openstack/fuel-qa
Code Issues Proposed changes
eb18c17903
fuel-qa/fuelweb_test/models/fuel_web_client.py
Vladimir Khlyunev ca7b2e49f6 Fetch next deployable release basen on OPENSTACK_RELEASE var 
Right now we have mitaka and mitaka+UCA releases which both uses
Ubuntu as operation system. We want to predict "which release will be
taken for next operations".
'octane upgrade-env' now requires release id so this patch is required
for further changes
https://review.openstack.org/#/c/358881/

Change-Id: Icc4c2f534a13e11995129bad62b4fc985633fb34
2016-08-24 09:01:41 +00:00

3226 lines
137 KiB
Python
Raw Blame History

# Copyright 2015 Mirantis, Inc.
#
# 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.
from __future__ import division
import logging
import re
import time
import traceback
from warnings import warn
import distutils
import devops
from devops.error import DevopsCalledProcessError
from devops.error import TimeoutError
from devops.helpers.helpers import wait_pass
from devops.helpers.helpers import wait
from devops.models.node import Node
try:
from devops.error import DevopsObjNotFound
except ImportError:
# pylint: disable=no-member
DevopsObjNotFound = Node.DoesNotExist
# pylint: enable=no-member
from keystoneauth1 import exceptions
import netaddr
from proboscis.asserts import assert_equal
from proboscis.asserts import assert_false
from proboscis.asserts import assert_is_not_none
from proboscis.asserts import assert_not_equal
from proboscis.asserts import assert_raises
from proboscis.asserts import assert_true
import yaml
from fuelweb_test import logger
from fuelweb_test import logwrap
from fuelweb_test import ostf_test_mapping
from fuelweb_test import QuietLogger
from fuelweb_test.helpers import ceph
from fuelweb_test.helpers import checkers
from fuelweb_test.helpers import replace_repos
from fuelweb_test.helpers.decorators import check_repos_management
from fuelweb_test.helpers.decorators import custom_repo
from fuelweb_test.helpers.decorators import download_astute_yaml
from fuelweb_test.helpers.decorators import download_packages_json
from fuelweb_test.helpers.decorators import duration
from fuelweb_test.helpers.decorators import retry
from fuelweb_test.helpers.decorators import update_fuel
from fuelweb_test.helpers.decorators import upload_manifests
from fuelweb_test.helpers.security import SecurityChecks
from fuelweb_test.helpers.ssh_manager import SSHManager
from fuelweb_test.helpers.ssl_helpers import change_cluster_ssl_config
from fuelweb_test.helpers.ssl_helpers import copy_cert_from_master
from fuelweb_test.helpers.uca import change_cluster_uca_config
from fuelweb_test.helpers.utils import get_node_hiera_roles
from fuelweb_test.helpers.utils import node_freemem
from fuelweb_test.helpers.utils import pretty_log
from fuelweb_test.models.nailgun_client import NailgunClient
import fuelweb_test.settings as help_data
from fuelweb_test.settings import ATTEMPTS
from fuelweb_test.settings import BONDING
from fuelweb_test.settings import DEPLOYMENT_MODE_HA
from fuelweb_test.settings import DISABLE_SSL
from fuelweb_test.settings import DNS_SUFFIX
from fuelweb_test.settings import iface_alias
from fuelweb_test.settings import KVM_USE
from fuelweb_test.settings import MULTIPLE_NETWORKS
from fuelweb_test.settings import NOVA_QUOTAS_ENABLED
from fuelweb_test.settings import NETWORK_PROVIDERS
from fuelweb_test.settings import NEUTRON
from fuelweb_test.settings import NEUTRON_SEGMENT
from fuelweb_test.settings import NEUTRON_SEGMENT_TYPE
from fuelweb_test.settings import NODEGROUPS
from fuelweb_test.settings import OPENSTACK_RELEASE
from fuelweb_test.settings import OPENSTACK_RELEASE_UBUNTU
from fuelweb_test.settings import OSTF_TEST_NAME
from fuelweb_test.settings import OSTF_TEST_RETRIES_COUNT
from fuelweb_test.settings import REPLACE_DEFAULT_REPOS
from fuelweb_test.settings import REPLACE_DEFAULT_REPOS_ONLY_ONCE
from fuelweb_test.settings import SSL_CN
from fuelweb_test.settings import TIMEOUT
from fuelweb_test.settings import UCA_ENABLED
from fuelweb_test.settings import USER_OWNED_CERT
from fuelweb_test.settings import VCENTER_DATACENTER
from fuelweb_test.settings import VCENTER_DATASTORE
from fuelweb_test.settings import VCENTER_IP
from fuelweb_test.settings import VCENTER_PASSWORD
from fuelweb_test.settings import VCENTER_USERNAME
class FuelWebClient29(object):
"""FuelWebClient.""" # TODO documentation
def __init__(self, environment):
self.ssh_manager = SSHManager()
self.admin_node_ip = self.ssh_manager.admin_ip
self.client = NailgunClient(self.ssh_manager.admin_ip)
self._environment = environment
self.security = SecurityChecks(self.client, self._environment)
super(FuelWebClient29, self).__init__()
@property
def environment(self):
"""Environment Model
:rtype: EnvironmentModel
"""
return self._environment
@staticmethod
@logwrap
def get_cluster_status(os_conn, smiles_count, networks_count=2):
checkers.verify_service_list_api(os_conn, service_count=smiles_count)
checkers.verify_glance_image_api(os_conn)
checkers.verify_network_list_api(os_conn, networks_count)
@logwrap
def _ostf_test_wait(self, cluster_id, timeout):
logger.info('Wait OSTF tests at cluster #%s for %s seconds',
cluster_id, timeout)
wait(
lambda: all([run['status'] == 'finished'
for run in
self.client.get_ostf_test_run(cluster_id)]),
timeout=timeout,
timeout_msg='OSTF tests run timeout '
'(cluster_id={})'.format(cluster_id))
return self.client.get_ostf_test_run(cluster_id)
@logwrap
def _tasks_wait(self, tasks, timeout):
return [self.task_wait(task, timeout) for task in tasks]
@logwrap
def add_syslog_server(self, cluster_id, host, port):
logger.info('Add syslog server %s:%s to cluster #%s',
host, port, cluster_id)
self.client.add_syslog_server(cluster_id, host, port)
@logwrap
def assert_cluster_floating_list(self, os_conn, cluster_id, expected_ips):
logger.info('Assert floating IPs on cluster #{0}. Expected {1}'.format(
cluster_id, expected_ips))
current_ips = self.get_cluster_floating_list(os_conn, cluster_id)
assert_equal(set(expected_ips), set(current_ips),
'Current floating IPs {0}'.format(current_ips))
@logwrap
def assert_cluster_ready(self, os_conn, smiles_count,
networks_count=2, timeout=300):
logger.info('Assert cluster services are UP')
# TODO(astudenov): add timeout_msg
wait_pass(
lambda: self.get_cluster_status(
os_conn,
smiles_count=smiles_count,
networks_count=networks_count),
timeout=timeout)
@logwrap
def assert_ha_services_ready(self, cluster_id, timeout=20 * 60,
should_fail=0):
"""Wait until HA services are UP.
Should be used before run any other check for services."""
if self.get_cluster_mode(cluster_id) == DEPLOYMENT_MODE_HA:
logger.info('Waiting {0} sec. for passed OSTF HA tests.'
.format(timeout))
with QuietLogger(logging.ERROR):
# TODO(astudenov): add timeout_msg
wait_pass(lambda: self.run_ostf(cluster_id,
test_sets=['ha'],
should_fail=should_fail),
interval=20, timeout=timeout)
logger.info('OSTF HA tests passed successfully.')
else:
logger.debug('Cluster {0} is not in HA mode, OSTF HA tests '
'skipped.'.format(cluster_id))
@logwrap
def assert_os_services_ready(self, cluster_id, timeout=5 * 60,
should_fail=0):
"""Wait until OpenStack services are UP.
Should be used before run any other check for services."""
logger.info('Waiting {0} sec. for passed OSTF Sanity checks.'
.format(timeout))
with QuietLogger():
# TODO(astudenov): add timeout_msg
wait_pass(lambda: self.run_ostf(cluster_id,
test_sets=['sanity'],
should_fail=should_fail),
interval=10, timeout=timeout)
logger.info('OSTF Sanity checks passed successfully.')
@logwrap
def assert_ostf_run_certain(self, cluster_id, tests_must_be_passed,
timeout=10 * 60):
"""Wait for OSTF tests to finish, check that the tests specified
in [tests_must_be_passed] are passed"""
logger.info('Assert OSTF tests are passed at cluster #{0}: {1}'.format(
cluster_id, pretty_log(tests_must_be_passed, indent=1)))
set_result_list = self._ostf_test_wait(cluster_id, timeout)
tests_pass_count = 0
tests_count = len(tests_must_be_passed)
fail_details = []
for set_result in set_result_list:
for test in set_result['tests']:
if test['id'] in tests_must_be_passed:
if test['status'] == 'success':
tests_pass_count += 1
logger.info('Passed OSTF test %s found', test['id'])
else:
details = ('%s (%s). Test status: %s, message: %s'
% (test['name'], test['id'], test['status'],
test['message']))
fail_details.append(details)
assert_true(tests_pass_count == tests_count,
'The following tests have not succeeded, while they '
'must have passed: {}'.format(pretty_log(fail_details,
indent=1)))
@logwrap
def assert_ostf_run(self, cluster_id, should_fail=0, failed_test_name=None,
timeout=15 * 60, test_sets=None):
"""Wait for OSTF tests to finish, check that there is no failed tests.
If [failed_test_name] tests are expected, ensure that these tests
are not passed"""
logger.info('Assert OSTF run at cluster #{0}. '
'Should fail {1} tests named {2}'.format(cluster_id,
should_fail,
failed_test_name))
set_result_list = self._ostf_test_wait(cluster_id, timeout)
failed_tests_res = []
failed = 0
actual_failed_names = []
test_result = {}
for set_result in set_result_list:
if set_result['testset'] not in test_sets:
continue
failed += len([test for test in set_result['tests']
if test['status'] in {'failure', 'error'}])
for test in set_result['tests']:
test_result.update({test['name']: test['status']})
if test['status'] not in ['success', 'disabled', 'skipped']:
actual_failed_names.append(test['name'])
key = ('{name:s} ({status:s})'
''.format(name=test['name'], status=test['status']))
failed_tests_res.append(
{key: test['message']})
logger.info('OSTF test statuses are :\n{}\n'.format(
pretty_log(test_result, indent=1)))
if failed_test_name:
for test_name in actual_failed_names:
assert_true(test_name in failed_test_name,
'WARNING! Unexpected fail: '
'expected {0}, actual {1}'.format(
failed_test_name, actual_failed_names)
)
assert_true(
failed <= should_fail, 'Failed {0} OSTF tests; should fail'
' {1} tests. Names of failed tests: {2}'
.format(failed,
should_fail,
pretty_log(failed_tests_res,
indent=1)))
def assert_release_state(self, release_name, state='available'):
logger.info('Assert release %s has state %s', release_name, state)
for release in self.client.get_releases():
if release["name"].lower().find(release_name) != -1:
assert_equal(release['state'], state,
'Release state {0}'.format(release['state']))
return release["id"]
def assert_release_role_present(self, release_name, role_name):
logger.info('Assert role %s is available in release %s',
role_name, release_name)
release_id = self.assert_release_state(release_name)
release_data = self.client.get_releases_details(release_id=release_id)
assert_equal(
True, role_name in release_data['roles'],
message='There is no {0} role in release id {1}'.format(
role_name, release_name))
@logwrap
def assert_fuel_version(self, fuel_version):
logger.info('Assert fuel version is {0}'.format(fuel_version))
version = self.client.get_api_version()
logger.debug('version get from api is {0}'.format(version['release']))
assert_equal(version['release'], fuel_version,
'Release state is not {0}'.format(fuel_version))
@logwrap
def assert_nailgun_upgrade_migration(self,
key='can_update_from_versions'):
for release in self.client.get_releases():
assert_true(key in release)
@logwrap
def assert_task_success(
self, task, timeout=130 * 60, interval=5, progress=None):
def _message(_task):
if 'message' in _task:
return _task['message']
else:
return ''
logger.info('Assert task %s is success', task)
if not progress:
task = self.task_wait(task, timeout, interval)
assert_equal(
task['status'], 'ready',
"Task '{0}' has incorrect status. {1} != {2}, '{3}'".format(
task["name"], task['status'], 'ready', _message(task)
)
)
else:
logger.info('Start to polling task progress')
task = self.task_wait_progress(
task, timeout=timeout, interval=interval, progress=progress)
assert_not_equal(
task['status'], 'error',
"Task '{0}' has error status. '{1}'"
.format(task['status'], _message(task)))
assert_true(
task['progress'] >= progress,
'Task has other progress{0}'.format(task['progress']))
@logwrap
def assert_task_failed(self, task, timeout=70 * 60, interval=5):
logger.info('Assert task %s is failed', task)
task = self.task_wait(task, timeout, interval)
assert_equal(
'error', task['status'],
"Task '{name}' has incorrect status. {} != {}".format(
task['status'], 'error', name=task["name"]
)
)
@logwrap
def assert_all_tasks_completed(self, cluster_id=None):
cluster_info_template = "\n\tCluster ID: {cluster}{info}\n"
all_tasks = sorted(
self.client.get_all_tasks_list(),
key=lambda _tsk: _tsk['id'],
reverse=True
)
not_ready_tasks, deploy_tasks = checkers.incomplete_tasks(
all_tasks, cluster_id)
not_ready_transactions = checkers.incomplete_deploy(
{
cluster: self.client.get_deployment_task_hist(task_id)
for cluster, task_id in deploy_tasks.items()})
if len(not_ready_tasks) > 0:
task_details_template = (
"\n"
"\t\tTask name: {name}\n"
"\t\t\tStatus: {status}\n"
"\t\t\tProgress: {progress}\n"
"\t\t\tResult: {result}\n"
"\t\t\tMessage: {message}\n"
"\t\t\tTask ID: {id}"
)
task_text = 'Not all tasks completed: {}'.format(
''.join(
cluster_info_template.format(
cluster=cluster,
info="".join(
task_details_template.format(**task)
for task in tasks))
for cluster, tasks in sorted(not_ready_tasks.items())
))
logger.error(task_text)
if len(not_ready_transactions) == 0:
# Else: we will raise assert with detailed info
# about deployment
assert_true(len(not_ready_tasks) == 0, task_text)
checkers.fail_deploy(not_ready_transactions)
def wait_node_is_online(self, devops_node, timeout=60 * 5):
logger.info(
'Wait for node {!r} online status'.format(devops_node.name))
wait(lambda: self.get_nailgun_node_by_devops_node(
devops_node)['online'],
timeout=timeout,
timeout_msg='Node {!r} failed to become online'
''.format(devops_node.name))
def wait_node_is_offline(self, devops_node, timeout=60 * 5):
logger.info(
'Wait for node {!r} offline status'.format(devops_node.name))
wait(lambda: not self.get_nailgun_node_by_devops_node(
devops_node)['online'],
timeout=timeout,
timeout_msg='Node {!r} failed to become offline'
''.format(devops_node.name))
@logwrap
def fqdn(self, devops_node):
logger.info('Get FQDN of a devops node %s', devops_node.name)
nailgun_node = self.get_nailgun_node_by_devops_node(devops_node)
if OPENSTACK_RELEASE_UBUNTU in OPENSTACK_RELEASE:
return nailgun_node['meta']['system']['fqdn']
return nailgun_node['fqdn']
@logwrap
def get_pcm_nodes(self, ctrl_node, pure=False):
nodes = {}
with self.get_ssh_for_node(ctrl_node) as remote:
pcs_status = remote.execute('pcs status nodes')['stdout']
pcm_nodes = yaml.load(''.join(pcs_status).strip())
for status in ('Online', 'Offline', 'Standby'):
list_nodes = (pcm_nodes['Pacemaker Nodes'][status] or '').split()
if not pure:
nodes[status] = [self.get_fqdn_by_hostname(x)
for x in list_nodes]
else:
nodes[status] = list_nodes
return nodes
@logwrap
def get_rabbit_running_nodes(self, ctrl_node):
"""
:param ctrl_node: str
:return: list
"""
ip = self.get_node_ip_by_devops_name(ctrl_node)
cmd = 'rabbitmqctl cluster_status'
# If any rabbitmq nodes failed, we have return(70) from rabbitmqctl
# Acceptable list:
# 0 | EX_OK | Self-explanatory
# 69 | EX_UNAVAILABLE | Failed to connect to node
# 70 | EX_SOFTWARE | Any other error discovered when running command
# | | against live node
# 75 | EX_TEMPFAIL | Temporary failure (e.g. something timed out)
rabbit_status = self.ssh_manager.execute_on_remote(
ip, cmd, raise_on_assert=False, assert_ec_equal=[0, 69, 70, 75]
)['stdout_str']
rabbit_status = re.sub(r',\n\s*', ',', rabbit_status)
found_nodes = re.search(
"\{running_nodes,\[([^\]]*)\]\}",
rabbit_status)
if not found_nodes:
logger.info(
'No running rabbitmq nodes found on {0}. Status:\n {1}'.format(
ctrl_node, rabbit_status))
return []
rabbit_nodes = found_nodes.group(1).replace("'", "").split(',')
logger.debug('rabbit nodes are {}'.format(rabbit_nodes))
nodes = [node.replace('rabbit@', "") for node in rabbit_nodes]
hostname_prefix = self.ssh_manager.execute_on_remote(
ip, 'hiera node_name_prefix_for_messaging', raise_on_assert=False
)['stdout_str']
if hostname_prefix not in ('', 'nil'):
nodes = [n.replace(hostname_prefix, "") for n in nodes]
return nodes
@logwrap
def assert_pacemaker(self, ctrl_node, online_nodes, offline_nodes):
logger.info('Assert pacemaker status at devops node %s', ctrl_node)
online = sorted([self.fqdn(n) for n in online_nodes])
offline = sorted([self.fqdn(n) for n in offline_nodes])
try:
wait(lambda: self.get_pcm_nodes(ctrl_node)['Online'] == online and
self.get_pcm_nodes(ctrl_node)['Offline'] == offline,
timeout=60)
except TimeoutError:
nodes = self.get_pcm_nodes(ctrl_node)
assert_true(nodes['Online'] == online,
'Online nodes: {0} ; should be online: {1}'
.format(nodes['Online'], online))
assert_true(nodes['Offline'] == offline,
'Offline nodes: {0} ; should be offline: {1}'
.format(nodes['Offline'], offline))
@logwrap
@upload_manifests
@update_fuel
def create_cluster(self,
name,
settings=None,
release_name=OPENSTACK_RELEASE,
mode=DEPLOYMENT_MODE_HA,
port=514,
release_id=None,
configure_ssl=True):
"""Creates a cluster
:param name:
:param release_name:
:param mode:
:param settings:
:param port:
:param configure_ssl:
:param cgroup_data:
:return: cluster_id
"""
logger.info('Create cluster with name %s', name)
if not release_id:
release_id = self.client.get_release_id(release_name=release_name)
logger.info('Release_id of %s is %s',
release_name, str(release_id))
if settings is None:
settings = {}
if REPLACE_DEFAULT_REPOS and not REPLACE_DEFAULT_REPOS_ONLY_ONCE:
self.replace_default_repos()
cluster_id = self.client.get_cluster_id(name)
if not cluster_id:
data = {
"name": name,
"release": str(release_id),
"mode": mode
}
if "net_provider" in settings:
data.update({'net_provider': settings["net_provider"]})
if "net_segment_type" in settings:
data.update({'net_segment_type': settings["net_segment_type"]})
# NEUTRON_SEGMENT_TYPE should not override any option
# configured from test, in case if test is going to set only
# 'net_provider' for a cluster.
if (NEUTRON_SEGMENT_TYPE and
"net_provider" not in settings and
"net_segment_type" not in settings):
data.update(
{
'net_provider': NEUTRON,
'net_segment_type': NEUTRON_SEGMENT[
NEUTRON_SEGMENT_TYPE]
}
)
self.client.create_cluster(data=data)
cluster_id = self.client.get_cluster_id(name)
logger.info('The cluster id is %s', cluster_id)
logger.info('Set cluster settings to {}'.format(
pretty_log(settings, indent=1)))
attributes = self.client.get_cluster_attributes(cluster_id)
for option in settings:
section = ''
if option in ('sahara', 'murano', 'ceilometer', 'mongo',
'ironic'):
section = 'additional_components'
elif option in {'mongo_db_name', 'mongo_replset', 'mongo_user',
'hosts_ip', 'mongo_password'}:
section = 'external_mongo'
elif option in {'volumes_ceph', 'images_ceph',
'ephemeral_ceph', 'objects_ceph',
'osd_pool_size', 'volumes_lvm',
'images_vcenter'}:
section = 'storage'
elif option in {'tenant', 'password', 'user'}:
section = 'access'
elif option == 'assign_to_all_nodes':
section = 'public_network_assignment'
elif option in {'neutron_l3_ha', 'neutron_dvr',
'neutron_l2_pop'}:
section = 'neutron_advanced_configuration'
elif option in {'dns_list'}:
section = 'external_dns'
elif option in {'ntp_list'}:
section = 'external_ntp'
elif option in {'propagate_task_deploy'}:
section = 'common'
if section:
try:
attributes['editable'][section][option]['value'] =\
settings[option]
except KeyError:
if section not in attributes['editable']:
raise KeyError(
"Section '{0}' not in "
"attributes['editable']: {1}".format(
section, attributes['editable'].keys()))
raise KeyError(
"Option {0} not in attributes['editable'][{1}]: "
"{2}".format(
option, section,
attributes['editable'][section].keys()))
# we should check DVR limitations
section = 'neutron_advanced_configuration'
if attributes['editable'][section]['neutron_dvr']['value']:
if attributes['editable'][section]['neutron_l3_ha']['value']:
raise Exception("Neutron DVR and Neutron L3 HA can't be"
" used simultaneously.")
if 'net_segment_type' in settings:
net_segment_type = settings['net_segment_type']
elif NEUTRON_SEGMENT_TYPE:
net_segment_type = NEUTRON_SEGMENT[NEUTRON_SEGMENT_TYPE]
else:
net_segment_type = None
if not attributes['editable'][section]['neutron_l2_pop'][
'value'] and net_segment_type == 'tun':
raise Exception("neutron_l2_pop is not enabled but "
"it is required for VxLAN DVR "
"network configuration.")
public_gw = self.get_public_gw()
if help_data.FUEL_USE_LOCAL_NTPD\
and ('ntp_list' not in settings)\
and checkers.is_ntpd_active(
self.ssh_manager.admin_ip, public_gw):
attributes['editable']['external_ntp']['ntp_list']['value'] =\
[public_gw]
logger.info("Configuring cluster #{0}"
"to use NTP server {1}"
.format(cluster_id, public_gw))
if help_data.FUEL_USE_LOCAL_DNS and ('dns_list' not in settings):
attributes['editable']['external_dns']['dns_list']['value'] =\
[public_gw]
logger.info("Configuring cluster #{0} to use DNS server {1}"
.format(cluster_id, public_gw))
logger.info('Set DEBUG MODE to %s', help_data.DEBUG_MODE)
attributes['editable']['common']['debug']['value'] = \
help_data.DEBUG_MODE
if KVM_USE:
logger.info('Set Hypervisor type to KVM')
hpv_data = attributes['editable']['common']['libvirt_type']
hpv_data['value'] = "kvm"
if help_data.VCENTER_USE:
logger.info('Enable Dual Hypervisors Mode')
hpv_data = attributes['editable']['common']['use_vcenter']
hpv_data['value'] = True
if NOVA_QUOTAS_ENABLED:
logger.info('Enable Nova quotas')
nova_quotas = attributes['editable']['common']['nova_quota']
nova_quotas['value'] = True
if not help_data.TASK_BASED_ENGINE:
logger.info('Switch to Granular deploy')
attributes['editable']['common']['task_deploy']['value'] =\
False
# Updating attributes is needed before updating
# networking configuration because additional networks
# may be created by new components like ironic
self.client.update_cluster_attributes(cluster_id, attributes)
self.nodegroups_configure(cluster_id)
logger.debug("Try to update cluster "
"with next attributes {0}".format(attributes))
self.client.update_cluster_attributes(cluster_id, attributes)
if configure_ssl:
self.ssl_configure(cluster_id)
if UCA_ENABLED or settings.get('uca_enabled', False):
self.enable_uca(cluster_id)
if not cluster_id:
raise Exception("Could not get cluster '{:s}'".format(name))
# TODO: rw105719
# self.client.add_syslog_server(
# cluster_id, self.environment.get_host_node_ip(), port)
return cluster_id
@logwrap
def get_public_gw(self):
return self.environment.d_env.router(router_name="public")
@logwrap
def nodegroups_configure(self, cluster_id):
"""Update nodegroups configuration
"""
if not MULTIPLE_NETWORKS:
return
ng = {rack['name']: [] for rack in NODEGROUPS}
self.update_nodegroups(cluster_id=cluster_id, node_groups=ng)
self.update_nodegroups_network_configuration(cluster_id, NODEGROUPS)
@logwrap
def ssl_configure(self, cluster_id):
attributes = self.client.get_cluster_attributes(cluster_id)
change_cluster_ssl_config(attributes, SSL_CN)
logger.debug("Try to update cluster "
"with next attributes {0}".format(attributes))
self.client.update_cluster_attributes(cluster_id, attributes)
@logwrap
def enable_uca(self, cluster_id):
attributes = self.client.get_cluster_attributes(cluster_id)
change_cluster_uca_config(attributes)
logger.debug("Try to update cluster "
"with next attributes {0}".format(attributes))
self.client.update_cluster_attributes(cluster_id, attributes)
@logwrap
def vcenter_configure(self, cluster_id, vcenter_value=None,
multiclusters=None, vc_glance=None,
target_node_1='controllers',
target_node_2='controllers'):
if not vcenter_value:
vcenter_value = {
"glance": {
"vcenter_username": "",
"datacenter": "",
"vcenter_host": "",
"vcenter_password": "",
"datastore": "", },
"availability_zones": [
{"vcenter_username": VCENTER_USERNAME,
"nova_computes": [
{"datastore_regex": ".*",
"vsphere_cluster": "Cluster1",
"service_name": "vmcluster1",
"target_node": {
"current": {"id": target_node_1,
"label": target_node_1},
"options": [{"id": "controllers",
"label": "controllers"}, ]},
},
],
"vcenter_host": VCENTER_IP,
"az_name": "vcenter",
"vcenter_password": VCENTER_PASSWORD,
}],
"network": {"esxi_vlan_interface": "vmnic0"}
}
if multiclusters:
multiclusters =\
vcenter_value["availability_zones"][0]["nova_computes"]
multiclusters.append(
{"datastore_regex": ".*",
"vsphere_cluster": "Cluster2",
"service_name": "vmcluster2",
"target_node": {
"current": {"id": target_node_2,
"label": target_node_2},
"options": [{"id": "controllers",
"label": "controllers"}, ]},
})
if vc_glance:
vcenter_value["glance"]["vcenter_username"] = VCENTER_USERNAME
vcenter_value["glance"]["datacenter"] = VCENTER_DATACENTER
vcenter_value["glance"]["vcenter_host"] = VCENTER_IP
vcenter_value["glance"]["vcenter_password"] = VCENTER_PASSWORD
vcenter_value["glance"]["datastore"] = VCENTER_DATASTORE
if help_data.VCENTER_USE:
logger.info('Configuring vCenter...')
vmware_attributes = \
self.client.get_cluster_vmware_attributes(cluster_id)
vcenter_data = vmware_attributes['editable']
vcenter_data['value'] = vcenter_value
logger.debug("Try to update cluster with next "
"vmware_attributes {0}".format(vmware_attributes))
self.client.update_cluster_vmware_attributes(cluster_id,
vmware_attributes)
logger.debug("Attributes of cluster were updated")
def add_local_ubuntu_mirror(self, cluster_id, name='Auxiliary',
path=help_data.LOCAL_MIRROR_UBUNTU,
suite='auxiliary', section='main',
priority=help_data.EXTRA_DEB_REPOS_PRIORITY):
# Append new mirror to attributes of currently creating Ubuntu cluster
mirror_url = path.replace('/var/www/nailgun',
'http://{0}:8080'.format(self.admin_node_ip))
mirror = '{0},deb {1} {2} {3}'.format(name, mirror_url, suite, section)
attributes = self.client.get_cluster_attributes(cluster_id)
repos_attr = attributes['editable']['repo_setup']['repos']
repos_attr['value'] = replace_repos.add_ubuntu_extra_mirrors(
repos=repos_attr['value'],
prefix=suite,
mirrors=mirror,
priority=priority)
replace_repos.report_ubuntu_repos(repos_attr['value'])
self.client.update_cluster_attributes(cluster_id, attributes)
def add_local_centos_mirror(self, cluster_id, repo_name='auxiliary',
path=help_data.LOCAL_MIRROR_CENTOS,
priority=help_data.EXTRA_RPM_REPOS_PRIORITY):
# Append new mirror to attributes of currently creating CentOS cluster
mirror_url = path.replace('/var/www/nailgun',
'http://{0}:8080'.format(self.admin_node_ip))
mirror = '{0},{1}'.format(repo_name, mirror_url)
attributes = self.client.get_cluster_attributes(cluster_id)
repos_attr = attributes['editable']['repo_setup']['repos']
repos_attr['value'] = replace_repos.add_centos_extra_mirrors(
repos=repos_attr['value'],
mirrors=mirror,
priority=priority)
replace_repos.report_centos_repos(repos_attr['value'])
self.client.update_cluster_attributes(cluster_id, attributes)
def replace_default_repos(self):
# Replace Ubuntu default repositories for the release
logger.info("Replace default repository list.")
ubuntu_id = self.client.get_release_id(
release_name=help_data.OPENSTACK_RELEASE_UBUNTU)
ubuntu_release = self.client.get_release(ubuntu_id)
ubuntu_meta = ubuntu_release["attributes_metadata"]
repos_ubuntu = ubuntu_meta["editable"]["repo_setup"]["repos"]
repos_ubuntu["value"] = replace_repos.replace_ubuntu_repos(
repos_ubuntu, upstream_host='archive.ubuntu.com')
self.client.put_release(ubuntu_id, ubuntu_release)
replace_repos.report_ubuntu_repos(repos_ubuntu["value"])
# Replace CentOS default repositories for the release
centos_id = self.client.get_release_id(
release_name=help_data.OPENSTACK_RELEASE_CENTOS)
centos_release = self.client.get_release(centos_id)
centos_meta = centos_release["attributes_metadata"]
repos_centos = centos_meta["editable"]["repo_setup"]["repos"]
repos_centos["value"] = replace_repos.replace_centos_repos(
repos_centos, upstream_host=self.admin_node_ip)
self.client.put_release(centos_id, centos_release)
replace_repos.report_centos_repos(repos_centos["value"])
def get_cluster_repos(self, cluster_id):
attributes = self.client.get_cluster_attributes(cluster_id)
return attributes['editable']['repo_setup']['repos']
def check_deploy_state(self, cluster_id, check_services=True,
check_tasks=True):
if check_tasks:
self.assert_all_tasks_completed(cluster_id=cluster_id)
if check_services:
self.assert_ha_services_ready(cluster_id)
self.assert_os_services_ready(cluster_id)
if not DISABLE_SSL and not USER_OWNED_CERT:
with self.environment.d_env.get_admin_remote() as admin_remote:
copy_cert_from_master(admin_remote, cluster_id)
n_nodes = self.client.list_cluster_nodes(cluster_id)
for n in filter(lambda n: 'ready' in n['status'], n_nodes):
node = self.get_devops_node_by_nailgun_node(n)
if node:
node_name = node.name
with self.get_ssh_for_node(node_name) as remote:
free = node_freemem(remote)
hiera_roles = get_node_hiera_roles(remote)
node_status = {
node_name:
{
'Host': n['hostname'],
'Roles':
{
'Nailgun': n['roles'],
'Hiera': hiera_roles,
},
'Memory':
{
'RAM': free['mem'],
'SWAP': free['swap'],
},
},
}
logger.info('Node status: {}'.format(pretty_log(node_status,
indent=1)))
@download_packages_json
@download_astute_yaml
@duration
@check_repos_management
@custom_repo
def deploy_cluster_wait(self, cluster_id, is_feature=False,
timeout=help_data.DEPLOYMENT_TIMEOUT, interval=30,
check_services=True, check_tasks=False,
allow_partially_deploy=False):
warn_txt = ('Temporary: flag check_tasks is set to False, '
'until bugs LP#1578218 and LP#1578257 fixed')
logger.warning(warn_txt)
warn(warn_txt, UserWarning)
cluster_attributes = self.client.get_cluster_attributes(cluster_id)
self.client.assign_ip_address_before_deploy_start(cluster_id)
network_settings = self.client.get_networks(cluster_id)
if not is_feature and help_data.DEPLOYMENT_RETRIES == 1:
logger.info('Deploy cluster %s', cluster_id)
task = self.deploy_cluster(cluster_id)
self.assert_task_success(task, interval=interval, timeout=timeout)
self.check_cluster_status(cluster_id, allow_partially_deploy)
self.check_deploy_state(cluster_id, check_services, check_tasks)
return
logger.info('Provision nodes of a cluster %s', cluster_id)
task = self.client.provision_nodes(cluster_id)
self.assert_task_success(task, timeout=timeout, interval=interval)
for retry_number in range(help_data.DEPLOYMENT_RETRIES):
logger.info('Deploy nodes of a cluster %s, run: %s',
cluster_id, str(retry_number + 1))
task = self.client.deploy_nodes(cluster_id)
self.assert_task_success(task, timeout=timeout, interval=interval)
self.check_cluster_status(cluster_id, allow_partially_deploy)
self.check_deploy_state(cluster_id, check_services, check_tasks)
self.check_cluster_settings(cluster_id, cluster_attributes)
self.check_network_settings(cluster_id, network_settings)
self.check_deployment_info_save_for_task(cluster_id)
def check_cluster_status(self, cluster_id, allow_partially_deploy):
cluster_info = self.client.get_cluster(cluster_id)
cluster_status = cluster_info['status']
error_msg = \
"Cluster is not deployed: some nodes are in the Error state"
check = 'operational' in cluster_status
if not check and allow_partially_deploy:
logger.warning(error_msg)
elif not check:
assert_true(check, error_msg)
else:
logger.info("Cluster with id {} is in Operational state".format(
cluster_id))
@logwrap
def check_cluster_settings(self, cluster_id, cluster_attributes):
task_id = self.get_last_task_id(cluster_id, 'deployment')
cluster_settings = \
self.client.get_cluster_settings_for_deployment_task(task_id)
logger.debug('Cluster settings before deploy {}'.format(
cluster_attributes))
logger.debug('Cluster settings after deploy {}'.format(
cluster_settings))
assert_equal(cluster_attributes, cluster_settings,
message='Cluster attributes before deploy are not equal'
' with cluster settings after deploy')
@logwrap
def check_network_settings(self, cluster_id, network_settings):
task_id = self.get_last_task_id(cluster_id, 'deployment')
network_configuration = \
self.client.get_network_configuration_for_deployment_task(task_id)
logger.debug('Network settings before deploy {}'.format(
network_settings))
logger.debug('Network settings after deploy {}'.format(
network_configuration))
assert_equal(network_settings, network_configuration,
message='Network settings from cluster configuration '
'and deployment task are not equal')
@logwrap
def check_deployment_info_save_for_task(self, cluster_id):
try:
task_id = self.get_last_task_id(cluster_id, 'deployment')
self.client.get_deployment_info_for_task(task_id)
except Exception:
logger.error(
"Cannot get information about deployment for task {}".format(
task_id))
@logwrap
def get_last_task_id(self, cluster_id, task_name):
tasks = self.client.get_tasks()
tasks_ids = []
for task in tasks:
if task['cluster'] == cluster_id and task['name'] == task_name:
tasks_ids.append(task['id'])
return min(tasks_ids)
def deploy_cluster_wait_progress(self, cluster_id, progress,
return_task=None):
task = self.deploy_cluster(cluster_id)
self.assert_task_success(task, interval=30, progress=progress)
if return_task:
return task
def redeploy_cluster_changes_wait_progress(self, cluster_id, progress,
data=None, return_task=None):
logger.info('Re-deploy cluster {}'
' to apply the changed settings'.format(cluster_id))
if data is None:
data = {}
task = self.client.redeploy_cluster_changes(cluster_id, data)
self.assert_task_success(task, interval=30, progress=progress)
if return_task:
return task
@logwrap
def deploy_cluster(self, cluster_id):
"""Return hash with task description."""
logger.info('Launch deployment of a cluster #%s', cluster_id)
return self.client.deploy_cluster_changes(cluster_id)
@logwrap
def get_cluster_predefined_networks_name(self, cluster_id):
net_params = self.client.get_networks(
cluster_id)['networking_parameters']
return {'private_net': net_params.get('internal_name', 'net04'),
'external_net': net_params.get('floating_name', 'net04_ext')}
@logwrap
def get_cluster_floating_list(self, os_conn, cluster_id):
logger.info('Get floating IPs list at cluster #{0}'.format(cluster_id))
subnet = os_conn.get_subnet('{0}__subnet'.format(
self.get_cluster_predefined_networks_name(
cluster_id)['external_net']))
ret = []
for pool in subnet['allocation_pools']:
ret.extend([str(ip) for ip in
netaddr.iter_iprange(pool['start'], pool['end'])])
return ret
@logwrap
def get_cluster_block_devices(self, node_name):
logger.info('Get %s node block devices (lsblk)', node_name)
with self.get_ssh_for_node(node_name) as remote:
ret = remote.check_call('/bin/lsblk')
return ''.join(ret['stdout'])
@logwrap
def get_pacemaker_status(self, controller_node_name):
logger.info('Get pacemaker status at %s node', controller_node_name)
with self.get_ssh_for_node(controller_node_name) as remote:
return ''.join(remote.check_call('crm_mon -1')['stdout'])
@logwrap
def get_pacemaker_config(self, controller_node_name):
logger.info('Get pacemaker config at %s node', controller_node_name)
with self.get_ssh_for_node(controller_node_name) as remote:
return ''.join(remote.check_call('crm_resource --list')['stdout'])
@logwrap
def get_pacemaker_resource_location(self, controller_node_name,
resource_name):
"""Get devops nodes where the resource is running."""
logger.info('Get pacemaker resource %s life status at %s node',
resource_name, controller_node_name)
hosts = []
with self.get_ssh_for_node(controller_node_name) as remote:
for line in remote.check_call(
'crm_resource --resource {0} '
'--locate --quiet'.format(resource_name))['stdout']:
hosts.append(
self.get_devops_node_by_nailgun_fqdn(line.strip()))
return hosts
@logwrap
def get_last_created_cluster(self):
# return id of last created cluster
logger.info('Get ID of a last created cluster')
clusters = self.client.list_clusters()
if len(clusters) > 0:
return clusters.pop()['id']
return None
@logwrap
def get_nailgun_node_roles(self, nodes_dict):
nailgun_node_roles = []
for node_name in nodes_dict:
slave = self.environment.d_env.get_node(name=node_name)
node = self.get_nailgun_node_by_devops_node(slave)
nailgun_node_roles.append((node, nodes_dict[node_name]))
return nailgun_node_roles
@logwrap
def get_nailgun_node_by_name(self, node_name):
logger.info('Get nailgun node by %s devops node', node_name)
return self.get_nailgun_node_by_devops_node(
self.environment.d_env.get_node(name=node_name))
@logwrap
def get_nailgun_node_by_base_name(self, base_node_name):
logger.debug('Get nailgun node by "{0}" base '
'node name.'.format(base_node_name))
nodes = self.client.list_nodes()
for node in nodes:
if base_node_name in node['name']:
return node
@logwrap
def get_nailgun_node_by_devops_node(self, devops_node):
"""Return slave node description.
Returns dict with nailgun slave node description if node is
registered. Otherwise return None.
"""
d_macs = {netaddr.EUI(i.mac_address) for i in devops_node.interfaces}
logger.debug('Verify that nailgun api is running')
attempts = ATTEMPTS
nodes = []
while attempts > 0:
logger.debug(
'current timeouts is {0} count of '
'attempts is {1}'.format(TIMEOUT, attempts))
try:
nodes = self.client.list_nodes()
logger.debug('Got nodes %s', nodes)
attempts = 0
except Exception:
logger.debug(traceback.format_exc())
attempts -= 1
time.sleep(TIMEOUT)
logger.debug('Look for nailgun node by macs %s', d_macs)
for nailgun_node in nodes:
node_nics = self.client.get_node_interfaces(nailgun_node['id'])
macs = {netaddr.EUI(nic['mac'])
for nic in node_nics if nic['type'] == 'ether'}
logger.debug('Look for macs returned by nailgun {0}'.format(macs))
# Because our HAproxy may create some interfaces
if d_macs.issubset(macs):
nailgun_node['devops_name'] = devops_node.name
return nailgun_node
# On deployed environment MAC addresses of bonded network interfaces
# are changes and don't match addresses associated with devops node
if BONDING:
return self.get_nailgun_node_by_base_name(devops_node.name)
@logwrap
def get_nailgun_node_by_fqdn(self, fqdn):
"""Return nailgun node with fqdn
:type fqdn: String
:rtype: Dict
"""
for nailgun_node in self.client.list_nodes():
if nailgun_node['meta']['system']['fqdn'] == fqdn:
return nailgun_node
@logwrap
def get_nailgun_node_by_status(self, status):
"""Return nailgun nodes with status
:type status: String
:rtype: List
"""
returned_nodes = []
for nailgun_node in self.client.list_nodes():
if nailgun_node['status'] == status:
returned_nodes.append(nailgun_node)
return returned_nodes
@logwrap
def find_devops_node_by_nailgun_fqdn(self, fqdn, devops_nodes):
"""Return devops node by nailgun fqdn
:type fqdn: String
:type devops_nodes: List
:rtype: Devops Node or None
"""
nailgun_node = self.get_nailgun_node_by_fqdn(fqdn)
macs = {netaddr.EUI(i['mac']) for i in
nailgun_node['meta']['interfaces']}
for devops_node in devops_nodes:
devops_macs = {netaddr.EUI(i.mac_address) for i in
devops_node.interfaces}
if devops_macs == macs:
return devops_node
@logwrap
def get_devops_node_by_mac(self, mac_address):
"""Return devops node by nailgun node
:type mac_address: String
:rtype: Node or None
"""
for node in self.environment.d_env.nodes():
for iface in node.interfaces:
if netaddr.EUI(iface.mac_address) == netaddr.EUI(mac_address):
return node
@logwrap
def get_devops_nodes_by_nailgun_nodes(self, nailgun_nodes):
"""Return devops node by nailgun node
:type nailgun_nodes: List
:rtype: list of Nodes or None
"""
d_nodes = [self.get_devops_node_by_nailgun_node(n) for n
in nailgun_nodes]
d_nodes = [n for n in d_nodes if n is not None]
return d_nodes if len(d_nodes) == len(nailgun_nodes) else None
@logwrap
def get_devops_node_by_nailgun_node(self, nailgun_node):
"""Return devops node by nailgun node
:type nailgun_node: Dict
:rtype: Node or None
"""
if nailgun_node:
return self.get_devops_node_by_mac(nailgun_node['mac'])
@logwrap
def get_devops_node_by_nailgun_fqdn(self, fqdn):
"""Return devops node with nailgun fqdn
:type fqdn: String
:rtype: Devops Node or None
"""
return self.get_devops_node_by_nailgun_node(
self.get_nailgun_node_by_fqdn(fqdn))
@logwrap
def get_nailgun_cluster_nodes_by_roles(self, cluster_id, roles,
role_status='roles'):
"""Return list of nailgun nodes from cluster with cluster_id which have
a roles
:type cluster_id: Int
:type roles: list
:rtype: list
"""
nodes = self.client.list_cluster_nodes(cluster_id=cluster_id)
return [n for n in nodes if set(roles) <= set(n[role_status])]
@logwrap
def get_node_ip_by_devops_name(self, node_name):
"""Get node ip by it's devops name (like "slave-01" and etc)
:param node_name: str
:return: str
"""
# TODO: This method should be part of fuel-devops
try:
node = self.get_nailgun_node_by_devops_node(
self.environment.d_env.get_node(name=node_name))
except DevopsObjNotFound:
node = self.get_nailgun_node_by_fqdn(node_name)
assert_true(node is not None,
'Node with name "{0}" not found!'.format(node_name))
return node['ip']
@logwrap
def get_ssh_for_node(self, node_name):
return self.environment.d_env.get_ssh_to_remote(
self.get_node_ip_by_devops_name(node_name))
@logwrap
def get_ssh_for_role(self, nodes_dict, role):
node_name = sorted(filter(lambda name: role in nodes_dict[name],
nodes_dict.keys()))[0]
return self.get_ssh_for_node(node_name)
@logwrap
def get_ssh_for_nailgun_node(self, nailgun_node):
return self.environment.d_env.get_ssh_to_remote(nailgun_node['ip'])
@logwrap
def is_node_discovered(self, nailgun_node):
return any(
map(lambda node:
node['mac'] == nailgun_node['mac'] and
node['status'] == 'discover', self.client.list_nodes()))
def wait_node_is_discovered(self, nailgun_node, timeout=6 * 60):
logger.info('Wait for node {!r} to become discovered'
''.format(nailgun_node['name']))
wait(lambda: self.is_node_discovered(nailgun_node),
timeout=timeout,
timeout_msg='Node {!r} failed to become discovered'
''.format(nailgun_node['name']))
@logwrap
def run_network_verify(self, cluster_id):
logger.info('Run network verification on the cluster %s', cluster_id)
return self.client.verify_networks(cluster_id)
@logwrap
def run_ostf(self, cluster_id, test_sets=None,
should_fail=0, tests_must_be_passed=None,
timeout=None, failed_test_name=None):
"""Run specified OSTF test set(s), check that all of them
or just [tests_must_be_passed] are passed"""
test_sets = test_sets or ['smoke', 'sanity']
timeout = timeout or 30 * 60
self.client.ostf_run_tests(cluster_id, test_sets)
if tests_must_be_passed:
self.assert_ostf_run_certain(
cluster_id,
tests_must_be_passed,
timeout)
else:
logger.info('Try to run assert ostf with '
'expected fail name {0}'.format(failed_test_name))
self.assert_ostf_run(
cluster_id,
should_fail=should_fail, timeout=timeout,
failed_test_name=failed_test_name, test_sets=test_sets)
@logwrap
def return_ostf_results(self, cluster_id, timeout, test_sets):
"""Filter and return OSTF results for further analysis"""
set_result_list = self._ostf_test_wait(cluster_id, timeout)
tests_res = []
for set_result in set_result_list:
for test in set_result['tests']:
if (test['testset'] in test_sets and
test['status'] != 'disabled'):
tests_res.append({test['name']: test['status']})
logger.info('OSTF test statuses are : {0}'
.format(pretty_log(tests_res, indent=1)))
return tests_res
@logwrap
def run_single_ostf_test(self, cluster_id,
test_sets=None, test_name=None,
retries=None, timeout=15 * 60):
"""Run a single OSTF test"""
self.client.ostf_run_singe_test(cluster_id, test_sets, test_name)
if retries:
return self.return_ostf_results(cluster_id, timeout=timeout,
test_sets=test_sets)
else:
self.assert_ostf_run_certain(cluster_id,
tests_must_be_passed=[test_name],
timeout=timeout)
@logwrap
def task_wait(self, task, timeout, interval=5):
logger.info('Wait for task {0} seconds: {1}'.format(
timeout, pretty_log(task, indent=1)))
start = time.time()
wait(
lambda: (self.client.get_task(task['id'])['status']
not in ('pending', 'running')),
interval=interval,
timeout=timeout,
timeout_msg='Waiting task {0!r} timeout {1} sec '
'was exceeded'.format(task['name'], timeout))
took = time.time() - start
task = self.client.get_task(task['id'])
logger.info('Task finished. Took {0} seconds. {1}'.format(
took,
pretty_log(task, indent=1)))
return task
@logwrap
def task_wait_progress(self, task, timeout, interval=5, progress=None):
logger.info('start to wait with timeout {0} '
'interval {1}'.format(timeout, interval))
wait(
lambda: self.client.get_task(
task['id'])['progress'] >= progress,
interval=interval,
timeout=timeout,
timeout_msg='Waiting task {0!r} timeout {1} sec '
'was exceeded'.format(task["name"], timeout))
return self.client.get_task(task['id'])
# TODO(ddmitriev): this method will be replaced
# after switching to fuel-devops3.0
# pylint: disable=no-self-use
def get_node_group_and_role(self, node_name, nodes_dict):
if MULTIPLE_NETWORKS:
node_roles = nodes_dict[node_name][0]
node_group = nodes_dict[node_name][1]
else:
node_roles = nodes_dict[node_name]
node_group = 'default'
return node_group, node_roles
# pylint: enable=no-self-use
@logwrap
def update_nodes(self, cluster_id, nodes_dict,
pending_addition=True, pending_deletion=False,
update_nodegroups=False, custom_names=None,
update_interfaces=True):
failed_nodes = {}
for node_name, node_roles in nodes_dict.items():
try:
self.environment.d_env.get_node(name=node_name)
except DevopsObjNotFound:
failed_nodes[node_name] = node_roles
if failed_nodes:
text = 'Some nodes is inaccessible:\n'
for node_name, node_roles in sorted(failed_nodes.items()):
text += '\t{name} for roles: {roles!s}\n'.format(
name=node_name,
roles=['{}'.format(node) for node in sorted(node_roles)])
text += 'Impossible to continue!'
logger.error(text)
raise KeyError(sorted(list(failed_nodes.keys())))
# update nodes in cluster
nodes_data = []
nodes_groups = {}
updated_nodes = []
for node_name in nodes_dict:
devops_node = self.environment.d_env.get_node(name=node_name)
node_group, node_roles = self.get_node_group_and_role(node_name,
nodes_dict)
wait(lambda:
self.get_nailgun_node_by_devops_node(devops_node)['online'],
timeout=60 * 2)
node = self.get_nailgun_node_by_devops_node(devops_node)
assert_true(node['online'],
'Node {0} is offline'.format(node['mac']))
if custom_names:
name = custom_names.get(node_name,
'{}_{}'.format(
node_name,
"_".join(node_roles)))
else:
name = '{0}_{1}'.format(node_name, "_".join(node_roles))
node_data = {
'cluster_id': cluster_id,
'id': node['id'],
'pending_addition': pending_addition,
'pending_deletion': pending_deletion,
'pending_roles': node_roles,
'name': name
}
nodes_data.append(node_data)
if node_group not in nodes_groups.keys():
nodes_groups[node_group] = []
nodes_groups[node_group].append(node)
updated_nodes.append(node)
# assume nodes are going to be updated for one cluster only
cluster_id = nodes_data[-1]['cluster_id']
node_ids = [str(node_info['id']) for node_info in nodes_data]
self.client.update_nodes(nodes_data)
nailgun_nodes = self.client.list_cluster_nodes(cluster_id)
cluster_node_ids = [str(_node['id']) for _node in nailgun_nodes]
assert_true(
all([node_id in cluster_node_ids for node_id in node_ids]))
if update_interfaces and not pending_deletion:
self.update_nodes_interfaces(cluster_id, updated_nodes)
if update_nodegroups:
self.update_nodegroups(cluster_id=cluster_id,
node_groups=nodes_groups)
return nailgun_nodes
@logwrap
def delete_node(self, node_id, interval=30, timeout=600):
task = self.client.delete_node(node_id)
logger.debug("task info is {}".format(task))
self.assert_task_success(task, interval=interval, timeout=timeout)
@logwrap
def update_node_networks(self, node_id, interfaces_dict,
raw_data=None,
override_ifaces_params=None):
interfaces = self.client.get_node_interfaces(node_id)
if raw_data is not None:
interfaces.extend(raw_data)
def get_bond_ifaces():
# Filter out all interfaces to be bonded
ifaces = []
for bond in [i for i in interfaces if i['type'] == 'bond']:
ifaces.extend(s['name'] for s in bond['slaves'])
return ifaces
# fuelweb_admin is always on 1st iface unless the iface is not bonded
iface = iface_alias('eth0')
if iface not in get_bond_ifaces():
interfaces_dict[iface] = interfaces_dict.get(iface,
[])
if 'fuelweb_admin' not in interfaces_dict[iface]:
interfaces_dict[iface].append('fuelweb_admin')
def get_iface_by_name(ifaces, name):
iface = [_iface for _iface in ifaces if _iface['name'] == name]
assert_true(len(iface) > 0,
"Interface with name {} is not present on "
"node. Please check override params.".format(name))
return iface[0]
if override_ifaces_params is not None:
for interface in override_ifaces_params:
get_iface_by_name(interfaces, interface['name']).\
update(interface)
all_networks = dict()
for interface in interfaces:
all_networks.update(
{net['name']: net for net in interface['assigned_networks']})
for interface in interfaces:
name = interface["name"]
interface['assigned_networks'] = \
[all_networks[i] for i in interfaces_dict.get(name, []) if
i in all_networks.keys()]
self.client.put_node_interfaces(
[{'id': node_id, 'interfaces': interfaces}])
@logwrap
def update_node_disk(self, node_id, disks_dict):
disks = self.client.get_node_disks(node_id)
for disk in disks:
dname = disk['name']
if dname not in disks_dict:
continue
for volume in disk['volumes']:
vname = volume['name']
if vname in disks_dict[dname]:
volume['size'] = disks_dict[dname][vname]
self.client.put_node_disks(node_id, disks)
@logwrap
def get_node_disk_size(self, node_id, disk_name):
disks = self.client.get_node_disks(node_id)
size = 0
for disk in disks:
if disk['name'] == disk_name:
for volume in disk['volumes']:
size += volume['size']
return size
def get_node_partition_size(self, node_id, partition_name):
disks = self.client.get_node_disks(node_id)
size = 0
logger.debug('Disks of node-{}: \n{}'.format(node_id,
pretty_log(disks)))
for disk in disks:
for volume in disk['volumes']:
if volume['name'] == partition_name:
size += volume['size']
return size
@logwrap
def update_node_partitioning(self, node, disk='vdc',
node_role='cinder', unallocated_size=11116):
node_size = self.get_node_disk_size(node['id'], disk)
disk_part = {
disk: {
node_role: node_size - unallocated_size
}
}
self.update_node_disk(node['id'], disk_part)
return node_size - unallocated_size
@logwrap
def update_vlan_network_fixed(
self, cluster_id, amount=1, network_size=256):
self.client.update_network(
cluster_id,
networking_parameters={
"net_manager": help_data.NETWORK_MANAGERS['vlan'],
"fixed_network_size": network_size,
"fixed_networks_amount": amount
}
)
@retry(count=2, delay=20)
@logwrap
def verify_network(self, cluster_id, timeout=60 * 5, success=True):
def _report_verify_network_result(task):
# Report verify_network results using style like on UI
if task['status'] == 'error' and 'result' in task:
msg = "Network verification failed:\n"
if task['result']:
msg += ("{0:30} | {1:20} | {2:15} | {3}\n"
.format("Node Name", "Node MAC address",
"Node Interface",
"Expected VLAN (not received)"))
for res in task['result']:
name = None
mac = None
interface = None
absent_vlans = []
if 'name' in res:
name = res['name']
if 'mac' in res:
mac = res['mac']
if 'interface' in res:
interface = res['interface']
if 'absent_vlans' in res:
absent_vlans = res['absent_vlans']
msg += ("{0:30} | {1:20} | {2:15} | {3}\n".format(
name or '-', mac or '-', interface or '-',
[x or 'untagged' for x in absent_vlans]))
logger.error(''.join([msg, task['message']]))
# TODO(apanchenko): remove this hack when network verification begins
# TODO(apanchenko): to work for environments with multiple net groups
if len(self.client.get_nodegroups()) > 1:
logger.warning('Network verification is temporary disabled when '
'"multiple cluster networks" feature is used')
return
try:
task = self.run_network_verify(cluster_id)
with QuietLogger():
if success:
self.assert_task_success(task, timeout, interval=10)
else:
self.assert_task_failed(task, timeout, interval=10)
logger.info("Network verification of cluster {0} finished"
.format(cluster_id))
except AssertionError:
# Report the result of network verify.
task = self.client.get_task(task['id'])
_report_verify_network_result(task)
raise
@logwrap
def update_nodes_interfaces(self, cluster_id, nailgun_nodes=None):
if nailgun_nodes is None:
nailgun_nodes = []
net_provider = self.client.get_cluster(cluster_id)['net_provider']
if NEUTRON == net_provider:
assigned_networks = {
iface_alias('eth0'): ['fuelweb_admin'],
iface_alias('eth1'): ['public'],
iface_alias('eth2'): ['management'],
iface_alias('eth3'): ['private'],
iface_alias('eth4'): ['storage'],
}
else:
assigned_networks = {
iface_alias('eth1'): ['public'],
iface_alias('eth2'): ['management'],
iface_alias('eth3'): ['fixed'],
iface_alias('eth4'): ['storage'],
}
baremetal_iface = iface_alias('eth5')
if self.get_cluster_additional_components(cluster_id).get(
'ironic', False):
assigned_networks[baremetal_iface] = ['baremetal']
logger.info('Assigned networks are: {}'.format(str(assigned_networks)))
if not nailgun_nodes:
nailgun_nodes = self.client.list_cluster_nodes(cluster_id)
for node in nailgun_nodes:
self.update_node_networks(node['id'], assigned_networks)
@logwrap
def get_offloading_modes(self, node_id, interfaces):
offloading_types = []
for i in self.client.get_node_interfaces(node_id):
for interface in interfaces:
if i['name'] == interface:
for offloading_type in i['offloading_modes']:
offloading_types.append(offloading_type['name'])
return offloading_types
@logwrap
def update_offloads(self, node_id, update_values, interface_to_update):
interfaces = self.client.get_node_interfaces(node_id)
for i in interfaces:
if i['name'] == interface_to_update:
for new_mode in update_values['offloading_modes']:
for mode in i['offloading_modes']:
if mode['name'] == new_mode['name']:
mode.update(new_mode)
break
else:
raise Exception("Offload type '{0}' is not applicable"
" for interface {1}".format(
new_mode['name'],
interface_to_update))
self.client.put_node_interfaces(
[{'id': node_id, 'interfaces': interfaces}])
def change_default_network_settings(self):
def fetch_networks(networks):
"""Parse response from api/releases/1/networks and return dict with
networks' settings - need for avoiding hardcode"""
result = {}
for net in networks:
if (net['name'] == 'private' and
net.get('seg_type', '') == 'tun'):
result['private_tun'] = net
elif (net['name'] == 'private' and
net.get('seg_type', '') == 'gre'):
result['private_gre'] = net
elif net['name'] == 'public':
result['public'] = net
elif net['name'] == 'management':
result['management'] = net
elif net['name'] == 'storage':
result['storage'] = net
elif net['name'] == 'baremetal':
result['baremetal'] = net
return result
default_networks = {}
for n in ('public', 'management', 'storage', 'private'):
if self.environment.d_env.get_networks(name=n):
default_networks[n] = self.environment.d_env.get_network(
name=n).ip
logger.info("Applying default network settings")
for _release in self.client.get_releases():
if (_release['is_deployable'] is False and
_release['state'] != 'available'):
logger.info("Release {!r} (version {!r}) is not available for "
"deployment; skipping default network "
"replacement".format(_release['name'],
_release['version']))
continue
logger.info(
'Applying changes for release: {}'.format(
_release['name']))
net_settings = \
self.client.get_release_default_net_settings(
_release['id'])
for net_provider in NETWORK_PROVIDERS:
if net_provider not in net_settings:
continue
networks = fetch_networks(
net_settings[net_provider]['networks'])
networks['public']['cidr'] = str(default_networks['public'])
networks['public']['gateway'] = str(
default_networks['public'].network + 1)
networks['public']['notation'] = 'ip_ranges'
# use the first half of public network as static public range
networks['public']['ip_range'] = self.get_range(
default_networks['public'], ip_range=-1)[0]
# use the second half of public network as floating range
net_settings[net_provider]['config']['floating_ranges'] = \
self.get_range(default_networks['public'], ip_range=1)
devops_env = self.environment.d_env
# NOTE(akostrikov) possible break.
if 'baremetal' in networks and \
devops_env.get_networks(name='ironic'):
ironic_net = self.environment.d_env.get_network(
name='ironic').ip
prefix = netaddr.IPNetwork(
str(ironic_net.cidr)
).prefixlen
subnet1, subnet2 = tuple(ironic_net.subnet(prefix + 1))
networks['baremetal']['cidr'] = str(ironic_net)
net_settings[net_provider]['config'][
'baremetal_gateway'] = str(ironic_net[-2])
networks['baremetal']['ip_range'] = [
str(subnet1[2]), str(subnet2[0])]
net_settings[net_provider]['config']['baremetal_range'] =\
[str(subnet2[1]), str(subnet2[-3])]
networks['baremetal']['vlan_start'] = None
if BONDING:
# leave defaults for mgmt, storage and private if
# BONDING is enabled
continue
for net, cidr in default_networks.items():
if net in ('public', 'private'):
continue
networks[net]['cidr'] = str(cidr)
networks[net]['ip_range'] = self.get_range(cidr)[0]
networks[net]['notation'] = 'ip_ranges'
networks[net]['vlan_start'] = None
if net_provider == 'neutron':
networks['private_tun']['cidr'] = str(
default_networks['private'])
networks['private_gre']['cidr'] = str(
default_networks['private'])
net_settings[net_provider]['config']['internal_cidr'] = \
str(default_networks['private'])
net_settings[net_provider]['config']['internal_gateway'] =\
str(default_networks['private'][1])
elif net_provider == 'nova_network':
net_settings[net_provider]['config'][
'fixed_networks_cidr'] = str(
default_networks['private'])
self.client.put_release_default_net_settings(
_release['id'], net_settings)
@logwrap
def update_nodegroups_network_configuration(self, cluster_id,
nodegroups=None):
net_config = self.client.get_networks(cluster_id)
new_settings = net_config
for nodegroup in nodegroups:
logger.info('Update network settings of cluster %s, '
'nodegroup %s', cluster_id, nodegroup['name'])
new_settings = self.update_nodegroup_net_settings(new_settings,
nodegroup,
cluster_id)
self.client.update_network(
cluster_id=cluster_id,
networking_parameters=new_settings["networking_parameters"],
networks=new_settings["networks"]
)
@staticmethod
def _get_true_net_name(name, net_pools):
"""Find a devops network name in net_pools"""
for net in net_pools:
if name in net:
return {name: net_pools[net]}
def update_nodegroup_net_settings(self, network_configuration, nodegroup,
cluster_id=None):
seg_type = network_configuration.get('networking_parameters', {}) \
.get('segmentation_type')
nodegroup_id = self.get_nodegroup(cluster_id, nodegroup['name'])['id']
for net in network_configuration.get('networks'):
if net['group_id'] == nodegroup_id:
# Do not overwrite default PXE admin network configuration
if nodegroup['name'] == 'default' and \
net['name'] == 'fuelweb_admin':
continue
self.set_network(net_config=net,
net_name=net['name'],
net_devices=nodegroup['networks'],
seg_type=seg_type)
# For all admin/pxe networks except default use master
# node as router
# TODO(mstrukov): find way to get admin node networks only
if net['name'] != 'fuelweb_admin':
continue
for devops_network in self.environment.d_env.get_networks():
if str(devops_network.ip_network) == net['cidr']:
net['gateway'] = \
self.environment.d_env.nodes().\
admin.get_ip_address_by_network_name(
devops_network.name)
logger.info('Set master node ({0}) as '
'router for admin network '
'in nodegroup {1}.'.format(
net['gateway'], nodegroup_id))
return network_configuration
def set_network(self, net_config, net_name, net_devices=None,
seg_type=None):
nets_wo_floating = ['public', 'management', 'storage', 'baremetal']
if (seg_type == NEUTRON_SEGMENT['tun'] or
seg_type == NEUTRON_SEGMENT['gre']):
nets_wo_floating.append('private')
if not net_devices:
if not BONDING:
if 'floating' == net_name:
self.net_settings(net_config, 'public', floating=True)
elif net_name in nets_wo_floating:
self.net_settings(net_config, net_name)
else:
ip_obj = self.environment.d_env.get_network(name="public").ip
pub_subnets = list(ip_obj.subnet(new_prefix=27))
if "floating" == net_name:
self.net_settings(net_config, pub_subnets[0],
floating=True, jbond=True)
elif net_name in nets_wo_floating:
i = nets_wo_floating.index(net_name)
self.net_settings(net_config, pub_subnets[i], jbond=True)
else:
if not BONDING:
if 'floating' == net_name:
self.net_settings(net_config, net_devices['public'],
floating=True)
self.net_settings(net_config, net_devices[net_name])
else:
ip_obj = self.environment.d_env.get_network(
name=net_devices['public']).ip
pub_subnets = list(ip_obj.subnet(new_prefix=27))
if "floating" == net_name:
self.net_settings(net_config, pub_subnets[0],
floating=True, jbond=True)
elif net_name in nets_wo_floating:
i = nets_wo_floating.index(net_name)
self.net_settings(net_config, pub_subnets[i], jbond=True)
elif net_name in 'fuelweb_admin':
self.net_settings(net_config, net_devices['fuelweb_admin'])
if 'ip_ranges' in net_config:
if net_config['ip_ranges']:
net_config['meta']['notation'] = 'ip_ranges'
def net_settings(self, net_config, net_name, floating=False, jbond=False):
if jbond:
if net_config['name'] == 'public':
net_config['gateway'] = self.environment.d_env.router('public')
ip_network = net_name
elif net_config['name'] == 'baremetal':
baremetal_net = self.environment.d_env.get_network(
name='ironic').ip_network
net_config['gateway'] = str(
list(netaddr.IPNetwork(str(baremetal_net)))[-2])
ip_network = baremetal_net
else:
ip_network = net_name
else:
net_config['vlan_start'] = None
if net_config['name'] == 'baremetal':
baremetal_net = self.environment.d_env.get_network(
name='ironic').ip_network
net_config['gateway'] = str(
list(netaddr.IPNetwork(str(baremetal_net)))[-2])
ip_network = baremetal_net
else:
net_config['gateway'] = self.environment.d_env.router(net_name)
ip_network = self.environment.d_env.get_network(
name=net_name).ip_network
net_config['cidr'] = str(ip_network)
if 'admin' in net_config['name']:
net_config['ip_ranges'] = self.get_range(ip_network, 2)
elif floating:
net_config['ip_ranges'] = self.get_range(ip_network, 1)
else:
net_config['ip_ranges'] = self.get_range(ip_network, -1)
@staticmethod
def get_range(ip_network, ip_range=0):
net = list(netaddr.IPNetwork(str(ip_network)))
half = len(net) // 2
if ip_range == 0:
return [[str(net[2]), str(net[-2])]]
elif ip_range == 1:
return [[str(net[half]), str(net[-2])]]
elif ip_range == -1:
return [[str(net[2]), str(net[half - 1])]]
elif ip_range == 2:
return [[str(net[3]), str(net[half - 1])]]
elif ip_range == 3:
return [[str(net[half]), str(net[-3])]]
def get_floating_ranges(self, network_set=''):
net_name = 'public{0}'.format(network_set)
net = list(self.environment.d_env.get_network(name=net_name).ip)
ip_ranges, expected_ips = [], []
for i in [0, -20, -40]:
l = []
for k in range(11):
l.append(str(net[-12 + i + k]))
expected_ips.append(l)
e, s = str(net[-12 + i]), str(net[-2 + i])
ip_ranges.append([e, s])
return ip_ranges, expected_ips
def warm_shutdown_nodes(self, devops_nodes, timeout=4 * 60):
logger.info('Shutting down (warm) nodes %s',
[n.name for n in devops_nodes])
for node in devops_nodes:
logger.debug('Shutdown node %s', node.name)
with self.get_ssh_for_node(node.name) as remote:
remote.check_call('/sbin/shutdown -Ph now')
for node in devops_nodes:
self.wait_node_is_offline(node, timeout=timeout)
node.destroy()
def warm_start_nodes(self, devops_nodes, timeout=4 * 60):
logger.info('Starting nodes %s', [n.name for n in devops_nodes])
for node in devops_nodes:
node.start()
self.wait_nodes_get_online_state(devops_nodes, timeout=timeout)
def warm_restart_nodes(self, devops_nodes, timeout=4 * 60):
logger.info('Reboot (warm restart) nodes %s',
[n.name for n in devops_nodes])
self.warm_shutdown_nodes(devops_nodes, timeout=timeout)
self.warm_start_nodes(devops_nodes, timeout=timeout)
def cold_restart_nodes(self, devops_nodes,
wait_offline=True, wait_online=True,
wait_after_destroy=None, timeout=4 * 60):
logger.info('Cold restart nodes %s',
[n.name for n in devops_nodes])
for node in devops_nodes:
logger.info('Destroy node %s', node.name)
node.destroy()
for node in devops_nodes:
if wait_offline:
self.wait_node_is_offline(node, timeout=timeout)
if wait_after_destroy:
time.sleep(wait_after_destroy)
for node in devops_nodes:
logger.info('Start %s node', node.name)
node.start()
if wait_online:
for node in devops_nodes:
self.wait_node_is_online(node, timeout=timeout)
self.environment.sync_time()
@logwrap
def ip_address_show(self, node_name, interface, namespace=None):
"""Return ip on interface in node with node_name inside namespace
:type node_name: String
:type namespace: String
:type interface: String
:rtype: String on None
"""
try:
if namespace:
cmd = 'ip netns exec {0} ip -4 ' \
'-o address show {1}'.format(namespace, interface)
else:
cmd = 'ip -4 -o address show {1}'.format(interface)
with self.get_ssh_for_node(node_name) as remote:
ret = remote.check_call(cmd)
ip_search = re.search(
'inet (?P<ip>\d+\.\d+\.\d+.\d+/\d+).*scope .* '
'{0}'.format(interface), ' '.join(ret['stdout']))
if ip_search is None:
logger.debug("Ip show output does not match in regex. "
"Current value is None. On node {0} in netns "
"{1} for interface {2}".format(node_name,
namespace,
interface))
return None
return ip_search.group('ip')
except DevopsCalledProcessError as err:
logger.error(err)
return None
@logwrap
def ip_address_del(self, node_name, namespace, interface, ip):
logger.info('Delete %s ip address of %s interface at %s node',
ip, interface, node_name)
with self.get_ssh_for_node(node_name) as remote:
remote.check_call(
'ip netns exec {0} ip addr'
' del {1} dev {2}'.format(namespace, ip, interface))
@logwrap
def provisioning_cluster_wait(self, cluster_id, progress=None):
logger.info('Start cluster #%s provisioning', cluster_id)
task = self.client.provision_nodes(cluster_id)
self.assert_task_success(task, progress=progress)
@logwrap
def deploy_custom_graph_wait(self,
cluster_id,
graph_type,
node_ids=None,
progress=None):
"""Deploy custom graph of a given type.
:param cluster_id: Id of a cluster to deploy
:param graph_type: Custom graph type to deploy
:param node_ids: Ids of nodes to deploy. None means all
:param progress: Progress at which count deployment as a success.
"""
logger.info('Start cluster #{cid} custom type "{type}" '
'graph deployment on nodes: {nodes}. '
'None means on all nodes.'.format(
cid=cluster_id,
type=graph_type,
nodes=node_ids
))
task = self.client.deploy_custom_graph(cluster_id,
graph_type,
node_ids)
self.assert_task_success(task, progress=progress)
@logwrap
def deploy_task_wait(self, cluster_id, progress=None):
logger.info('Start cluster #%s deployment', cluster_id)
task = self.client.deploy_nodes(cluster_id)
self.assert_task_success(
task, progress=progress)
@logwrap
def stop_deployment_wait(self, cluster_id):
logger.info('Stop cluster #%s deployment', cluster_id)
task = self.client.stop_deployment(cluster_id)
self.assert_task_success(task, timeout=50 * 60, interval=30)
@logwrap
def stop_reset_env_wait(self, cluster_id):
logger.info('Reset cluster #%s', cluster_id)
task = self.client.reset_environment(cluster_id)
self.assert_task_success(task, timeout=50 * 60, interval=30)
@logwrap
def delete_env_wait(self, cluster_id, timeout=10 * 60):
logger.info('Removing cluster with id={0}'.format(cluster_id))
self.client.delete_cluster(cluster_id)
tasks = self.client.get_tasks()
delete_tasks = [t for t in tasks if t['status']
in ('pending', 'running') and
t['name'] == 'cluster_deletion' and
t['cluster'] == cluster_id]
if delete_tasks:
for task in delete_tasks:
logger.info('Task found: {}'.format(task))
task = delete_tasks[0]
logger.info('Selected task: {}'.format(task))
# Task will be removed with the cluster, so we will get 404 error
assert_raises(
exceptions.NotFound,
self.assert_task_success, task, timeout)
else:
assert 'No cluster_deletion task found!'
@logwrap
def wait_nodes_get_online_state(self, nodes, timeout=4 * 60):
for node in nodes:
self.wait_node_is_online(node, timeout=timeout)
@logwrap
def wait_mysql_galera_is_up(self, node_names, timeout=60 * 4):
def _get_galera_status(_remote):
cmd = ("mysql --connect_timeout=5 -sse \"SELECT VARIABLE_VALUE "
"FROM information_schema.GLOBAL_STATUS WHERE VARIABLE_NAME"
" = 'wsrep_ready';\"")
result = _remote.execute(cmd)
if result['exit_code'] == 0:
return ''.join(result['stdout']).strip()
else:
return ''.join(result['stderr']).strip()
for node_name in node_names:
with self.get_ssh_for_node(node_name) as remote:
try:
wait(lambda: _get_galera_status(remote) == 'ON',
timeout=timeout)
logger.info("MySQL Galera is up on {host} node.".format(
host=node_name))
except TimeoutError:
logger.error("MySQL Galera isn't ready on {0}: {1}"
.format(node_name,
_get_galera_status(remote)))
raise TimeoutError(
"MySQL Galera isn't ready on {0}: {1}".format(
node_name, _get_galera_status(remote)))
return True
@logwrap
def wait_cinder_is_up(self, node_names):
logger.info("Waiting for all Cinder services up.")
for node_name in node_names:
node = self.get_nailgun_node_by_name(node_name)
wait(lambda: checkers.check_cinder_status(node['ip']),
timeout=300,
timeout_msg='Cinder services not ready')
logger.info("All Cinder services up.")
return True
def run_ostf_repeatably(self, cluster_id, test_name=None,
test_retries=None, checks=None):
res = []
passed_count = []
failed_count = []
test_name_to_run = test_name or OSTF_TEST_NAME
retries = test_retries or OSTF_TEST_RETRIES_COUNT
test_path = ostf_test_mapping.OSTF_TEST_MAPPING.get(test_name_to_run)
logger.info('Test path is {0}'.format(test_path))
for _ in range(retries):
result = self.run_single_ostf_test(
cluster_id=cluster_id, test_sets=['smoke', 'sanity'],
test_name=test_path,
retries=True)
res.append(result)
logger.info('res is {0}'.format(res))
logger.info('full res is {0}'.format(res))
for element in res:
for test in element:
if test.get(test_name) == 'success':
passed_count.append(test)
elif test.get(test_name) in {'failure', 'error'}:
failed_count.append(test)
if not checks:
assert_true(
len(passed_count) == test_retries,
'not all retries were successful,'
' fail {0} retries'.format(len(failed_count)))
else:
return failed_count
def get_nailgun_version(self):
response = self.client.get_api_version()
logger.info("ISO version: {}".format(pretty_log(
response, indent=1)))
return response
@logwrap
def run_ceph_task(self, cluster_id, offline_nodes):
ceph_id = [n['id'] for n in self.client.list_cluster_nodes(cluster_id)
if 'ceph-osd' in n['roles'] and
n['id'] not in offline_nodes]
res = self.client.put_deployment_tasks_for_cluster(
cluster_id, data=['top-role-ceph-osd'],
node_id=str(ceph_id).strip('[]'))
logger.debug('res info is {0}'.format(res))
self.assert_task_success(task=res)
@retry(count=3)
def check_ceph_time_skew(self, cluster_id, offline_nodes):
ceph_nodes = self.get_nailgun_cluster_nodes_by_roles(
cluster_id, ['ceph-osd'])
online_ceph_nodes = [
n for n in ceph_nodes if n['id'] not in offline_nodes]
# Let's find nodes where are a time skew. It can be checked on
# an arbitrary one.
logger.debug("Looking up nodes with a time skew and try to fix them")
with self.environment.d_env.get_ssh_to_remote(
online_ceph_nodes[0]['ip']) as remote:
if ceph.is_clock_skew(remote):
skewed = ceph.get_node_fqdns_w_clock_skew(remote)
logger.warning("Time on nodes {0} are to be "
"re-synchronized".format(skewed))
nodes_to_sync = [
n for n in online_ceph_nodes
if n['fqdn'].split('.')[0] in skewed]
self.environment.sync_time(nodes_to_sync)
try:
wait(lambda: not ceph.is_clock_skew(remote),
timeout=120)
except TimeoutError:
skewed = ceph.get_node_fqdns_w_clock_skew(remote)
logger.error("Time on Ceph nodes {0} is still skewed. "
"Restarting Ceph monitor on these "
"nodes".format(', '.join(skewed)))
for node in skewed:
fqdn = self.get_fqdn_by_hostname(node)
d_node = self.get_devops_node_by_nailgun_fqdn(fqdn)
logger.debug("Establish SSH connection to first Ceph "
"monitor node %s", fqdn)
with self.get_ssh_for_node(d_node.name) as remote_to_mon:
logger.debug("Restart Ceph monitor service "
"on node %s", fqdn)
ceph.restart_monitor(remote_to_mon)
wait(lambda: not ceph.is_clock_skew(remote), timeout=120,
timeout_msg='check ceph time skew timeout')
@logwrap
def check_ceph_status(self, cluster_id, offline_nodes=(),
recovery_timeout=360):
ceph_nodes = self.get_nailgun_cluster_nodes_by_roles(
cluster_id, ['ceph-osd'])
online_ceph_nodes = [
n for n in ceph_nodes if n['id'] not in offline_nodes]
logger.info('Waiting until Ceph service become up...')
for node in online_ceph_nodes:
with self.environment.d_env\
.get_ssh_to_remote(node['ip']) as remote:
wait(lambda: ceph.check_service_ready(remote) is True,
interval=20, timeout=600,
timeout_msg='Ceph service is not properly started'
' on {0}'.format(node['name']))
logger.info('Ceph service is ready. Checking Ceph Health...')
self.check_ceph_time_skew(cluster_id, offline_nodes)
node = online_ceph_nodes[0]
with self.environment.d_env.get_ssh_to_remote(node['ip']) as remote:
if not ceph.is_health_ok(remote):
if ceph.is_pgs_recovering(remote) and len(offline_nodes) > 0:
logger.info('Ceph is being recovered after osd node(s)'
' shutdown.')
try:
wait(lambda: ceph.is_health_ok(remote),
interval=30, timeout=recovery_timeout)
except TimeoutError:
result = ceph.health_detail(remote)
msg = 'Ceph HEALTH is not OK on {0}. Details: {1}'\
.format(node['name'], result)
logger.error(msg)
raise TimeoutError(msg)
else:
result = ceph.health_detail(remote)
msg = 'Ceph HEALTH is not OK on {0}. Details: {1}'.format(
node['name'], result)
assert_true(ceph.is_health_ok(remote), msg)
logger.info('Checking Ceph OSD Tree...')
ceph.check_disks(remote, [n['id'] for n in online_ceph_nodes])
logger.info('Ceph cluster status is OK')
@logwrap
def get_releases_list_for_os(self, release_name, release_version=None):
full_list = self.client.get_releases()
release_ids = []
for release in full_list:
if release_version:
if release_name in release['name'].lower() \
and release_version == release['version']:
logger.debug('release data is {0}'.format(release))
release_ids.append(release['id'])
else:
if release_name in release['name'].lower():
release_ids.append(release['id'])
return release_ids
@logwrap
def get_next_deployable_release_id(self, release_id):
releases = self.client.get_releases()
release_details = self.client.get_releases_details(release_id)
for release in releases:
if (release["id"] > release_id and
release["operating_system"] ==
release_details["operating_system"] and
release["is_deployable"] and
OPENSTACK_RELEASE in release["name"].lower()):
return release["id"]
return None
@logwrap
def update_cluster(self, cluster_id, data):
logger.debug(
"Try to update cluster with data {0}".format(data))
self.client.update_cluster(cluster_id, data)
@logwrap
def run_update(self, cluster_id, timeout, interval):
logger.info("Run update..")
task = self.client.run_update(cluster_id)
logger.debug("Invocation of update runs with result {0}".format(task))
self.assert_task_success(task, timeout=timeout, interval=interval)
@logwrap
def get_cluster_release_id(self, cluster_id):
data = self.client.get_cluster(cluster_id)
return data['release_id']
def assert_nodes_in_ready_state(self, cluster_id):
for nailgun_node in self.client.list_cluster_nodes(cluster_id):
assert_equal(nailgun_node['status'], 'ready',
'Nailgun node status is not ready but {0}'.format(
nailgun_node['status']))
@staticmethod
@logwrap
def modify_python_file(remote, modification, filename):
remote.execute('sed -i "{0}" {1}'.format(modification, filename))
@staticmethod
def backup_master(remote):
# FIXME(kozhukalov): This approach is outdated
# due to getting rid of docker containers.
logger.info("Backup of the master node is started.")
remote.check_call(
"echo CALC_MY_MD5SUM > /etc/fuel/data",
error_info='command calc_my_mdsum failed')
remote.check_call(
"iptables-save > /etc/fuel/iptables-backup",
error_info='can not save iptables in iptables-backup')
remote.check_call(
"md5sum /etc/fuel/data | cut -d' ' -f1 > /etc/fuel/sum",
error_info='failed to create sum file')
remote.check_call('dockerctl backup')
remote.check_call(
'rm -f /etc/fuel/data',
error_info='Can not remove /etc/fuel/data')
logger.info("Backup of the master node is complete.")
@logwrap
def restore_master(self, ip):
# FIXME(kozhukalov): This approach is outdated
# due to getting rid of docker containers.
logger.info("Restore of the master node is started.")
path = checkers.find_backup(ip)
self.ssh_manager.execute_on_remote(
ip=ip,
cmd='dockerctl restore {0}'.format(path))
logger.info("Restore of the master node is complete.")
@logwrap
def restore_check_nailgun_api(self):
logger.info("Restore check nailgun api")
info = self.client.get_api_version()
os_version = info["openstack_version"]
assert_true(os_version, 'api version returned empty data')
@logwrap
def get_nailgun_cidr_nova(self, cluster_id):
return self.client.get_networks(cluster_id).\
get("networking_parameters").get("fixed_networks_cidr")
@logwrap
def get_nailgun_cidr_neutron(self, cluster_id):
return self.client.get_networks(cluster_id).\
get("networking_parameters").get("internal_cidr")
@logwrap
def check_fixed_network_cidr(self, cluster_id, os_conn):
net_provider = self.client.get_cluster(cluster_id)['net_provider']
if net_provider == 'nova_network':
nailgun_cidr = self.get_nailgun_cidr_nova(cluster_id)
logger.debug('nailgun cidr is {0}'.format(nailgun_cidr))
net = os_conn.nova_get_net('novanetwork')
logger.debug('nova networks: {0}'.format(
net))
assert_equal(nailgun_cidr, net.cidr.rstrip(),
'Cidr after deployment is not equal'
' to cidr by default')
elif net_provider == 'neutron':
nailgun_cidr = self.get_nailgun_cidr_neutron(cluster_id)
logger.debug('nailgun cidr is {0}'.format(nailgun_cidr))
private_net_name = self.get_cluster_predefined_networks_name(
cluster_id)['private_net']
subnet = os_conn.get_subnet('{0}__subnet'.format(private_net_name))
logger.debug('subnet of pre-defined fixed network: {0}'.format(
subnet))
assert_true(subnet, '{0}__subnet does not exists'.format(
private_net_name))
logger.debug('cidr {0}__subnet: {1}'.format(
private_net_name, subnet['cidr']))
assert_equal(nailgun_cidr, subnet['cidr'].rstrip(),
'Cidr after deployment is not equal'
' to cidr by default')
@staticmethod
@logwrap
def check_fixed_nova_splited_cidr(os_conn, nailgun_cidr):
logger.debug('Nailgun cidr for nova: {0}'.format(nailgun_cidr))
subnets_list = [net.cidr for net in os_conn.get_nova_network_list()]
logger.debug('Nova subnets list: {0}'.format(subnets_list))
# Check that all subnets are included in nailgun_cidr
for subnet in subnets_list:
logger.debug("Check that subnet {0} is part of network {1}"
.format(subnet, nailgun_cidr))
assert_true(netaddr.IPNetwork(str(subnet)) in
netaddr.IPNetwork(str(nailgun_cidr)),
'Something goes wrong. Seems subnet {0} is out '
'of net {1}'.format(subnet, nailgun_cidr))
# Check that any subnet doesn't include any other subnet
subnets_pairs = [(subnets_list[x1], subnets_list[x2])
for x1 in range(len(subnets_list))
for x2 in range(len(subnets_list))
if x1 != x2]
for subnet1, subnet2 in subnets_pairs:
logger.debug("Check if the subnet {0} is part of the subnet {1}"
.format(subnet1, subnet2))
assert_true(netaddr.IPNetwork(str(subnet1)) not in
netaddr.IPNetwork(str(subnet2)),
"Subnet {0} is part of subnet {1}"
.format(subnet1, subnet2))
def update_internal_network(self, cluster_id, cidr, gateway=None):
net_provider = self.client.get_cluster(cluster_id)['net_provider']
net_config = self.client.get_networks(cluster_id)
data = (cluster_id, net_config["networking_parameters"],
net_config["networks"])
if net_provider == 'nova_network':
net_config["networking_parameters"]['fixed_networks_cidr']\
= cidr
self.client.update_network(*data)
elif net_provider == 'neutron':
net_config["networking_parameters"]['internal_cidr']\
= cidr
net_config["networking_parameters"]['internal_gateway']\
= gateway
self.client.update_network(*data)
def get_cluster_mode(self, cluster_id):
return self.client.get_cluster(cluster_id)['mode']
def get_public_ip(self, cluster_id):
# Find a controller and get it's IP for public network
network_data = [
node['network_data']
for node in self.client.list_cluster_nodes(cluster_id)
if "controller" in node['roles']][0]
pub_ip = [net['ip'] for net in network_data
if "public" in net['name']][0]
return pub_ip.split('/')[0]
def get_public_vip(self, cluster_id):
if self.get_cluster_mode(cluster_id) == DEPLOYMENT_MODE_HA:
return self.client.get_networks(
cluster_id)['vips']['public']['ipaddr']
else:
logger.error("Public VIP for cluster '{0}' not found, searching "
"for public IP on the controller".format(cluster_id))
ip = self.get_public_ip(cluster_id)
logger.info("Public IP found: {0}".format(ip))
return ip
def get_management_vrouter_vip(self, cluster_id):
return self.client.get_networks(
cluster_id)['vips']['vrouter']['ipaddr']
def get_mgmt_vip(self, cluster_id):
return self.client.get_networks(
cluster_id)['vips']['management']['ipaddr']
def get_public_vrouter_vip(self, cluster_id):
return self.client.get_networks(
cluster_id)['vips']['vrouter_pub']['ipaddr']
@logwrap
def get_controller_with_running_service(self, slave, service_name):
ret = self.get_pacemaker_status(slave.name)
logger.debug("pacemaker status is {0}".format(ret))
node_name = re.search(service_name, ret).group(1)
logger.debug("node name is {0}".format(node_name))
fqdn = self.get_fqdn_by_hostname(node_name)
devops_node = self.find_devops_node_by_nailgun_fqdn(
fqdn, self.environment.d_env.nodes().slaves)
return devops_node
@staticmethod
@logwrap
def get_fqdn_by_hostname(hostname):
return (
hostname + DNS_SUFFIX if DNS_SUFFIX not in hostname else hostname
)
def get_nodegroup(self, cluster_id, name='default', group_id=None):
ngroups = self.client.get_nodegroups()
for group in ngroups:
if group['cluster_id'] == cluster_id and group['name'] == name:
if group_id and group['id'] != group_id:
continue
return group
return None
def update_nodegroups(self, cluster_id, node_groups):
for ngroup in node_groups:
if not self.get_nodegroup(cluster_id, name=ngroup):
self.client.create_nodegroup(cluster_id, ngroup)
# Assign nodes to nodegroup if nodes are specified
if len(node_groups[ngroup]) > 0:
ngroup_id = self.get_nodegroup(cluster_id, name=ngroup)['id']
self.client.assign_nodegroup(ngroup_id, node_groups[ngroup])
@logwrap
def get_nailgun_primary_node(self, slave, role='primary-controller'):
# returns controller or mongo that is primary in nailgun
with self.get_ssh_for_node(slave.name) as remote:
data = yaml.load(''.join(
remote.execute('cat /etc/astute.yaml')['stdout']))
nodes = data['network_metadata']['nodes']
node_name = [node['fqdn'] for node in nodes.values()
if role in node['node_roles']][0]
logger.debug("node name is {0}".format(node_name))
fqdn = self.get_fqdn_by_hostname(node_name)
devops_node = self.get_devops_node_by_nailgun_fqdn(fqdn)
return devops_node
@logwrap
def get_rabbit_master_node(self, node, fqdn_needed=False):
with self.get_ssh_for_node(node) as remote:
cmd = 'crm resource status master_p_rabbitmq-server'
output = ''.join(remote.execute(cmd)['stdout'])
master_node = re.search(
'resource master_p_rabbitmq-server is running on: (.*) Master',
output).group(1)
if fqdn_needed:
return master_node
else:
devops_node = self.find_devops_node_by_nailgun_fqdn(
master_node, self.environment.d_env.nodes().slaves)
return devops_node
def check_plugin_exists(self, cluster_id, plugin_name, section='editable'):
attr = self.client.get_cluster_attributes(cluster_id)[section]
return plugin_name in attr
def update_plugin_data(self, cluster_id, plugin_name, data):
attr = self.client.get_cluster_attributes(cluster_id)
# Do not re-upload anything, except selected plugin data
plugin_attributes = {
'editable': {plugin_name: attr['editable'][plugin_name]}}
for option, value in data.items():
plugin_data = plugin_attributes['editable'][plugin_name]
path = option.split("/")
"""Key 'metadata' can be in section
plugin_data['metadata']['versions']
For enable/disable plugin value must be set in
plugin_data['metadata']['enabled']
"""
if 'metadata' in path:
plugin_data['metadata'][path[-1]] = value
elif 'versions' in plugin_data['metadata']:
for version in plugin_data['metadata']['versions']:
for p in path[:-1]:
version = version[p]
version[path[-1]] = value
else:
for p in path[:-1]:
plugin_data = plugin_data[p]
plugin_data[path[-1]] = value
self.client.update_cluster_attributes(cluster_id, plugin_attributes)
def get_plugin_data(self, cluster_id, plugin_name, version):
"""Return data (settings) for specified version of plugin
:param cluster_id: int
:param plugin_name: string
:param version: string
:return: dict
"""
attr = self.client.get_cluster_attributes(cluster_id)
plugin_data = attr['editable'][plugin_name]
plugin_versions = plugin_data['metadata']['versions']
for p in plugin_versions:
if p['metadata']['plugin_version'] == version:
return p
raise AssertionError("Plugin {0} version {1} is not "
"found".format(plugin_name, version))
def update_plugin_settings(self, cluster_id, plugin_name, version, data,
enabled=True):
"""Update settings for specified version of plugin
:param plugin_name: string
:param version: string
:param data: dict - settings for the plugin
:return: None
"""
attr = self.client.get_cluster_attributes(cluster_id)
plugin_versions = attr['editable'][plugin_name]['metadata']['versions']
if enabled:
attr['editable'][plugin_name]['metadata']['enabled'] = True
plugin_data = None
for item in plugin_versions:
if item['metadata']['plugin_version'] == version:
plugin_data = item
break
assert_true(plugin_data is not None, "Plugin {0} version {1} is not "
"found".format(plugin_name, version))
for option, value in data.items():
path = option.split("/")
for p in path[:-1]:
plugin_settings = plugin_data[p]
plugin_settings[path[-1]] = value
self.client.update_cluster_attributes(cluster_id, attr)
@staticmethod
@logwrap
def prepare_ceph_to_delete(remote_ceph):
hostname = ''.join(remote_ceph.execute(
"hostname -s")['stdout']).strip()
osd_tree = ceph.get_osd_tree(remote_ceph)
logger.debug("osd tree is {0}".format(osd_tree))
ids = []
for osd in osd_tree['nodes']:
if hostname in osd['name']:
ids = osd['children']
logger.debug("ids are {}".format(ids))
assert_true(ids, "osd ids for {} weren't found".format(hostname))
for osd_id in ids:
remote_ceph.execute("ceph osd out {}".format(osd_id))
wait(lambda: ceph.is_health_ok(remote_ceph),
interval=30, timeout=10 * 60,
timeout_msg='ceph helth ok timeout')
for osd_id in ids:
if OPENSTACK_RELEASE_UBUNTU in OPENSTACK_RELEASE:
remote_ceph.execute("stop ceph-osd id={}".format(osd_id))
else:
remote_ceph.execute("service ceph stop osd.{}".format(osd_id))
remote_ceph.execute("ceph osd crush remove osd.{}".format(osd_id))
remote_ceph.execute("ceph auth del osd.{}".format(osd_id))
remote_ceph.execute("ceph osd rm osd.{}".format(osd_id))
# remove ceph node from crush map
remote_ceph.execute("ceph osd crush remove {}".format(hostname))
@logwrap
def get_rabbit_slaves_node(self, node, fqdn_needed=False):
with self.get_ssh_for_node(node) as remote:
cmd = 'crm resource status master_p_rabbitmq-server'
list_output = ''.join(remote.execute(cmd)['stdout']).split('\n')
filtered_list = [el for el in list_output
if el and not el.endswith('Master')]
slaves_nodes = []
for el in filtered_list:
slaves_nodes.append(
re.search('resource master_p_rabbitmq-server is running on:'
' (.*)', el).group(1).strip())
if fqdn_needed:
return slaves_nodes
else:
devops_nodes = [self.find_devops_node_by_nailgun_fqdn(
slave_node, self.environment.d_env.nodes().slaves)
for slave_node in slaves_nodes]
return devops_nodes
@logwrap
def run_deployment_tasks(self, cluster_id, nodes, tasks):
self.client.put_deployment_tasks_for_cluster(
cluster_id=cluster_id, data=tasks,
node_id=','.join(map(str, nodes)))
tasks = self.client.get_tasks()
deploy_tasks = [t for t in tasks if t['status']
in ('pending', 'running') and
t['name'] == 'deployment' and
t['cluster'] == cluster_id]
for task in deploy_tasks:
if min([t['progress'] for t in deploy_tasks]) == task['progress']:
return task
@logwrap
def wait_deployment_tasks(self, cluster_id, nodes, tasks, timeout=60 * 10):
task = self.run_deployment_tasks(cluster_id, nodes, tasks)
assert_is_not_none(task,
'Got empty result after running deployment tasks!')
self.assert_task_success(task, timeout)
@logwrap
def get_alive_proxy(self, cluster_id, port='8888'):
online_controllers = [node for node in
self.get_nailgun_cluster_nodes_by_roles(
cluster_id,
roles=['controller', ]) if node['online']]
with self.environment.d_env.get_admin_remote() as admin_remote:
check_proxy_cmd = ('[[ $(curl -s -w "%{{http_code}}" '
'{0} -o /dev/null) -eq 200 ]]')
for controller in online_controllers:
proxy_url = 'http://{0}:{1}/'.format(controller['ip'], port)
logger.debug('Trying to connect to {0} from master node...'
.format(proxy_url))
if admin_remote.execute(
check_proxy_cmd.format(proxy_url))['exit_code'] == 0:
return proxy_url
assert_true(len(online_controllers) > 0,
'There are no online controllers available '
'to provide HTTP proxy!')
assert_false(len(online_controllers) == 0,
'There are online controllers available ({0}), '
'but no HTTP proxy is accessible from master '
'node'.format(online_controllers))
@logwrap
def get_cluster_credentials(self, cluster_id):
attributes = self.client.get_cluster_attributes(cluster_id)
username = attributes['editable']['access']['user']['value']
password = attributes['editable']['access']['password']['value']
tenant = attributes['editable']['access']['tenant']['value']
return {'username': username,
'password': password,
'tenant': tenant}
@logwrap
def get_cluster_additional_components(self, cluster_id):
components = {}
attributes = self.client.get_cluster_attributes(cluster_id)
add_comps = attributes['editable']['additional_components'].items()
for comp, opts in add_comps:
# exclude metadata
if 'metadata' not in comp:
components[comp] = opts['value']
return components
@logwrap
def get_cluster_ibp_packages(self, cluster_id):
attributes = self.client.get_cluster_attributes(cluster_id)
pkgs = attributes['editable']['provision']['packages']['value']
return set(pkgs.splitlines())
@logwrap
def update_cluster_ibp_packages(self, cluster_id, pkgs):
attributes = self.client.get_cluster_attributes(cluster_id)
attributes['editable']['provision']['packages']['value'] = '\n'.join(
pkgs)
self.client.update_cluster_attributes(cluster_id, attributes)
return self.get_cluster_ibp_packages(cluster_id)
@logwrap
def spawn_vms_wait(self, cluster_id, timeout=60 * 60, interval=30):
logger.info('Spawn VMs of a cluster %s', cluster_id)
task = self.client.spawn_vms(cluster_id)
self.assert_task_success(task, timeout=timeout, interval=interval)
@logwrap
def get_all_ostf_set_names(self, cluster_id):
sets = self.client.get_ostf_test_sets(cluster_id)
return [s['id'] for s in sets]
@logwrap
def update_network_cidr(self, cluster_id, network_name):
"""Simple method for changing default network cidr
(just use its subnet with 2x smaller network mask)
:param cluster_id: int
:param network_name: str
:return: None
"""
networks = self.client.get_networks(cluster_id)['networks']
params = self.client.get_networks(cluster_id)['networking_parameters']
for network in networks:
if network['name'] != network_name:
continue
old_cidr = netaddr.IPNetwork(str(network['cidr']))
new_cidr = list(old_cidr.subnet(old_cidr.prefixlen + 1))[0]
assert_not_equal(old_cidr, new_cidr,
'Can\t create a subnet using default cidr {0} '
'for {1} network!'.format(old_cidr, network_name))
network['cidr'] = str(new_cidr)
logger.debug('CIDR for {0} network was changed from {1} to '
'{2}.'.format(network_name, old_cidr, new_cidr))
if network['meta']['notation'] != 'ip_ranges':
continue
if network['name'] == 'public':
network['ip_ranges'] = self.get_range(new_cidr, ip_range=-1)
params['floating_ranges'] = self.get_range(new_cidr,
ip_range=1)
else:
network['ip_ranges'] = self.get_range(new_cidr, ip_range=0)
self.client.update_network(cluster_id, params, networks)
@logwrap
def wait_task_success(self, task_name='', interval=30,
timeout=help_data.DEPLOYMENT_TIMEOUT):
"""Wait provided task to finish
:param task_name: str
:param interval: int
:param timeout: int
:return: None
"""
all_tasks = self.client.get_tasks()
tasks = [task for task in all_tasks if task['name'] == task_name]
latest_task = sorted(tasks, key=lambda k: k['id'])[-1]
self.assert_task_success(latest_task, interval=interval,
timeout=timeout)
def deploy_cluster_changes_wait(
self, cluster_id, data=None,
timeout=help_data.DEPLOYMENT_TIMEOUT,
interval=30):
"""Redeploy cluster to apply changes in its settings
:param cluster_id: int, env ID to apply changes for
:param data: dict, changed env settings
:param timeout: int, time (in seconds) to wait for deployment end
:param interval: int, time (in seconds) between deployment
status queries
:return:
"""
logger.info('Re-deploy cluster {} to apply the changed '
'settings'.format(cluster_id))
if data is None:
data = {}
task = self.client.redeploy_cluster_changes(cluster_id, data)
self.assert_task_success(task, interval=interval, timeout=timeout)
def execute_task_on_node(self, task_name, node_id,
cluster_id, force_exception=False,
force_execution=True):
"""Execute deployment task against the corresponding node
:param task_name: str, name of a task to execute
:param node_id: int, node ID to execute task on
:param cluster_id: int, cluster ID
:param force_exception: bool, indication whether exceptions on task
execution are ignored
:param force_execution: bool, run particular task on nodes
and do not care if there were changes or not
:return: None
"""
try:
logger.info("Trying to execute {!r} task on node {!r}"
.format(task_name, node_id))
task = self.client.put_deployment_tasks_for_cluster(
cluster_id=cluster_id,
data=[task_name],
node_id=node_id,
force=force_execution)
self.assert_task_success(task, timeout=30 * 60)
except (AssertionError, TimeoutError):
logger.exception("Failed to run task {!r}".format(task_name))
if force_exception:
raise
class FuelWebClient30(FuelWebClient29):
"""FuelWebClient that works with fuel-devops 3.0
"""
@logwrap
def get_default_node_group(self):
return self.environment.d_env.get_group(name='default')
@logwrap
def get_public_gw(self):
default_node_group = self.get_default_node_group()
pub_pool = default_node_group.get_network_pool(name='public')
return str(pub_pool.gateway)
@logwrap
def nodegroups_configure(self, cluster_id):
# Add node groups with networks
if len(self.environment.d_env.get_groups()) > 1:
ng = {rack.name: [] for rack in
self.environment.d_env.get_groups()}
ng_nets = []
for rack in self.environment.d_env.get_groups():
nets = {'name': rack.name}
nets['networks'] = {r.name: r.address_pool.name for
r in rack.get_network_pools(
name__in=['fuelweb_admin',
'public',
'management',
'storage',
'private'])}
ng_nets.append(nets)
self.update_nodegroups(cluster_id=cluster_id,
node_groups=ng)
self.update_nodegroups_network_configuration(cluster_id,
ng_nets)
def change_default_network_settings(self):
api_version = self.client.get_api_version()
if int(api_version["release"][0]) < 6:
return
def fetch_networks(networks):
"""Parse response from api/releases/1/networks and return dict with
networks' settings - need for avoiding hardcode"""
result = {}
for net in networks:
if (net['name'] == 'private' and
net.get('seg_type', '') == 'tun'):
result['private_tun'] = net
elif (net['name'] == 'private' and
net.get('seg_type', '') == 'gre'):
result['private_gre'] = net
elif net['name'] == 'public':
result['public'] = net
elif net['name'] == 'management':
result['management'] = net
elif net['name'] == 'storage':
result['storage'] = net
elif net['name'] == 'baremetal':
result['baremetal'] = net
return result
default_node_group = self.get_default_node_group()
logger.info("Default node group has {} name".format(
default_node_group.name))
logger.info("Applying default network settings")
for _release in self.client.get_releases():
logger.info(
'Applying changes for release: {}'.format(
_release['name']))
net_settings = \
self.client.get_release_default_net_settings(
_release['id'])
for net_provider in NETWORK_PROVIDERS:
if net_provider not in net_settings:
# TODO(ddmitriev): should show warning if NETWORK_PROVIDERS
# are not match providers in net_settings.
continue
networks = fetch_networks(
net_settings[net_provider]['networks'])
pub_pool = default_node_group.get_network_pool(
name='public')
networks['public']['cidr'] = str(pub_pool.net)
networks['public']['gateway'] = str(pub_pool.gateway)
networks['public']['notation'] = 'ip_ranges'
networks['public']['vlan_start'] = \
pub_pool.vlan_start if pub_pool.vlan_start else None
networks['public']['ip_range'] = list(
pub_pool.ip_range(relative_start=2, relative_end=-16))
net_settings[net_provider]['config']['floating_ranges'] = [
list(pub_pool.ip_range('floating',
relative_start=-15,
relative_end=-2))]
if 'baremetal' in networks and \
default_node_group.get_network_pools(name='ironic'):
ironic_net_pool = default_node_group.get_network_pool(
name='ironic')
networks['baremetal']['cidr'] = ironic_net_pool.net
net_settings[net_provider]['config'][
'baremetal_gateway'] = ironic_net_pool.gateway
networks['baremetal']['ip_range'] = \
list(ironic_net_pool.ip_range())
net_settings[net_provider]['config']['baremetal_range'] = \
list(ironic_net_pool.ip_range('baremetal'))
for pool in default_node_group.get_network_pools(
name__in=['storage', 'management']):
networks[pool.name]['cidr'] = str(pool.net)
networks[pool.name]['ip_range'] = self.get_range(
pool.net)[0]
networks[pool.name]['notation'] = 'ip_ranges'
networks[pool.name]['vlan_start'] = pool.vlan_start
if net_provider == 'neutron':
private_net_pool = default_node_group.get_network_pool(
name='private')
networks['private_tun']['cidr'] = str(private_net_pool.net)
networks['private_gre']['cidr'] = str(private_net_pool.net)
networks['private_tun']['vlan_start'] = \
private_net_pool.vlan_start or None
networks['private_gre']['vlan_start'] = \
private_net_pool.vlan_start or None
net_settings[net_provider]['config']['internal_cidr'] = \
'192.168.0.0/24'
net_settings[net_provider]['config']['internal_gateway'] =\
'192.168.0.1'
elif net_provider == 'nova_network':
private_net_pool = default_node_group.get_network_pool(
name='private')
net_settings[net_provider]['config'][
'fixed_networks_cidr'] = \
str(private_net_pool.net) or None
net_settings[net_provider]['config'][
'fixed_networks_vlan_start'] = \
private_net_pool.vlan_start or None
self.client.put_release_default_net_settings(
_release['id'], net_settings)
def get_node_group_and_role(self, node_name, nodes_dict):
devops_node = self.environment.d_env.get_node(name=node_name)
node_group = devops_node.group.name
if isinstance(nodes_dict[node_name][0], list):
# Backwards compatibility
node_roles = nodes_dict[node_name][0]
else:
node_roles = nodes_dict[node_name]
return node_group, node_roles
@logwrap
def update_nodes_interfaces(self, cluster_id, nailgun_nodes=None):
nailgun_nodes = nailgun_nodes or []
if not nailgun_nodes:
nailgun_nodes = self.client.list_cluster_nodes(cluster_id)
for node in nailgun_nodes:
assigned_networks = {}
interfaces = self.client.get_node_interfaces(node['id'])
interfaces = {iface['mac']: iface for iface in interfaces}
d_node = self.get_devops_node_by_nailgun_node(node)
for net in d_node.network_configs:
if net.aggregation is None: # Have some ifaces aggregation?
node_iface = d_node.interface_set.get(label=net.label)
assigned_networks[interfaces[
node_iface.mac_address]['name']] = net.networks
else:
assigned_networks[net.label] = net.networks
logger.info(
'Assigned networks for node {node} are: {networks}'
''.format(
node=node['name'],
networks=str(assigned_networks)
)
)
self.update_node_networks(node['id'], assigned_networks)
@logwrap
def update_node_networks(self, node_id, interfaces_dict,
raw_data=None,
override_ifaces_params=None):
interfaces = self.client.get_node_interfaces(node_id)
node = [n for n in self.client.list_nodes() if n['id'] == node_id][0]
d_node = self.get_devops_node_by_nailgun_node(node)
bonds = [n for n in d_node.network_configs
if n.aggregation is not None]
for bond in bonds:
macs = [i.mac_address.lower() for i in
d_node.interface_set.filter(label__in=bond.parents)]
parents = [{'name': iface['name']} for iface in interfaces
if iface['mac'].lower() in macs]
bond_config = {
'mac': None,
'mode': bond.aggregation,
'name': bond.label,
'slaves': parents,
'state': None,
'type': 'bond',
'assigned_networks': []
}
interfaces.append(bond_config)
if raw_data is not None:
interfaces.extend(raw_data)
def get_iface_by_name(ifaces, name):
iface = [i for i in ifaces if i['name'] == name]
assert_true(len(iface) > 0,
"Interface with name {} is not present on "
"node. Please check override params.".format(name))
return iface[0]
if override_ifaces_params is not None:
for interface in override_ifaces_params:
get_iface_by_name(interfaces, interface['name']).\
update(interface)
all_networks = dict()
for interface in interfaces:
all_networks.update(
{net['name']: net for net in interface['assigned_networks']})
for interface in interfaces:
name = interface["name"]
interface['assigned_networks'] = \
[all_networks[i] for i in interfaces_dict.get(name, []) if
i in all_networks.keys()]
self.client.put_node_interfaces(
[{'id': node_id, 'interfaces': interfaces}])
def update_nodegroup_net_settings(self, network_configuration, nodegroup,
cluster_id=None):
# seg_type = network_configuration.get('networking_parameters', {}) \
# .get('segmentation_type')
nodegroup_id = self.get_nodegroup(cluster_id, nodegroup['name'])['id']
for net in network_configuration.get('networks'):
if nodegroup['name'] == 'default' and \
net['name'] == 'fuelweb_admin':
continue
if net['group_id'] == nodegroup_id:
group = self.environment.d_env.get_group(
name=nodegroup['name'])
net_pool = group.networkpool_set.get(name=net['name'])
net['cidr'] = net_pool.net
# if net['meta']['use_gateway']:
# net['gateway'] = net_pool.gateway
# else:
# net['gateway'] = None
net['gateway'] = net_pool.gateway
if net['gateway']:
net['meta']['use_gateway'] = True
net['meta']['gateway'] = net['gateway']
else:
net['meta']['use_gateway'] = False
net['vlan_start'] = net_pool.vlan_start
net['meta']['notation'] = 'ip_ranges'
net['ip_ranges'] = [list(net_pool.ip_range())]
return network_configuration
# TODO(ddmitriev): this code will be removed after moving to fuel-devops3.0
# pylint: disable=no-member
# noinspection PyUnresolvedReferences
if (distutils.version.LooseVersion(devops.__version__) <
distutils.version.LooseVersion('3')):
logger.info("Use FuelWebClient compatible to fuel-devops 2.9")
FuelWebClient = FuelWebClient29
else:
logger.info("Use FuelWebClient compatible to fuel-devops 3.0")
FuelWebClient = FuelWebClient30
# pylint: enable=no-member
View Git Blame Copy Permalink