openstack
/
sahara
Code Issues Proposed changes

Force deletion of clusters 

* Add force delete operation to engine using stack abandon
* Basic unit tests for force delete
* Necessary removal of "direct engine" code
* Change schema of APIv2 cluster delete to allow force delete
* Unit tests for new cluster delete schema
* Basic docs about force delete
* Release note

bp sahara-force-delete
Partial-Bug: #1647411

Change-Id: Ida72677c0a4110cb78edf9d62d8330cd4608ff76
This commit is contained in:
Jeremy Freudberg 2017-12-22 05:38:55 +00:00
parent 5535350aff
commit 6850bb8ba8
14 changed files with 178 additions and 230 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
doc/source/user/statuses.rst
View File

@ -144,6 +144,13 @@ Deleting an existing Cluster
The only step, that releases all Cluster's resources and removes it from the
database.
2. Force Deleting
~~~~~~~~~~~~~~~~~
In extreme cases the regular "Deleting" step may hang. Sahara APIv2 introduces
the ability to force delete a Cluster. This prevents deleting from hanging but
comes with the risk of orphaned resources.
Error State
-----------

4
releasenotes/notes/force-delete-apiv2-e372392bbc8639f8.yaml Normal file
View File

@ -0,0 +1,4 @@
---
features:
- The ability to force delete clusters is exposed in Sahara APIv2. The Heat
service must support Stack Abandon for force delete to function properly.

6
sahara/api/v2/clusters.py
View File

@ -86,7 +86,9 @@ def clusters_update(cluster_id, data):
@rest.delete('/clusters/<cluster_id>')
@acl.enforce("data-processing:clusters:delete")
@v.check_exists(api.get_cluster, 'cluster_id')
@v.validate(None, v_c.check_cluster_delete)
@v.validate(v_c_schema.CLUSTER_DELETE_SCHEMA_V2, v_c.check_cluster_delete)
def clusters_delete(cluster_id):
api.terminate_cluster(cluster_id)
data = u.request_data()
force = data.get('force', False)
api.terminate_cluster(cluster_id, force=force)
return u.render()

4
sahara/service/api/v2/clusters.py
View File

@ -149,14 +149,14 @@ def _add_ports_for_auto_sg(ctx, cluster, plugin):
conductor.node_group_update(ctx, ng, ports)
def terminate_cluster(id):
def terminate_cluster(id, force=False):
context.set_current_cluster_id(id)
cluster = c_u.change_cluster_status(id, c_u.CLUSTER_STATUS_DELETING)
if cluster is None:
return
api.OPS.terminate_cluster(id)
api.OPS.terminate_cluster(id, force)
sender.status_notify(cluster.id, cluster.name, cluster.status,
"delete")

115
sahara/service/engine.py
View File

@ -18,7 +18,6 @@ import abc
import datetime
import string
from novaclient import exceptions as nova_exceptions
from oslo_config import cfg
from oslo_log import log as logging
import six
@ -27,14 +26,11 @@ from sahara import conductor as c
from sahara import context
from sahara.i18n import _
from sahara.service import networks
from sahara.service import volumes
from sahara.utils import cluster as cluster_utils
from sahara.utils import cluster_progress_ops as cpo
from sahara.utils import edp
from sahara.utils import general as g
from sahara.utils.openstack import base as b
from sahara.utils.openstack import images as sahara_images
from sahara.utils.openstack import nova
from sahara.utils import poll_utils
from sahara.utils import remote
@ -55,7 +51,7 @@ class Engine(object):
pass
@abc.abstractmethod
def shutdown_cluster(self, cluster):
def shutdown_cluster(self, cluster, force):
pass
@abc.abstractmethod
@ -229,115 +225,6 @@ sed '/^Defaults requiretty*/ s/^/#/' -i /etc/sudoers\n
"end_time": datetime.datetime.now()})
conductor.job_execution_update(ctx, je, update)
def _delete_auto_security_group(self, node_group):
if not node_group.auto_security_group:
return
if not node_group.security_groups:
# node group has no security groups
# nothing to delete
return
name = node_group.security_groups[-1]
try:
client = nova.client().security_groups
security_group = b.execute_with_retries(client.get, name)
if (security_group.name !=
g.generate_auto_security_group_name(node_group)):
LOG.warning("Auto security group for node group {name} is "
"not found".format(name=node_group.name))
return
b.execute_with_retries(client.delete, name)
except Exception:
LOG.warning("Failed to delete security group {name}".format(
name=name))
def _delete_aa_server_groups(self, cluster):
if cluster.anti_affinity:
for i in range(1, cluster.anti_affinity_ratio):
server_group_name = g.generate_aa_group_name(cluster.name, i)
client = nova.client().server_groups
server_groups = b.execute_with_retries(client.findall,
name=server_group_name)
if len(server_groups) == 1:
b.execute_with_retries(client.delete, server_groups[0].id)
'''In case the server group is created
using mitaka or older version'''
old_server_group_name = server_group_name.rsplit('-', 1)[0]
server_groups_old = b.execute_with_retries(
client.findall,
name=old_server_group_name)
if len(server_groups_old) == 1:
b.execute_with_retries(client.delete,
server_groups_old[0].id)
def _shutdown_instance(self, instance):
# Heat dissociates and deletes upon deletion of resources
# See OS::Neutron::FloatingIP and OS::Neutron::FloatingIPAssociation
if instance.node_group.floating_ip_pool:
pass
try:
volumes.detach_from_instance(instance)
except Exception:
LOG.warning("Detaching volumes from instance failed")
try:
b.execute_with_retries(nova.client().servers.delete,
instance.instance_id)
except nova_exceptions.NotFound:
LOG.warning("Attempted to delete non-existent instance")
conductor.instance_remove(context.ctx(), instance)
@cpo.event_wrapper(mark_successful_on_exit=False)
def _check_if_deleted(self, instance):
try:
nova.get_instance_info(instance)
except nova_exceptions.NotFound:
return True
return False
@poll_utils.poll_status(
'delete_instances_timeout',
_("Wait for instances to be deleted"), sleep=1)
def _check_deleted(self, deleted_ids, cluster, instances):
if not cluster_utils.check_cluster_exists(cluster):
return True
for instance in instances:
if instance.id not in deleted_ids:
with context.set_current_instance_id(instance.instance_id):
if self._check_if_deleted(instance):
LOG.debug("Instance is deleted")
deleted_ids.add(instance.id)
cpo.add_successful_event(instance)
return len(deleted_ids) == len(instances)
def _await_deleted(self, cluster, instances):
"""Await all instances are deleted."""
if not instances:
return
cpo.add_provisioning_step(
cluster.id, _("Wait for instances to be deleted"), len(instances))
deleted_ids = set()
self._check_deleted(deleted_ids, cluster, instances)
def _shutdown_instances(self, cluster):
for node_group in cluster.node_groups:
for instance in node_group.instances:
with context.set_current_instance_id(instance.instance_id):
self._shutdown_instance(instance)
self._await_deleted(cluster, node_group.instances)
self._delete_auto_security_group(node_group)
def _remove_db_objects(self, cluster):
ctx = context.ctx()
cluster = conductor.cluster_get(ctx, cluster)

25
sahara/service/heat/heat_engine.py
View File

@ -191,21 +191,22 @@ class HeatEngine(e.Engine):
self._launch_instances(cluster, rollback_count, ROLLBACK_STAGES,
update_stack=True)
def shutdown_cluster(self, cluster):
def shutdown_cluster(self, cluster, force=False):
"""Shutdown specified cluster and all related resources."""
if force:
heat_shutdown = heat.abandon_stack
else:
heat_shutdown = heat.delete_stack
try:
heat.delete_stack(cluster)
heat_shutdown(cluster)
except heat_exc.HTTPNotFound:
LOG.warning('Did not find stack for cluster. Trying to delete '
'cluster manually.')
# Stack not found. Trying to delete cluster like direct engine
# do it
self._shutdown_instances(cluster)
self._delete_aa_server_groups(cluster)
self._clean_job_executions(cluster)
self._remove_db_objects(cluster)
LOG.warning('Did not find stack for cluster.')
except heat_exc.BadRequest:
LOG.error("Can't force delete cluster.", exc_info=True)
finally:
self._clean_job_executions(cluster)
self._remove_db_objects(cluster)
@cpo.event_wrapper(
True, step=_('Create Heat stack'), param=('cluster', 1))

16
sahara/service/ops.py
View File

@ -60,9 +60,9 @@ class LocalOps(object):
_provision_scaled_cluster, cluster_id, node_group_id_map,
node_group_instance_map)
def terminate_cluster(self, cluster_id):
def terminate_cluster(self, cluster_id, force=False):
context.spawn("cluster-terminating-%s" % cluster_id,
terminate_cluster, cluster_id)
terminate_cluster, cluster_id, force)
def run_edp_job(self, job_execution_id):
context.spawn("Starting Job Execution %s" % job_execution_id,
@ -102,8 +102,8 @@ class RemoteOps(rpc_utils.RPCClient):
node_group_id_map=node_group_id_map,
node_group_instance_map=node_group_instance_map)
def terminate_cluster(self, cluster_id):
self.cast('terminate_cluster', cluster_id=cluster_id)
def terminate_cluster(self, cluster_id, force=False):
self.cast('terminate_cluster', cluster_id=cluster_id, force=force)
def run_edp_job(self, job_execution_id):
self.cast('run_edp_job', job_execution_id=job_execution_id)
@ -153,8 +153,8 @@ class OpsServer(rpc_utils.RPCServer):
node_group_instance_map)
@request_context
def terminate_cluster(self, cluster_id):
terminate_cluster(cluster_id)
def terminate_cluster(self, cluster_id, force=False):
terminate_cluster(cluster_id, force)
@request_context
def run_edp_job(self, job_execution_id):
@ -394,7 +394,7 @@ def _get_random_instance_from_ng(instances, instances_to_delete):
@ops_error_handler(
_("Terminating cluster failed for the following reason(s): {reason}"))
def terminate_cluster(cluster_id):
def terminate_cluster(cluster_id, force=False):
ctx = context.ctx()
_setup_trust_for_cluster(cluster_id)
@ -406,7 +406,7 @@ def terminate_cluster(cluster_id):
plugin.on_terminate_cluster(cluster)
context.set_step_type(_("Engine: shutdown cluster"))
INFRA.shutdown_cluster(cluster)
INFRA.shutdown_cluster(cluster, force)
trusts.delete_trust_from_cluster(cluster)

11
sahara/service/validations/clusters_schema.py
View File

@ -146,3 +146,14 @@ CLUSTER_SCALING_SCHEMA_V2['properties']['resize_node_groups'][
},
}
})
CLUSTER_DELETE_SCHEMA_V2 = {
"type": "object",
"properties": {
"force": {
"type": "boolean"
}
},
"additionalProperties": False,
"required": []
}

2
sahara/tests/unit/service/api/v2/test_clusters.py
View File

@ -74,7 +74,7 @@ class FakeOps(object):
conductor.node_group_update(context.ctx(), ng, {'count': count})
conductor.cluster_update(context.ctx(), id, {'status': 'Scaled'})
def terminate_cluster(self, id):
def terminate_cluster(self, id, force):
self.calls_order.append('ops.terminate_cluster')
def _get_instance(self, cluster, instances_to_delete):

135
sahara/tests/unit/service/test_engine.py
View File

@ -20,7 +20,6 @@ import mock
from sahara.service import engine
from sahara.service.heat import heat_engine
from sahara.tests.unit import base
from sahara.utils import general as g
class EngineTest(engine.Engine):
@ -97,84 +96,11 @@ class TestDeletion(base.SaharaTestCase):
super(TestDeletion, self).setUp()
self.engine = EngineTest()
@mock.patch('sahara.utils.openstack.nova.client')
def test_delete_auto_security_group(self, nova_client):
ng = mock.Mock(id="16fd2706-8baf-433b-82eb-8c7fada847da",
auto_security_group=True)
ng.name = "ngname"
ng.cluster.name = "cluster"
auto_name = g.generate_auto_security_group_name(ng)
ng.security_groups = [auto_name]
client = mock.Mock()
nova_client.return_value = client
client.security_groups.get.side_effect = lambda x: SecurityGroup(x)
self.engine._delete_auto_security_group(ng)
client.security_groups.delete.assert_called_once_with(auto_name)
@mock.patch('sahara.utils.openstack.nova.client')
def test_delete_auto_security_group_other_groups(self, nova_client):
ng = mock.Mock(id="16fd2706-8baf-433b-82eb-8c7fada847da",
auto_security_group=True)
ng.name = "ngname"
ng.cluster.name = "cluster"
auto_name = g.generate_auto_security_group_name(ng)
ng.security_groups = ['1', '2', auto_name]
client = mock.Mock()
nova_client.return_value = client
client.security_groups.get.side_effect = lambda x: SecurityGroup(x)
self.engine._delete_auto_security_group(ng)
client.security_groups.delete.assert_called_once_with(auto_name)
@mock.patch('sahara.utils.openstack.nova.client')
def test_delete_auto_security_group_no_groups(self, nova_client):
ng = mock.Mock(id="16fd2706-8baf-433b-82eb-8c7fada847da",
auto_security_group=True)
ng.name = "ngname"
ng.cluster.name = "cluster"
ng.security_groups = []
client = mock.Mock()
nova_client.return_value = client
client.security_groups.get.side_effect = lambda x: SecurityGroup(x)
self.engine._delete_auto_security_group(ng)
self.assertEqual(0, client.security_groups.delete.call_count)
@mock.patch('sahara.utils.openstack.nova.client')
def test_delete_auto_security_group_wrong_group(self, nova_client):
ng = mock.Mock(id="16fd2706-8baf-433b-82eb-8c7fada847da",
auto_security_group=True)
ng.name = "ngname"
ng.cluster.name = "cluster"
ng.security_groups = ['1', '2']
client = mock.Mock()
nova_client.return_value = client
client.security_groups.get.side_effect = lambda x: SecurityGroup(x)
self.engine._delete_auto_security_group(ng)
self.assertEqual(0, client.security_groups.delete.call_count)
@mock.patch('sahara.service.engine.Engine._delete_aa_server_groups')
@mock.patch('sahara.service.engine.Engine._shutdown_instances')
@mock.patch('sahara.service.engine.Engine._remove_db_objects')
@mock.patch('sahara.service.engine.Engine._clean_job_executions')
@mock.patch('sahara.utils.openstack.heat.client')
@mock.patch('sahara.service.heat.heat_engine.LOG.warning')
def test_calls_order(self, logger, heat_client, _job_ex, _db_ob,
_shutdown, _delete_aa):
def test_calls_order(self, logger, heat_client, _job_ex, _db_ob):
class FakeHeatEngine(heat_engine.HeatEngine):
def __init__(self):
super(FakeHeatEngine, self).__init__()
@ -188,26 +114,55 @@ class TestDeletion(base.SaharaTestCase):
self.order.append('remove_db_objects')
super(FakeHeatEngine, self)._remove_db_objects(cluster)
def _shutdown_instances(self, cluster):
self.order.append('shutdown_instances')
super(FakeHeatEngine, self)._shutdown_instances(cluster)
def _delete_aa_server_groups(self, cluster):
self.order.append('delete_aa_server_groups')
super(FakeHeatEngine, self)._delete_aa_server_groups(cluster)
fake_cluster = mock.Mock()
heat_client.side_effect = heat_exc.HTTPNotFound()
engine = FakeHeatEngine()
engine.shutdown_cluster(fake_cluster)
self.assertEqual(['shutdown_instances', 'delete_aa_server_groups',
'clean_job_executions', 'remove_db_objects'],
self.assertEqual(['clean_job_executions', 'remove_db_objects'],
engine.order)
self.assertEqual(
[mock.call('Did not find stack for cluster. Trying to '
'delete cluster manually.')], logger.call_args_list)
[mock.call('Did not find stack for cluster.')],
logger.call_args_list
)
@mock.patch('sahara.service.heat.heat_engine.LOG.error')
@mock.patch('sahara.utils.openstack.heat.delete_stack')
@mock.patch('sahara.utils.openstack.heat.abandon_stack')
@mock.patch('sahara.service.engine.Engine._remove_db_objects')
@mock.patch('sahara.service.engine.Engine._clean_job_executions')
def test_force_delete_calls(self, cj, rdb, abandon, delete, log_err):
engine = heat_engine.HeatEngine()
class SecurityGroup(object):
def __init__(self, name):
self.name = name
# Force delete when Heat service support abandon
engine.shutdown_cluster(mock.Mock(), force=True)
self.assertEqual(delete.call_count, 0)
self.assertEqual(abandon.call_count, 1)
self.assertEqual(cj.call_count, 1)
self.assertEqual(rdb.call_count, 1)
delete.reset_mock()
abandon.reset_mock()
rdb.reset_mock()
cj.reset_mock()
# Regular delete
engine.shutdown_cluster(mock.Mock(), force=False)
self.assertEqual(delete.call_count, 1)
self.assertEqual(abandon.call_count, 0)
self.assertEqual(cj.call_count, 1)
self.assertEqual(rdb.call_count, 1)
delete.reset_mock()
abandon.reset_mock()
rdb.reset_mock()
cj.reset_mock()
# Force delete when stack abandon unavailable
abandon.side_effect = heat_exc.BadRequest()
engine.shutdown_cluster(mock.Mock(), force=True)
self.assertEqual(delete.call_count, 0)
self.assertEqual(abandon.call_count, 1)
self.assertEqual(cj.call_count, 1)
self.assertEqual(rdb.call_count, 1)
log_err.assert_called_once_with(
"Can't force delete cluster.", exc_info=True)

2
sahara/tests/unit/service/test_ops.py
View File

@ -70,7 +70,7 @@ class FakeINFRA(object):
TestOPS.SEQUENCE.append('INFRA.scale_cluster')
return True
def shutdown_cluster(self, cluster):
def shutdown_cluster(self, cluster, force):
TestOPS.SEQUENCE.append('shutdown_cluster')
def rollback_cluster(self, cluster, reason):

33
sahara/tests/unit/service/validation/test_cluster_delete_validation.py
View File

@ -17,7 +17,9 @@ import mock
import testtools
from sahara import exceptions as ex
from sahara.service import validation as v
from sahara.service.validations import clusters as c_val
from sahara.service.validations import clusters_schema as c_schema
from sahara.tests.unit.service.validation import utils as u
from sahara.tests.unit import testutils as tu
@ -52,3 +54,34 @@ class TestClusterDeleteValidation(u.ValidationTestCase):
except ex.DeletionFailed as e:
self.assert_created_in_another_tenant_exception(e)
raise e
class TestClusterDeleteValidationV2(testtools.TestCase):
@mock.patch("sahara.utils.api.request_data")
@mock.patch("sahara.utils.api.bad_request")
def _validate_body(self, request, br, rd):
m_func = mock.Mock()
m_func.__name__ = "m_func"
rd.return_value = request
validator = v.validate(c_schema.CLUSTER_DELETE_SCHEMA_V2, m_func)
validator(m_func)(data=request)
return not br.call_count
def test_delete_schema_empty_body(self):
request = {}
self.assertTrue(self._validate_body(request))
def test_delete_schema_wrong_type(self):
request = {"force": "True"}
self.assertFalse(self._validate_body(request))
def test_delete_schema_extra_fields(self):
request = {"force": True, "just_kidding": False}
self.assertFalse(self._validate_body(request))
def test_delete_schema_good(self):
request = {"force": True}
self.assertTrue(self._validate_body(request))

39
sahara/tests/unit/utils/openstack/test_heat.py Normal file
View File

@ -0,0 +1,39 @@
# Copyright (c) 2017 Massachusetts Open Cloud
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
# implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import mock
import testtools
from sahara.utils.openstack import heat as heat_u
class HeatClientTest(testtools.TestCase):
@mock.patch('heatclient.client.Client')
@mock.patch('sahara.utils.openstack.base.url_for')
@mock.patch('sahara.service.sessions.cache')
@mock.patch('sahara.context.ctx')
def test_abandon(self, ctx, cache, url_for, heat):
class _FakeHeatStacks(object):
def delete(self, stack):
call_list.append("delete")
def abandon(self, stack):
call_list.append("abandon")
heat.return_value.stacks = _FakeHeatStacks()
call_list = []
heat_u.abandon_stack(mock.Mock())
self.assertEqual(call_list, ["delete", "abandon"])

9
sahara/utils/openstack/heat.py
View File

@ -84,6 +84,15 @@ def delete_stack(cluster):
stack = get_stack(stack_name, raise_on_missing=False)
def abandon_stack(cluster):
'''Put stack in deleting state if not already, then abandon it.'''
heat_client = client()
stack_name = cluster.stack_name
base.execute_with_retries(heat_client.stacks.delete, stack_name)
# TODO(jfreud): sleep between calls?
base.execute_with_retries(heat_client.stacks.abandon, stack_name)
def get_stack_outputs(cluster):
stack = get_stack(cluster.stack_name)
stack.get()