openstack
/
nova
Code Issues Proposed changes

Set task_state=None when booting instance failed 

When booting instance failed in nova-compute, the
instance is ERROR state, but the instance.task_state
is not None. According to the definition of task_state,
instance task states represent what is happening to
the instance at the current moment. The users will be confused
when they find a instance with vm_state=ERROR and task_state=
spawning.

Change-Id: I6e8d1bd549052317a8b60052e2a0ffbf800af13f
Closes-Bug: #1459496
This commit is contained in:
Rui Chen 2015-06-01 17:20:42 +08:00
parent f1689c67e9
commit 67f29b196c
3 changed files with 42 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
nova/compute/manager.py
View File

@ -749,9 +749,12 @@ class ComputeManager(manager.Manager):
instance.host = None
instance.node = None
def _set_instance_obj_error_state(self, context, instance):
def _set_instance_obj_error_state(self, context, instance,
clean_task_state=False):
try:
instance.vm_state = vm_states.ERROR
if clean_task_state:
instance.task_state = None
instance.save()
except exception.InstanceNotFound:
LOG.debug('Instance has been destroyed from under us while '
@ -1911,7 +1914,8 @@ class ComputeManager(manager.Manager):
compute_utils.add_instance_fault_from_exc(context,
instance, e, sys.exc_info())
self._nil_out_instance_obj_host_and_node(instance)
self._set_instance_obj_error_state(context, instance)
self._set_instance_obj_error_state(context, instance,
clean_task_state=True)
return build_results.FAILED
LOG.debug(e.format_message(), instance=instance)
retry['exc'] = traceback.format_exception(*sys.exc_info())
@ -1953,7 +1957,8 @@ class ComputeManager(manager.Manager):
compute_utils.add_instance_fault_from_exc(context, instance,
e, sys.exc_info())
self._nil_out_instance_obj_host_and_node(instance)
self._set_instance_obj_error_state(context, instance)
self._set_instance_obj_error_state(context, instance,
clean_task_state=True)
return build_results.FAILED
except Exception as e:
# Should not reach here.
@ -1966,7 +1971,8 @@ class ComputeManager(manager.Manager):
compute_utils.add_instance_fault_from_exc(context, instance,
e, sys.exc_info())
self._nil_out_instance_obj_host_and_node(instance)
self._set_instance_obj_error_state(context, instance)
self._set_instance_obj_error_state(context, instance,
clean_task_state=True)
return build_results.FAILED
def _build_and_run_instance(self, context, instance, image, injected_files,

20
nova/tests/unit/compute/test_compute.py
View File

@ -1685,7 +1685,7 @@ class ComputeTestCase(BaseTestCase):
"""block device mapping failure test.
Make sure that when there is a block device mapping problem,
the instance goes to ERROR state, keeping the task state
the instance goes to ERROR state, cleaning the task state
"""
def fake(*args, **kwargs):
raise exception.InvalidBDM()
@ -1700,10 +1700,10 @@ class ComputeTestCase(BaseTestCase):
node=None)
# check state is failed even after the periodic poll
self._assert_state({'vm_state': vm_states.ERROR,
'task_state': task_states.BLOCK_DEVICE_MAPPING})
'task_state': None})
self.compute.periodic_tasks(context.get_admin_context())
self._assert_state({'vm_state': vm_states.ERROR,
'task_state': task_states.BLOCK_DEVICE_MAPPING})
'task_state': None})
@mock.patch('nova.compute.manager.ComputeManager._prep_block_device',
side_effect=exception.OverQuota(overs='volumes'))
@ -1711,8 +1711,8 @@ class ComputeTestCase(BaseTestCase):
"""block device mapping over quota failure test.
Make sure when we're over volume quota according to Cinder client, the
appropriate exception is raised and the instances to ERROR state, keep
the task state.
appropriate exception is raised and the instances to ERROR state,
cleaning the task state.
"""
instance = self._create_fake_instance_obj()
self.compute.build_and_run_instance(
@ -1722,17 +1722,17 @@ class ComputeTestCase(BaseTestCase):
node=None, block_device_mapping=[], image={})
# check state is failed even after the periodic poll
self._assert_state({'vm_state': vm_states.ERROR,
'task_state': task_states.BLOCK_DEVICE_MAPPING})
'task_state': None})
self.compute.periodic_tasks(context.get_admin_context())
self._assert_state({'vm_state': vm_states.ERROR,
'task_state': task_states.BLOCK_DEVICE_MAPPING})
'task_state': None})
self.assertTrue(mock_prep_block_dev.called)
def test_run_instance_spawn_fail(self):
"""spawn failure test.
Make sure that when there is a spawning problem,
the instance goes to ERROR state, keeping the task state.
the instance goes to ERROR state, cleaning the task state.
"""
def fake(*args, **kwargs):
raise test.TestingException()
@ -1745,10 +1745,10 @@ class ComputeTestCase(BaseTestCase):
block_device_mapping=[], image={}, node=None)
# check state is failed even after the periodic poll
self._assert_state({'vm_state': vm_states.ERROR,
'task_state': task_states.SPAWNING})
'task_state': None})
self.compute.periodic_tasks(context.get_admin_context())
self._assert_state({'vm_state': vm_states.ERROR,
'task_state': task_states.SPAWNING})
'task_state': None})
def test_run_instance_dealloc_network_instance_not_found(self):
"""spawn network deallocate test.

26
nova/tests/unit/compute/test_compute_mgr.py
View File

@ -2663,6 +2663,23 @@ class ComputeManagerUnitTestCase(test.NoDBTestCase):
self.assertIsInstance(mock_r.call_args_list[0][0][0],
objects.Instance)
def test_set_instance_obj_error_state_with_clean_task_state(self):
instance = fake_instance.fake_instance_obj(self.context,
vm_state=vm_states.BUILDING, task_state=task_states.SPAWNING)
with mock.patch.object(instance, 'save'):
self.compute._set_instance_obj_error_state(self.context, instance,
clean_task_state=True)
self.assertEqual(vm_states.ERROR, instance.vm_state)
self.assertIsNone(instance.task_state)
def test_set_instance_obj_error_state_by_default(self):
instance = fake_instance.fake_instance_obj(self.context,
vm_state=vm_states.BUILDING, task_state=task_states.SPAWNING)
with mock.patch.object(instance, 'save'):
self.compute._set_instance_obj_error_state(self.context, instance)
self.assertEqual(vm_states.ERROR, instance.vm_state)
self.assertEqual(task_states.SPAWNING, instance.task_state)
class ComputeManagerBuildInstanceTestCase(test.NoDBTestCase):
def setUp(self):
@ -2835,7 +2852,8 @@ class ComputeManagerBuildInstanceTestCase(test.NoDBTestCase):
compute_utils.add_instance_fault_from_exc(self.context,
self.instance, mox.IgnoreArg(), mox.IgnoreArg())
self.compute._nil_out_instance_obj_host_and_node(self.instance)
self.compute._set_instance_obj_error_state(self.context, self.instance)
self.compute._set_instance_obj_error_state(self.context, self.instance,
clean_task_state=True)
self._instance_action_events()
self.mox.ReplayAll()
@ -2996,8 +3014,8 @@ class ComputeManagerBuildInstanceTestCase(test.NoDBTestCase):
compute_utils.add_instance_fault_from_exc(self.context, self.instance,
mox.IgnoreArg(), mox.IgnoreArg())
self.compute._nil_out_instance_obj_host_and_node(self.instance)
self.compute._set_instance_obj_error_state(self.context,
self.instance)
self.compute._set_instance_obj_error_state(self.context, self.instance,
clean_task_state=True)
self._instance_action_events()
self.mox.ReplayAll()
@ -3135,7 +3153,7 @@ class ComputeManagerBuildInstanceTestCase(test.NoDBTestCase):
compute_utils.add_instance_fault_from_exc(self.context,
self.instance, mox.IgnoreArg(), mox.IgnoreArg())
self.compute._set_instance_obj_error_state(self.context,
self.instance)
self.instance, clean_task_state=True)
self._instance_action_events()
self.mox.ReplayAll()