the guest transfer in the hypervisor.
|
||||
|
||||
By default, this is False for two reasons:
|
||||
|
||||
1. Backward compatibility: deployments should test this out and ensure it works
|
||||
for them before enabling it.
|
||||
|
||||
2. The compute service cannot reliably determine which types of virtual
|
||||
interfaces (``port.binding:vif_type``) will send ``network-vif-plugged``
|
||||
events without an accompanying port ``binding:host_id`` change.
|
||||
Open vSwitch and linuxbridge should be OK, but OpenDaylight is at least
|
||||
one known backend that will not currently work in this case, see bug
|
||||
https://launchpad.net/bugs/1755890 for more details.
|
||||
|
||||
Possible values:
|
||||
|
||||
* True: wait for ``network-vif-plugged`` events before starting guest transfer
|
||||
* False: do not wait for ``network-vif-plugged`` events before starting guest
|
||||
transfer (this is how things have always worked before this option
|
||||
was introduced)
|
||||
|
||||
Related options:
|
||||
|
||||
* [DEFAULT]/vif_plugging_is_fatal: if ``live_migration_wait_for_vif_plug`` is
|
||||
True and ``vif_plugging_timeout`` is greater than 0, and a timeout is
|
||||
reached, the live migration process will fail with an error but the guest
|
||||
transfer will not have started to the destination host
|
||||
* [DEFAULT]/vif_plugging_timeout: if ``live_migration_wait_for_vif_plug`` is
|
||||
True, this controls the amount of time to wait before timing out and either
|
||||
failing if ``vif_plugging_is_fatal`` is True, or simply continuing with the
|
||||
live migration
|
||||
"""),
|
||||
]
|
||||
|
||||
interval_opts = [
|
||||
|
||||
@ -33,6 +33,11 @@ class LiveMigrateData(obj_base.NovaObject):
|
||||
# for each volume so they can be restored on a migration rollback. The
|
||||
# key is the volume_id, and the value is the attachment_id.
|
||||
'old_vol_attachment_ids': fields.DictOfStringsField(),
|
||||
# wait_for_vif_plugged is set in pre_live_migration on the destination
|
||||
# compute host based on the [compute]/live_migration_wait_for_vif_plug
|
||||
# config option value; a default value is not set here since the
|
||||
# default for the config option may change in the future
|
||||
'wait_for_vif_plugged': fields.BooleanField()
|
||||
}
|
||||
|
||||
def to_legacy_dict(self, pre_migration_result=False):
|
||||
@ -127,7 +132,8 @@ class LibvirtLiveMigrateData(LiveMigrateData):
|
||||
# Version 1.3: Added 'supported_perf_events'
|
||||
# Version 1.4: Added old_vol_attachment_ids
|
||||
# Version 1.5: Added src_supports_native_luks
|
||||
VERSION = '1.5'
|
||||
# Version 1.6: Added wait_for_vif_plugged
|
||||
VERSION = '1.6'
|
||||
|
||||
fields = {
|
||||
'filename': fields.StringField(),
|
||||
@ -153,6 +159,8 @@ class LibvirtLiveMigrateData(LiveMigrateData):
|
||||
super(LibvirtLiveMigrateData, self).obj_make_compatible(
|
||||
primitive, target_version)
|
||||
target_version = versionutils.convert_version_to_tuple(target_version)
|
||||
if target_version < (1, 6) and 'wait_for_vif_plugged' in primitive:
|
||||
del primitive['wait_for_vif_plugged']
|
||||
if target_version < (1, 5):
|
||||
if 'src_supports_native_luks' in primitive:
|
||||
del primitive['src_supports_native_luks']
|
||||
@ -248,7 +256,8 @@ class XenapiLiveMigrateData(LiveMigrateData):
|
||||
# Version 1.0: Initial version
|
||||
# Version 1.1: Added vif_uuid_map
|
||||
# Version 1.2: Added old_vol_attachment_ids
|
||||
VERSION = '1.2'
|
||||
# Version 1.3: Added wait_for_vif_plugged
|
||||
VERSION = '1.3'
|
||||
|
||||
fields = {
|
||||
'block_migration': fields.BooleanField(nullable=True),
|
||||
@ -300,6 +309,8 @@ class XenapiLiveMigrateData(LiveMigrateData):
|
||||
super(XenapiLiveMigrateData, self).obj_make_compatible(
|
||||
primitive, target_version)
|
||||
target_version = versionutils.convert_version_to_tuple(target_version)
|
||||
if target_version < (1, 3) and 'wait_for_vif_plugged' in primitive:
|
||||
del primitive['wait_for_vif_plugged']
|
||||
if target_version < (1, 2):
|
||||
if 'old_vol_attachment_ids' in primitive:
|
||||
del primitive['old_vol_attachment_ids']
|
||||
@ -313,7 +324,8 @@ class HyperVLiveMigrateData(LiveMigrateData):
|
||||
# Version 1.0: Initial version
|
||||
# Version 1.1: Added is_shared_instance_path
|
||||
# Version 1.2: Added old_vol_attachment_ids
|
||||
VERSION = '1.2'
|
||||
# Version 1.3: Added wait_for_vif_plugged
|
||||
VERSION = '1.3'
|
||||
|
||||
fields = {'is_shared_instance_path': fields.BooleanField()}
|
||||
|
||||
@ -321,6 +333,8 @@ class HyperVLiveMigrateData(LiveMigrateData):
|
||||
super(HyperVLiveMigrateData, self).obj_make_compatible(
|
||||
primitive, target_version)
|
||||
target_version = versionutils.convert_version_to_tuple(target_version)
|
||||
if target_version < (1, 3) and 'wait_for_vif_plugged' in primitive:
|
||||
del primitive['wait_for_vif_plugged']
|
||||
if target_version < (1, 2):
|
||||
if 'old_vol_attachment_ids' in primitive:
|
||||
del primitive['old_vol_attachment_ids']
|
||||
@ -346,7 +360,8 @@ class PowerVMLiveMigrateData(LiveMigrateData):
|
||||
# Version 1.0: Initial version
|
||||
# Version 1.1: Added the Virtual Ethernet Adapter VLAN mappings.
|
||||
# Version 1.2: Added old_vol_attachment_ids
|
||||
VERSION = '1.2'
|
||||
# Version 1.3: Added wait_for_vif_plugged
|
||||
VERSION = '1.3'
|
||||
|
||||
fields = {
|
||||
'host_mig_data': fields.DictOfNullableStringsField(),
|
||||
@ -363,6 +378,8 @@ class PowerVMLiveMigrateData(LiveMigrateData):
|
||||
super(PowerVMLiveMigrateData, self).obj_make_compatible(
|
||||
primitive, target_version)
|
||||
target_version = versionutils.convert_version_to_tuple(target_version)
|
||||
if target_version < (1, 3) and 'wait_for_vif_plugged' in primitive:
|
||||
del primitive['wait_for_vif_plugged']
|
||||
if target_version < (1, 2):
|
||||
if 'old_vol_attachment_ids' in primitive:
|
||||
del primitive['old_vol_attachment_ids']
|
||||
|
||||
@ -6136,15 +6136,17 @@ class ComputeTestCase(BaseTestCase,
|
||||
fake_notifier.NOTIFICATIONS = []
|
||||
migrate_data = objects.LibvirtLiveMigrateData(
|
||||
is_shared_instance_path=False)
|
||||
mock_pre.return_value = None
|
||||
mock_pre.return_value = migrate_data
|
||||
|
||||
with mock.patch.object(self.compute.network_api,
|
||||
'setup_networks_on_host') as mock_setup:
|
||||
self.flags(live_migration_wait_for_vif_plug=True, group='compute')
|
||||
ret = self.compute.pre_live_migration(c, instance=instance,
|
||||
block_migration=False,
|
||||
disk=None,
|
||||
migrate_data=migrate_data)
|
||||
self.assertIsNone(ret)
|
||||
self.assertIs(migrate_data, ret)
|
||||
self.assertTrue(ret.wait_for_vif_plugged, ret)
|
||||
self.assertEqual(len(fake_notifier.NOTIFICATIONS), 2)
|
||||
msg = fake_notifier.NOTIFICATIONS[0]
|
||||
self.assertEqual(msg.event_type,
|
||||
@ -6191,7 +6193,9 @@ class ComputeTestCase(BaseTestCase,
|
||||
|
||||
instance = self._create_fake_instance_obj(
|
||||
{'host': 'src_host',
|
||||
'task_state': task_states.MIGRATING})
|
||||
'task_state': task_states.MIGRATING,
|
||||
'info_cache': objects.InstanceInfoCache(
|
||||
network_info=network_model.NetworkInfo([]))})
|
||||
updated_instance = self._create_fake_instance_obj(
|
||||
{'host': 'fake-dest-host'})
|
||||
dest_host = updated_instance['host']
|
||||
@ -6276,7 +6280,9 @@ class ComputeTestCase(BaseTestCase,
|
||||
# Confirm live_migration() works as expected correctly.
|
||||
# creating instance testdata
|
||||
c = context.get_admin_context()
|
||||
instance = self._create_fake_instance_obj(context=c)
|
||||
params = {'info_cache': objects.InstanceInfoCache(
|
||||
network_info=network_model.NetworkInfo([]))}
|
||||
instance = self._create_fake_instance_obj(params=params, context=c)
|
||||
instance.host = self.compute.host
|
||||
dest = 'desthost'
|
||||
|
||||
@ -6330,7 +6336,9 @@ class ComputeTestCase(BaseTestCase,
|
||||
# Confirm live_migration() works as expected correctly.
|
||||
# creating instance testdata
|
||||
c = context.get_admin_context()
|
||||
instance = self._create_fake_instance_obj(context=c)
|
||||
params = {'info_cache': objects.InstanceInfoCache(
|
||||
network_info=network_model.NetworkInfo([]))}
|
||||
instance = self._create_fake_instance_obj(params=params, context=c)
|
||||
instance.host = self.compute.host
|
||||
dest = 'desthost'
|
||||
|
||||
|
||||
@ -20,6 +20,7 @@ from cinderclient import exceptions as cinder_exception
|
||||
from cursive import exception as cursive_exception
|
||||
import ddt
|
||||
from eventlet import event as eventlet_event
|
||||
from eventlet import timeout as eventlet_timeout
|
||||
import mock
|
||||
import netaddr
|
||||
from oslo_log import log as logging
|
||||
@ -7046,6 +7047,159 @@ class ComputeManagerMigrationTestCase(test.NoDBTestCase):
|
||||
new_attachment_id)
|
||||
_test()
|
||||
|
||||
def test_get_neutron_events_for_live_migration_empty(self):
|
||||
"""Tests the various ways that _get_neutron_events_for_live_migration
|
||||
will return an empty list.
|
||||
"""
|
||||
nw = network_model.NetworkInfo([network_model.VIF(uuids.port1)])
|
||||
# 1. no timeout
|
||||
self.flags(vif_plugging_timeout=0)
|
||||
self.assertEqual(
|
||||
[], self.compute._get_neutron_events_for_live_migration(nw))
|
||||
# 2. not neutron
|
||||
self.flags(vif_plugging_timeout=300, use_neutron=False)
|
||||
self.assertEqual(
|
||||
[], self.compute._get_neutron_events_for_live_migration(nw))
|
||||
# 3. no VIFs
|
||||
self.flags(vif_plugging_timeout=300, use_neutron=True)
|
||||
self.assertEqual(
|
||||
[], self.compute._get_neutron_events_for_live_migration([]))
|
||||
|
||||
@mock.patch('nova.compute.rpcapi.ComputeAPI.pre_live_migration')
|
||||
@mock.patch('nova.compute.manager.ComputeManager._post_live_migration')
|
||||
def test_live_migration_wait_vif_plugged(
|
||||
self, mock_post_live_mig, mock_pre_live_mig):
|
||||
"""Tests the happy path of waiting for network-vif-plugged events from
|
||||
neutron when pre_live_migration returns a migrate_data object with
|
||||
wait_for_vif_plugged=True.
|
||||
"""
|
||||
migrate_data = objects.LibvirtLiveMigrateData(
|
||||
wait_for_vif_plugged=True)
|
||||
mock_pre_live_mig.return_value = migrate_data
|
||||
self.instance.info_cache = objects.InstanceInfoCache(
|
||||
network_info=network_model.NetworkInfo([
|
||||
network_model.VIF(uuids.port1), network_model.VIF(uuids.port2)
|
||||
]))
|
||||
with mock.patch.object(self.compute.virtapi,
|
||||
'wait_for_instance_event') as wait_for_event:
|
||||
self.compute._do_live_migration(
|
||||
self.context, 'dest-host', self.instance, None, self.migration,
|
||||
migrate_data)
|
||||
self.assertEqual(2, len(wait_for_event.call_args[0][1]))
|
||||
self.assertEqual(CONF.vif_plugging_timeout,
|
||||
wait_for_event.call_args[1]['deadline'])
|
||||
mock_pre_live_mig.assert_called_once_with(
|
||||
self.context, self.instance, None, None, 'dest-host',
|
||||
migrate_data)
|
||||
|
||||
@mock.patch('nova.compute.rpcapi.ComputeAPI.pre_live_migration')
|
||||
@mock.patch('nova.compute.manager.ComputeManager._post_live_migration')
|
||||
@mock.patch('nova.compute.manager.LOG.debug')
|
||||
def test_live_migration_wait_vif_plugged_old_dest_host(
|
||||
self, mock_log_debug, mock_post_live_mig, mock_pre_live_mig):
|
||||
"""Tests the scenario that the destination compute returns a
|
||||
migrate_data with no wait_for_vif_plugged set because the dest compute
|
||||
doesn't have that code yet. In this case, we default to legacy behavior
|
||||
of not waiting.
|
||||
"""
|
||||
migrate_data = objects.LibvirtLiveMigrateData()
|
||||
mock_pre_live_mig.return_value = migrate_data
|
||||
self.instance.info_cache = objects.InstanceInfoCache(
|
||||
network_info=network_model.NetworkInfo([
|
||||
network_model.VIF(uuids.port1)]))
|
||||
with mock.patch.object(
|
||||
self.compute.virtapi, 'wait_for_instance_event'):
|
||||
self.compute._do_live_migration(
|
||||
self.context, 'dest-host', self.instance, None, self.migration,
|
||||
migrate_data)
|
||||
# This isn't awesome, but we need a way to assert that we
|
||||
# short-circuit'ed the wait_for_instance_event context manager.
|
||||
self.assertEqual(2, mock_log_debug.call_count)
|
||||
self.assertIn('Not waiting for events after pre_live_migration',
|
||||
mock_log_debug.call_args_list[0][0][0]) # first call/arg
|
||||
|
||||
@mock.patch('nova.compute.rpcapi.ComputeAPI.pre_live_migration')
|
||||
@mock.patch('nova.compute.manager.ComputeManager._rollback_live_migration')
|
||||
def test_live_migration_wait_vif_plugged_vif_plug_error(
|
||||
self, mock_rollback_live_mig, mock_pre_live_mig):
|
||||
"""Tests the scenario where wait_for_instance_event fails with
|
||||
VirtualInterfacePlugException.
|
||||
"""
|
||||
migrate_data = objects.LibvirtLiveMigrateData(
|
||||
wait_for_vif_plugged=True)
|
||||
mock_pre_live_mig.return_value = migrate_data
|
||||
self.instance.info_cache = objects.InstanceInfoCache(
|
||||
network_info=network_model.NetworkInfo([
|
||||
network_model.VIF(uuids.port1)]))
|
||||
with mock.patch.object(
|
||||
self.compute.virtapi,
|
||||
'wait_for_instance_event') as wait_for_event:
|
||||
wait_for_event.return_value.__enter__.side_effect = (
|
||||
exception.VirtualInterfacePlugException())
|
||||
self.assertRaises(
|
||||
exception.VirtualInterfacePlugException,
|
||||
self.compute._do_live_migration, self.context, 'dest-host',
|
||||
self.instance, None, self.migration, migrate_data)
|
||||
self.assertEqual('error', self.migration.status)
|
||||
mock_rollback_live_mig.assert_called_once_with(
|
||||
self.context, self.instance, 'dest-host', migrate_data)
|
||||
|
||||
@mock.patch('nova.compute.rpcapi.ComputeAPI.pre_live_migration')
|
||||
@mock.patch('nova.compute.manager.ComputeManager._rollback_live_migration')
|
||||
def test_live_migration_wait_vif_plugged_timeout_error(
|
||||
self, mock_rollback_live_mig, mock_pre_live_mig):
|
||||
"""Tests the scenario where wait_for_instance_event raises an
|
||||
eventlet Timeout exception and we're configured such that vif plugging
|
||||
failures are fatal (which is the default).
|
||||
"""
|
||||
migrate_data = objects.LibvirtLiveMigrateData(
|
||||
wait_for_vif_plugged=True)
|
||||
mock_pre_live_mig.return_value = migrate_data
|
||||
self.instance.info_cache = objects.InstanceInfoCache(
|
||||
network_info=network_model.NetworkInfo([
|
||||
network_model.VIF(uuids.port1)]))
|
||||
with mock.patch.object(
|
||||
self.compute.virtapi,
|
||||
'wait_for_instance_event') as wait_for_event:
|
||||
wait_for_event.return_value.__enter__.side_effect = (
|
||||
eventlet_timeout.Timeout())
|
||||
ex = self.assertRaises(
|
||||
exception.MigrationError, self.compute._do_live_migration,
|
||||
self.context, 'dest-host', self.instance, None,
|
||||
self.migration, migrate_data)
|
||||
self.assertIn('Timed out waiting for events', six.text_type(ex))
|
||||
self.assertEqual('error', self.migration.status)
|
||||
mock_rollback_live_mig.assert_called_once_with(
|
||||
self.context, self.instance, 'dest-host', migrate_data)
|
||||
|
||||
@mock.patch('nova.compute.rpcapi.ComputeAPI.pre_live_migration')
|
||||
@mock.patch('nova.compute.manager.ComputeManager._rollback_live_migration')
|
||||
@mock.patch('nova.compute.manager.ComputeManager._post_live_migration')
|
||||
def test_live_migration_wait_vif_plugged_timeout_non_fatal(
|
||||
self, mock_post_live_mig, mock_rollback_live_mig,
|
||||
mock_pre_live_mig):
|
||||
"""Tests the scenario where wait_for_instance_event raises an
|
||||
eventlet Timeout exception and we're configured such that vif plugging
|
||||
failures are NOT fatal.
|
||||
"""
|
||||
self.flags(vif_plugging_is_fatal=False)
|
||||
migrate_data = objects.LibvirtLiveMigrateData(
|
||||
wait_for_vif_plugged=True)
|
||||
mock_pre_live_mig.return_value = migrate_data
|
||||
self.instance.info_cache = objects.InstanceInfoCache(
|
||||
network_info=network_model.NetworkInfo([
|
||||
network_model.VIF(uuids.port1)]))
|
||||
with mock.patch.object(
|
||||
self.compute.virtapi,
|
||||
'wait_for_instance_event') as wait_for_event:
|
||||
wait_for_event.return_value.__enter__.side_effect = (
|
||||
eventlet_timeout.Timeout())
|
||||
self.compute._do_live_migration(
|
||||
self.context, 'dest-host', self.instance, None,
|
||||
self.migration, migrate_data)
|
||||
self.assertEqual('running', self.migration.status)
|
||||
mock_rollback_live_mig.assert_not_called()
|
||||
|
||||
def test_live_migration_force_complete_succeeded(self):
|
||||
migration = objects.Migration()
|
||||
migration.status = 'running'
|
||||
|
||||
@ -75,11 +75,15 @@ class _TestLiveMigrateData(object):
|
||||
props = {
|
||||
'serial_listen_addr': '127.0.0.1',
|
||||
'serial_listen_ports': [1000, 10001, 10002, 10003],
|
||||
'wait_for_vif_plugged': True
|
||||
}
|
||||
|
||||
obj = migrate_data.LibvirtLiveMigrateData(**props)
|
||||
primitive = obj.obj_to_primitive()
|
||||
self.assertIn('serial_listen_ports', primitive['nova_object.data'])
|
||||
self.assertIn('wait_for_vif_plugged', primitive['nova_object.data'])
|
||||
obj.obj_make_compatible(primitive['nova_object.data'], '1.5')
|
||||
self.assertNotIn('wait_for_vif_plugged', primitive['nova_object.data'])
|
||||
obj.obj_make_compatible(primitive['nova_object.data'], '1.1')
|
||||
self.assertNotIn('serial_listen_ports', primitive['nova_object.data'])
|
||||
|
||||
@ -362,12 +366,15 @@ class _TestXenapiLiveMigrateData(object):
|
||||
migrate_send_data={'key': 'val'},
|
||||
sr_uuid_map={'apple': 'banana'},
|
||||
vif_uuid_map={'orange': 'lemon'},
|
||||
old_vol_attachment_ids={uuids.volume: uuids.attachment})
|
||||
old_vol_attachment_ids={uuids.volume: uuids.attachment},
|
||||
wait_for_vif_plugged=True)
|
||||
primitive = obj.obj_to_primitive('1.0')
|
||||
self.assertNotIn('vif_uuid_map', primitive['nova_object.data'])
|
||||
primitive2 = obj.obj_to_primitive('1.1')
|
||||
self.assertIn('vif_uuid_map', primitive2['nova_object.data'])
|
||||
self.assertNotIn('old_vol_attachment_ids', primitive2)
|
||||
primitive3 = obj.obj_to_primitive('1.2')['nova_object.data']
|
||||
self.assertNotIn('wait_for_vif_plugged', primitive3)
|
||||
|
||||
|
||||
class TestXenapiLiveMigrateData(test_objects._LocalTest,
|
||||
@ -384,7 +391,8 @@ class _TestHyperVLiveMigrateData(object):
|
||||
def test_obj_make_compatible(self):
|
||||
obj = migrate_data.HyperVLiveMigrateData(
|
||||
is_shared_instance_path=True,
|
||||
old_vol_attachment_ids={'yes': 'no'})
|
||||
old_vol_attachment_ids={'yes': 'no'},
|
||||
wait_for_vif_plugged=True)
|
||||
|
||||
data = lambda x: x['nova_object.data']
|
||||
|
||||
@ -394,6 +402,8 @@ class _TestHyperVLiveMigrateData(object):
|
||||
self.assertNotIn('is_shared_instance_path', primitive)
|
||||
primitive = data(obj.obj_to_primitive(target_version='1.1'))
|
||||
self.assertNotIn('old_vol_attachment_ids', primitive)
|
||||
primitive = data(obj.obj_to_primitive(target_version='1.2'))
|
||||
self.assertNotIn('wait_for_vif_plugged', primitive)
|
||||
|
||||
def test_to_legacy_dict(self):
|
||||
obj = migrate_data.HyperVLiveMigrateData(
|
||||
@ -435,7 +445,8 @@ class _TestPowerVMLiveMigrateData(object):
|
||||
dest_proc_compat='POWER7',
|
||||
vol_data=dict(three=4),
|
||||
vea_vlan_mappings=dict(five=6),
|
||||
old_vol_attachment_ids=dict(seven=8))
|
||||
old_vol_attachment_ids=dict(seven=8),
|
||||
wait_for_vif_plugged=True)
|
||||
|
||||
@staticmethod
|
||||
def _mk_leg():
|
||||
@ -449,6 +460,7 @@ class _TestPowerVMLiveMigrateData(object):
|
||||
'vol_data': {'three': '4'},
|
||||
'vea_vlan_mappings': {'five': '6'},
|
||||
'old_vol_attachment_ids': {'seven': '8'},
|
||||
'wait_for_vif_plugged': True
|
||||
}
|
||||
|
||||
def test_migrate_data(self):
|
||||
@ -468,6 +480,8 @@ class _TestPowerVMLiveMigrateData(object):
|
||||
self.assertNotIn('vea_vlan_mappings', primitive)
|
||||
primitive = data(obj.obj_to_primitive(target_version='1.1'))
|
||||
self.assertNotIn('old_vol_attachment_ids', primitive)
|
||||
primitive = data(obj.obj_to_primitive(target_version='1.2'))
|
||||
self.assertNotIn('wait_for_vif_plugged', primitive)
|
||||
|
||||
def test_to_legacy_dict(self):
|
||||
self.assertEqual(self._mk_leg(), self._mk_obj().to_legacy_dict())
|
||||
|
||||
@ -1089,7 +1089,7 @@ object_data = {
|
||||
'FloatingIPList': '1.12-e4debd21fddb12cf40d36f737225fa9d',
|
||||
'HostMapping': '1.0-1a3390a696792a552ab7bd31a77ba9ac',
|
||||
'HostMappingList': '1.1-18ac2bfb8c1eb5545bed856da58a79bc',
|
||||
'HyperVLiveMigrateData': '1.2-bcb6dad687369348ffe0f41da6888704',
|
||||
'HyperVLiveMigrateData': '1.3-dae75414c337d3bfd7e4fbf7f74a3c04',
|
||||
'HVSpec': '1.2-de06bcec472a2f04966b855a49c46b41',
|
||||
'IDEDeviceBus': '1.0-29d4c9f27ac44197f01b6ac1b7e16502',
|
||||
'ImageMeta': '1.8-642d1b2eb3e880a367f37d72dd76162d',
|
||||
@ -1114,7 +1114,7 @@ object_data = {
|
||||
'InstancePCIRequest': '1.2-6344dd8bd1bf873e7325c07afe47f774',
|
||||
'InstancePCIRequests': '1.1-65e38083177726d806684cb1cc0136d2',
|
||||
'LibvirtLiveMigrateBDMInfo': '1.1-5f4a68873560b6f834b74e7861d71aaf',
|
||||
'LibvirtLiveMigrateData': '1.5-26f8beff5fe9489efe3dfd3ab7a9eaec',
|
||||
'LibvirtLiveMigrateData': '1.6-9c8e7200a6f80fa7a626b8855c5b394b',
|
||||
'KeyPair': '1.4-1244e8d1b103cc69d038ed78ab3a8cc6',
|
||||
'KeyPairList': '1.3-94aad3ac5c938eef4b5e83da0212f506',
|
||||
'MemoryDiagnostics': '1.0-2c995ae0f2223bb0f8e523c5cc0b83da',
|
||||
@ -1138,7 +1138,7 @@ object_data = {
|
||||
'PciDeviceList': '1.3-52ff14355491c8c580bdc0ba34c26210',
|
||||
'PciDevicePool': '1.1-3f5ddc3ff7bfa14da7f6c7e9904cc000',
|
||||
'PciDevicePoolList': '1.1-15ecf022a68ddbb8c2a6739cfc9f8f5e',
|
||||
'PowerVMLiveMigrateData': '1.2-b62cd242c5205a853545b1085b072340',
|
||||
'PowerVMLiveMigrateData': '1.3-79c635ecf61d1d70b5b9fa04bf778a91',
|
||||
'Quotas': '1.3-40fcefe522111dddd3e5e6155702cf4e',
|
||||
'QuotasNoOp': '1.3-347a039fc7cfee7b225b68b5181e0733',
|
||||
'RequestSpec': '1.9-e506ccb22cd7807a1207c22a3f179387',
|
||||
@ -1166,7 +1166,7 @@ object_data = {
|
||||
'VirtualInterfaceList': '1.0-9750e2074437b3077e46359102779fc6',
|
||||
'VolumeUsage': '1.0-6c8190c46ce1469bb3286a1f21c2e475',
|
||||
'XenDeviceBus': '1.0-272a4f899b24e31e42b2b9a7ed7e9194',
|
||||
'XenapiLiveMigrateData': '1.2-72b9b6e70de34a283689ec7126aa4879',
|
||||
'XenapiLiveMigrateData': '1.3-46659bb17e85ae74dce5e7eeef551e5f',
|
||||
}
|
||||
|
||||
|
||||
|
||||
@ -0,0 +1,20 @@
|
||||
---
|
||||
other:
|
||||
- |
|
||||
A new configuration option, ``[compute]/live_migration_wait_for_vif_plug``,
|
||||
has been added which can be used to configure compute services to wait
|
||||
for network interface plugging to complete on the destination host before
|
||||
starting the guest transfer on the source host during live migration.
|
||||
|
||||
Note that this option is read on the destination host of a live migration.
|
||||
If you set this option the same on all of your compute hosts, which you
|
||||
should do if you use the same networking backend universally, you do not
|
||||
have to worry about this.
|
||||
|
||||
This is disabled by default for backward compatibilty and because the
|
||||
compute service cannot reliably determine which types of virtual
|
||||
interfaces (``port.binding:vif_type``) will send ``network-vif-plugged``
|
||||
events without an accompanying port ``binding:host_id`` change.
|
||||
Open vSwitch and linuxbridge should be OK, but OpenDaylight is at least
|
||||
one known backend that will not currently work in this case, see bug
|
||||
https://launchpad.net/bugs/1755890 for more details.
|
||||
Loading…
Reference in New Issue
Block a user