265 lines
9.4 KiB
Python
265 lines
9.4 KiB
Python
from __future__ import absolute_import
|
|
|
|
import math
|
|
import random
|
|
import typing
|
|
|
|
from oslo_log import log
|
|
import testtools
|
|
|
|
import tobiko
|
|
from tobiko.openstack import neutron
|
|
from tobiko.openstack import tests
|
|
from tobiko.openstack import topology
|
|
from tobiko.tests.faults.ha import cloud_disruptions
|
|
from tobiko.tripleo import pacemaker
|
|
from tobiko.tripleo import processes
|
|
from tobiko.tripleo import containers
|
|
from tobiko.tripleo import nova
|
|
from tobiko.tripleo import undercloud
|
|
from tobiko.tripleo import validations
|
|
|
|
|
|
LOG = log.getLogger(__name__)
|
|
|
|
|
|
def overcloud_health_checks(passive_checks_only=False,
|
|
skip_mac_table_size_test=True):
|
|
# this method will be changed in future commit
|
|
check_pacemaker_resources_health()
|
|
check_overcloud_processes_health()
|
|
nova.check_nova_services_health()
|
|
tests.test_neutron_agents_are_alive()
|
|
if not passive_checks_only:
|
|
# create a uniq stack
|
|
check_vm_create()
|
|
nova.start_all_instances()
|
|
nova.check_computes_vms_running_via_virsh()
|
|
containers.list_node_containers.cache_clear()
|
|
containers.assert_all_tripleo_containers_running()
|
|
containers.assert_equal_containers_state()
|
|
containers.run_container_config_validations()
|
|
tests.test_ovn_dbs_validations()
|
|
# skip_mac_table_size_test has to be removed when BZ1695122 is resolved
|
|
# we need it for the moment because this validation should not be performed
|
|
# after any overcloud node is rebooted
|
|
if not skip_mac_table_size_test:
|
|
tests.test_ovs_bridges_mac_table_size()
|
|
validations.run_post_deployment_validations()
|
|
|
|
|
|
# check vm create with ssh and ping checks
|
|
def check_vm_create():
|
|
tests.test_server_creation()
|
|
|
|
|
|
# check cluster failed statuses
|
|
def check_pacemaker_resources_health():
|
|
return pacemaker.PacemakerResourcesStatus().all_healthy
|
|
|
|
|
|
def check_overcloud_processes_health():
|
|
procs = processes.OvercloudProcessesStatus()
|
|
return (procs.basic_overcloud_processes_running and
|
|
procs.ovn_overcloud_processes_validations)
|
|
|
|
|
|
class OvercloudHealthCheck(tobiko.SharedFixture):
|
|
|
|
skips: typing.FrozenSet[str] = frozenset()
|
|
|
|
@classmethod
|
|
def run_before(cls, **params):
|
|
cls.run(after=False, **params)
|
|
|
|
@classmethod
|
|
def run_after(cls, **params):
|
|
cls.run(after=True, **params)
|
|
|
|
@classmethod
|
|
def run(cls, after: bool, **params):
|
|
fixture = tobiko.get_fixture(cls)
|
|
params.setdefault('passive_checks_only', False)
|
|
params.setdefault('skip_mac_table_size_test', True)
|
|
skips = frozenset(k for k, v in params.items() if v)
|
|
if after or skips < fixture.skips:
|
|
# Force re-check
|
|
tobiko.cleanup_fixture(fixture)
|
|
else:
|
|
LOG.info("Will skip Overcloud health checks if already "
|
|
f"executed: {params}")
|
|
fixture.skips = skips
|
|
tobiko.setup_fixture(fixture)
|
|
|
|
def setup_fixture(self):
|
|
# run validations
|
|
params = {name: True
|
|
for name in self.skips}
|
|
LOG.info(f"Start executing Overcloud health checks: {params}.")
|
|
overcloud_health_checks(**params)
|
|
LOG.info(f"Overcloud health checks successfully executed: {params}.")
|
|
|
|
def cleanup_fixture(self):
|
|
self.skips = frozenset()
|
|
|
|
|
|
@undercloud.skip_if_missing_undercloud
|
|
class DisruptTripleoNodesTest(testtools.TestCase):
|
|
""" HA Tests: run health check -> disruptive action -> health check
|
|
disruptive_action: a function that runs some
|
|
disruptive scenario on a overcloud"""
|
|
|
|
def test_0vercloud_health_check(self):
|
|
OvercloudHealthCheck.run_before(skip_mac_table_size_test=False)
|
|
|
|
def test_hard_reboot_controllers_recovery(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.reset_all_controller_nodes()
|
|
OvercloudHealthCheck.run_after()
|
|
|
|
def test_soft_reboot_computes_recovery(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.reset_all_compute_nodes(hard_reset=False)
|
|
# verify VM status is updated after reboot
|
|
nova.wait_for_all_instances_status('SHUTOFF')
|
|
# start all VM instance
|
|
# otherwise sidecar containers will not run after computes reboot
|
|
nova.start_all_instances()
|
|
OvercloudHealthCheck.run_after(passive_checks_only=True)
|
|
|
|
# TODO(eolivare): the following test is skipped due to rhbz#1890895
|
|
# def test_hard_reboot_computes_recovery(self):
|
|
# OvercloudHealthCheck.run_before()
|
|
# cloud_disruptions.reset_all_compute_nodes(hard_reset=True)
|
|
# # verify VM status is updated after reboot
|
|
# nova.wait_for_all_instances_status('SHUTOFF')
|
|
# # start all VM instance
|
|
# # otherwise sidecar containers will not run after computes reboot
|
|
# nova.start_all_instances()
|
|
# OvercloudHealthCheck.run_after(passive_checks_only=True)
|
|
|
|
def test_reboot_controller_main_vip(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.reset_controller_main_vip()
|
|
OvercloudHealthCheck.run_after()
|
|
|
|
def test_reboot_controller_non_main_vip(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.reset_controllers_non_main_vip()
|
|
OvercloudHealthCheck.run_after()
|
|
|
|
def test_crash_controller_main_vip(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.crash_controller_main_vip()
|
|
OvercloudHealthCheck.run_after()
|
|
|
|
def test_crash_controller_non_main_vip(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.crash_controllers_non_main_vip()
|
|
OvercloudHealthCheck.run_after()
|
|
|
|
@pacemaker.skip_if_fencing_not_deployed
|
|
def test_network_disruptor_main_vip(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.network_disrupt_controller_main_vip()
|
|
OvercloudHealthCheck.run_after()
|
|
cloud_disruptions.network_undisrupt_controller_main_vip()
|
|
|
|
# @pacemaker.skip_if_fencing_not_deployed
|
|
# def test_network_disruptor_non_main_vip(self):
|
|
# OvercloudHealthCheck.run_before()
|
|
# cloud_disruptions.network_disrupt_controllers_non_main_vip()
|
|
# OvercloudHealthCheck.run_after()
|
|
# cloud_disruptions.network_undisrupt_controllers_non_main_vip()
|
|
|
|
@neutron.skip_unless_is_ovn()
|
|
def test_reset_ovndb_pcs_master_resource(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.reset_ovndb_pcs_master_resource()
|
|
OvercloudHealthCheck.run_after()
|
|
|
|
@neutron.skip_unless_is_ovn()
|
|
def test_reset_ovndb_pcs_resource(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.reset_ovndb_pcs_resource()
|
|
OvercloudHealthCheck.run_after()
|
|
|
|
@neutron.skip_unless_is_ovn()
|
|
def test_reset_ovndb_master_container(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.reset_ovndb_master_container()
|
|
OvercloudHealthCheck.run_after()
|
|
|
|
def test_kill_rabbitmq_service_one_controller(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.kill_rabbitmq_service()
|
|
OvercloudHealthCheck.run_after()
|
|
|
|
def test_kill_all_galera_services(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.kill_all_galera_services()
|
|
OvercloudHealthCheck.run_after()
|
|
|
|
def test_remove_all_grastate_galera(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.remove_all_grastate_galera()
|
|
OvercloudHealthCheck.run_after()
|
|
|
|
def test_remove_one_grastate_galera(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.remove_one_grastate_galera()
|
|
OvercloudHealthCheck.run_after()
|
|
|
|
def test_request_galera_sst(self):
|
|
OvercloudHealthCheck.run_before()
|
|
cloud_disruptions.request_galera_sst()
|
|
OvercloudHealthCheck.run_after()
|
|
|
|
def _test_controllers_shutdown(self, shutdown_method):
|
|
all_nodes = topology.list_openstack_nodes(group='controller')
|
|
if len(all_nodes) < 3:
|
|
self.skipTest('It requires at least three controller nodes')
|
|
LOG.info("Ensure all controller nodes are running...")
|
|
for node in all_nodes:
|
|
node.power_on_overcloud_node()
|
|
|
|
LOG.info("Verify can create VMs before controllers power off...")
|
|
tests.test_server_creation()
|
|
|
|
quorum_level = math.ceil(0.5 * len(all_nodes))
|
|
assert quorum_level >= len(all_nodes) - quorum_level
|
|
nodes = random.sample(all_nodes, quorum_level)
|
|
|
|
LOG.info(f"Power off {quorum_level} random controller nodes: "
|
|
f"{[node.name for node in nodes]}")
|
|
shutdown_method(nodes)
|
|
|
|
random.shuffle(nodes)
|
|
LOG.info("Power on controller nodes: "
|
|
f"{[node.name for node in nodes]}")
|
|
for node in nodes:
|
|
node.power_on_overcloud_node()
|
|
|
|
LOG.info("Wait until pacemaker resources will be healthy again...")
|
|
check_pacemaker_resources_health()
|
|
LOG.info("Verify can create VMs after controllers power on...")
|
|
tests.test_server_creation()
|
|
|
|
def test_controllers_shutdown(self):
|
|
|
|
def shutdown_nodes(nodes):
|
|
for node in nodes:
|
|
node.power_off_overcloud_node()
|
|
|
|
self._test_controllers_shutdown(shutdown_nodes)
|
|
|
|
def test_controllers_hard_power_down(self):
|
|
self._test_controllers_shutdown(
|
|
cloud_disruptions.hard_power_off_overcloud_nodes)
|
|
|
|
# [..]
|
|
# more tests to follow
|
|
# run health checks
|
|
# faults stop rabbitmq service on one controller
|
|
# run health checks again
|