openstack/charm-cinder
Code Issues Proposed changes
9a6ab28af2
charm-cinder/hooks/cinder_utils.py
Nobuto Murata 9a6ab28af2 Install lsscsi by default for zed or older 
lsscsi is a dependency of os-brick for volume related operations
including creating a boot volume with iSCSI/FC backends. However, it
wasn't installed by default until jammy-antelope (lunar) due to a
packaging bug. It hasn't been an obvious issue especially when the
cinder-volume unit and the nova-compute unit are co-located since
nova-compute charm installs the package. Instead of applying a
workaround to all backend charms, let's apply the workaround in the
cinder charm since lsscsi is installed by default for later OpenStack
releases anyway and it's a small package.

Closes-Bug: #1939390
Change-Id: I131e12c53c78957873b6a75a3b1b67b1725b43ed
2024-08-20 12:22:14 +09:00

1082 lines
37 KiB
Python
Raw Blame History

# Copyright 2016 Canonical Ltd
#
# 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 print_function
import os
import re
import subprocess
import uuid
from copy import deepcopy
from collections import OrderedDict
from copy import copy
from tempfile import NamedTemporaryFile
from charmhelpers.core.strutils import (
bytes_from_string
)
from charmhelpers.core.hookenv import (
config,
local_unit,
relation_get,
relation_set,
relation_ids,
related_units,
log,
DEBUG,
INFO,
ERROR,
hook_name,
)
from charmhelpers.fetch import (
apt_upgrade,
apt_update,
apt_install,
add_source,
apt_purge,
apt_autoremove,
filter_missing_packages,
)
from charmhelpers.core.host import (
CompareHostReleases,
lsb_release,
mkdir,
mounts,
umount,
service_restart,
service_stop,
service_start,
)
from charmhelpers.contrib.openstack.alternatives import install_alternative
from charmhelpers.contrib.hahelpers.cluster import (
is_elected_leader,
get_hacluster_config,
get_managed_services_and_ports,
)
from charmhelpers.contrib.storage.linux.utils import (
is_block_device,
zap_disk,
is_device_mounted,
)
from charmhelpers.contrib.storage.linux.lvm import (
create_lvm_physical_volume,
create_lvm_volume_group,
deactivate_lvm_volume_group,
extend_logical_volume_by_device,
is_lvm_physical_volume,
list_lvm_volume_group,
list_thin_logical_volume_pools,
remove_lvm_physical_volume,
)
from charmhelpers.contrib.storage.linux.loopback import (
ensure_loopback_device,
)
from charmhelpers.contrib.openstack import (
templating,
context,
)
from charmhelpers.contrib.openstack.utils import (
configure_installation_source,
get_os_codename_install_source,
get_subordinate_release_packages,
os_release,
reset_os_release,
make_assess_status_func,
pause_unit,
resume_unit,
is_unit_paused_set,
os_application_version_set,
token_cache_pkgs,
enable_memcache,
CompareOpenStackReleases,
)
from charmhelpers.core.decorators import (
retry_on_exception,
)
import cinder_contexts
from cinder_contexts import ceph_config_file
COMMON_PACKAGES = [
'cinder-common',
'gdisk',
'librbd1', # bug 1440948 vol-from-img
'python-jinja2',
'python-keystoneclient',
'python-mysqldb',
'python-psycopg2',
'qemu-utils',
'thin-provisioning-tools',
'nfs-common',
]
PY3_PACKAGES = [
'python3-cinder',
'python3-memcache',
'python3-rados',
'python3-rbd',
]
PY3_API_PACKAGES = ['libapache2-mod-wsgi-py3']
API_PACKAGES = [
'apache2',
'cinder-api',
'haproxy',
]
# lsscsi is installed as a dependency of os-brick for jammy-antelope
# (lunar) or later. Explicit installation by the charm is only necessary
# for zed or older releases. LP: #1939390
VOLUME_PACKAGES = ['cinder-volume', 'lsscsi']
SCHEDULER_PACKAGES = ['cinder-scheduler']
DEFAULT_LOOPBACK_SIZE = '5G'
# Cluster resource used to determine leadership when hacluster'd
CLUSTER_RES = 'grp_cinder_vips'
CINDER_DB_INIT_RKEY = 'cinder-db-initialised'
CINDER_DB_INIT_ECHO_RKEY = 'cinder-db-initialised-echo'
class CinderCharmError(Exception):
pass
CINDER_CONF_DIR = "/etc/cinder"
CINDER_CONF = '%s/cinder.conf' % CINDER_CONF_DIR
CINDER_API_CONF = '%s/api-paste.ini' % CINDER_CONF_DIR
CINDER_POLICY_JSON = '%s/policy.json' % CINDER_CONF_DIR
CINDER_AUDIT_MAP = '%s/api_audit_map.conf' % CINDER_CONF_DIR
CEPH_CONF = '/etc/ceph/ceph.conf'
HAPROXY_CONF = '/etc/haproxy/haproxy.cfg'
APACHE_PORTS_CONF = '/etc/apache2/ports.conf'
APACHE_SITE_CONF = '/etc/apache2/sites-available/openstack_https_frontend'
APACHE_SITE_24_CONF = '/etc/apache2/sites-available/' \
'openstack_https_frontend.conf'
MEMCACHED_CONF = '/etc/memcached.conf'
WSGI_CINDER_API_CONF = '/etc/apache2/sites-enabled/wsgi-openstack-api.conf'
PACKAGE_CINDER_API_CONF = '/etc/apache2/conf-enabled/cinder-wsgi.conf'
# max number of objects to include in a batch/query when running
# online_data_migrations.
MAX_OBJECTS_COUNT = '5000'
VERSION_PACKAGE = 'cinder-common'
TEMPLATES = 'templates/'
# The interface is said to be satisfied if anyone of the interfaces in
# the
# list has a complete context.
REQUIRED_INTERFACES = {
'database': ('shared-db',),
'messaging': ('amqp',),
'identity': ('identity-service',),
}
def required_interfaces():
"""Provide the required charm interfaces based on configured roles."""
_interfaces = copy(REQUIRED_INTERFACES)
if not service_enabled('api'):
# drop requirement for identity interface
_interfaces.pop('identity')
return _interfaces
# Map config files to hook contexts and services that will be associated
# with file in restart_on_changes()'s service map.
BASE_RESOURCE_MAP = OrderedDict([
(CINDER_CONF, {
'contexts': [context.SharedDBContext(ssl_dir=CINDER_CONF_DIR),
context.AMQPContext(ssl_dir=CINDER_CONF_DIR),
context.ImageServiceContext(),
context.OSConfigFlagContext(),
context.SyslogContext(),
cinder_contexts.CephContext(),
cinder_contexts.HAProxyContext(),
cinder_contexts.ImageServiceContext(),
cinder_contexts.CinderSubordinateConfigContext(
interface=['storage-backend', 'backup-backend'],
service='cinder',
config_file=CINDER_CONF),
cinder_contexts.StorageBackendContext(),
cinder_contexts.LoggingConfigContext(),
context.IdentityCredentialsContext(
service='cinder',
service_user='cinder'),
context.IdentityServiceContext(
service='cinder',
service_user='cinder'),
context.BindHostContext(),
context.WorkerConfigContext(),
cinder_contexts.RegionContext(),
context.InternalEndpointContext(),
cinder_contexts.VolumeUsageAuditContext(),
context.MemcacheContext(),
cinder_contexts.SectionalConfigContext(),
cinder_contexts.LVMContext(),
context.KeystoneAuditMiddleware(service='cinder')],
'services': ['cinder-api', 'cinder-volume', 'cinder-scheduler',
'haproxy']
}),
(CINDER_API_CONF, {
'contexts': [context.IdentityServiceContext(),
context.KeystoneAuditMiddleware(service='cinder')],
'services': ['cinder-api'],
}),
(CINDER_POLICY_JSON, {
'contexts': [],
'services': ['cinder-api']
}),
(CINDER_AUDIT_MAP, {
'contexts': [context.KeystoneAuditMiddleware(service='cinder')],
'services': ['cinder-api']
}),
(ceph_config_file(), {
'contexts': [context.CephContext()],
'services': ['cinder-volume']
}),
(HAPROXY_CONF, {
'contexts': [context.HAProxyContext(singlenode_mode=True),
cinder_contexts.HAProxyContext()],
'services': ['haproxy'],
}),
(APACHE_SITE_CONF, {
'contexts': [cinder_contexts.ApacheSSLContext()],
'services': ['apache2'],
}),
(APACHE_SITE_24_CONF, {
'contexts': [cinder_contexts.ApacheSSLContext()],
'services': ['apache2'],
}),
(APACHE_PORTS_CONF, {
'contexts': [],
'services': ['apache2'],
}),
])
def register_configs(release=None):
"""Register config files with their respective contexts.
Regstration of some configs may not be required depending on
existing of certain relations.
"""
# if called without anything installed (eg during install hook)
# just default to earliest supported release. configs dont get touched
# till post-install, anyway.
release = release or os_release('cinder-common', base='icehouse')
configs = templating.OSConfigRenderer(templates_dir=TEMPLATES,
openstack_release=release)
for cfg, rscs in resource_map().items():
configs.register(cfg, rscs['contexts'])
return configs
def resource_map(release=None):
"""
Dynamically generate a map of resources that will be managed for a single
hook execution.
"""
resource_map = deepcopy(BASE_RESOURCE_MAP)
release = release or os_release('cinder-common', base='icehouse')
if relation_ids('backup-backend'):
resource_map[CINDER_CONF]['services'].append('cinder-backup')
resource_map[ceph_config_file()]['services'].append('cinder-backup')
if relation_ids('ceph') and hook_name() != 'ceph-relation-broken':
# need to create this early, new peers will have a relation during
# registration # before they've run the ceph hooks to create the
# directory.
# !!! FIX: These side effects seem inappropriate for this method
mkdir(os.path.dirname(CEPH_CONF))
mkdir(os.path.dirname(ceph_config_file()))
# Install ceph config as an alternative for co-location with
# ceph and ceph-osd charm - cinder ceph.conf will be
# lower priority than both of these but thats OK
if not os.path.exists(ceph_config_file()):
# touch file for pre-templated generation
open(ceph_config_file(), 'w').close()
install_alternative(os.path.basename(CEPH_CONF),
CEPH_CONF, ceph_config_file())
else:
resource_map.pop(ceph_config_file())
if os.path.exists('/etc/apache2/conf-available'):
resource_map.pop(APACHE_SITE_CONF)
else:
resource_map.pop(APACHE_SITE_24_CONF)
# Remove services from map which are not enabled by user config
for cfg in resource_map.keys():
resource_map[cfg]['services'] = \
filter_services(resource_map[cfg]['services'])
if enable_memcache(source=config()['openstack-origin']):
resource_map[MEMCACHED_CONF] = {
'contexts': [context.MemcacheContext()],
'services': ['memcached']}
if not service_enabled('api'):
# haproxy and apache2 are related to cinder-api
cfg_files = {
CINDER_API_CONF,
HAPROXY_CONF,
APACHE_SITE_CONF,
APACHE_SITE_24_CONF,
APACHE_PORTS_CONF,
}
for cfg in cfg_files.intersection(resource_map.keys()):
resource_map.pop(cfg)
elif run_in_apache():
for cfile in resource_map:
svcs = resource_map[cfile]['services']
if 'cinder-api' in svcs:
svcs.remove('cinder-api')
if 'apache2' not in svcs:
svcs.append('apache2')
wsgi_script = "/usr/bin/cinder-wsgi"
resource_map[WSGI_CINDER_API_CONF] = {
'contexts': [context.WSGIWorkerConfigContext(name="cinder",
script=wsgi_script),
cinder_contexts.HAProxyContext()],
'services': ['apache2']
}
if release and CompareOpenStackReleases(release) < 'queens':
resource_map.pop(CINDER_POLICY_JSON)
return resource_map
def filter_services(svcs):
'''Remove services not enabled by user config from a list of services
@param svcs: List of services
@returns : List of enabled services
'''
deps = {
'haproxy': 'api',
'apache2': 'api',
}
return [s for s in svcs
if service_enabled(s.lstrip('cinder-')) or
(deps.get(s) and service_enabled(deps[s]))]
def juju_log(msg):
log('[cinder] %s' % msg)
def determine_packages():
'''Determine list of packages required for the currently enabled services.
:returns: list of package names
'''
pkgs = copy(COMMON_PACKAGES)
for s, p in [('api', API_PACKAGES),
('volume', VOLUME_PACKAGES),
('scheduler', SCHEDULER_PACKAGES)]:
if service_enabled(s):
pkgs += p
pkgs.extend(token_cache_pkgs(source=config()['openstack-origin']))
release = os_release('cinder')
if CompareOpenStackReleases(release) >= 'rocky':
pkgs = [p for p in pkgs if not p.startswith('python-')]
pkgs.extend(PY3_PACKAGES)
if service_enabled('api'):
pkgs.extend(PY3_API_PACKAGES)
pkgs = set(pkgs).union(get_subordinate_release_packages(
release).install)
return sorted(pkgs)
def determine_purge_packages():
'''
Determine list of packages that where previously installed which are no
longer needed.
:returns: list of package names
'''
pkgs = []
release = os_release('cinder')
if CompareOpenStackReleases(release) >= 'rocky':
pkgs = [p for p in COMMON_PACKAGES if p.startswith('python-')]
pkgs.append('python-cinder')
pkgs.append('python-memcache')
pkgs = set(pkgs).union(
get_subordinate_release_packages(release).purge)
return sorted(pkgs)
def remove_old_packages():
'''Purge any packages that need ot be removed.
:returns: bool Whether packages were removed.
'''
installed_packages = filter_missing_packages(determine_purge_packages())
if installed_packages:
apt_purge(installed_packages, fatal=True)
apt_autoremove(purge=True, fatal=True)
return bool(installed_packages)
def service_enabled(service):
'''Determine if a specific cinder service is enabled in
charm configuration.
:param service: str: cinder service name to query (volume, scheduler, api,
all)
:returns: boolean: True if service is enabled in config, False if not.
'''
enabled = config()['enabled-services']
if enabled == 'all':
return True
return service in enabled
def restart_map():
'''Determine the correct resource map to be passed to
charmhelpers.core.restart_on_change() based on the services configured.
:returns: dict: A dictionary mapping config file to lists of services
that should be restarted when file changes.
'''
return OrderedDict([(cfg, v['services'])
for cfg, v in resource_map().items()
if v['services']])
def enabled_services():
m = restart_map()
svcs = set()
for t in m.items():
svcs.update(t[1])
return list(svcs)
def services():
''' Returns a list of services associate with this charm '''
_services = set()
for v in restart_map().values():
_services.update(v)
return list(sorted(_services))
def reduce_lvm_volume_group_missing(volume_group, extra_args=None):
'''
Remove all missing physical volumes from the volume group, if there
are no logical volumes allocated on them.
:param volume_group: str: Name of volume group to reduce.
:param extra_args: list: List of extra args to pass to vgreduce
'''
if extra_args is None:
extra_args = []
command = ['vgreduce', '--removemissing'] + extra_args + [volume_group]
subprocess.check_call(command)
def extend_lvm_volume_group(volume_group, block_device):
'''
Extend an LVM volume group onto a given block device.
Assumes block device has already been initialized as an LVM PV.
:param volume_group: str: Name of volume group to create.
:block_device: str: Full path of PV-initialized block device.
'''
subprocess.check_call(['vgextend', volume_group, block_device])
def lvm_volume_group_exists(volume_group):
"""Check for the existence of a volume group.
:param volume_group: str: Name of volume group.
"""
try:
subprocess.check_call(['vgdisplay', volume_group])
except subprocess.CalledProcessError:
return False
else:
return True
def remove_lvm_volume_group(volume_group):
"""Remove a volume group.
:param volume_group: str: Name of volume group to remove.
"""
subprocess.check_call(['vgremove', '--force', volume_group])
def ensure_lvm_volume_group_non_existent(volume_group):
"""Remove volume_group if it exists.
:param volume_group: str: Name of volume group.
"""
if not lvm_volume_group_exists(volume_group):
return
remove_lvm_volume_group(volume_group)
def log_lvm_info():
"""Log some useful information about how LVM is setup."""
try:
pvscan_output = subprocess.check_output(['pvscan']).decode('UTF-8')
juju_log('pvscan: {}'.format(pvscan_output))
except subprocess.CalledProcessError:
juju_log('pvscan did not complete successfully; may not be setup yet')
def configure_lvm_storage(block_devices, volume_group, overwrite=False,
remove_missing=False, remove_missing_force=False):
''' Configure LVM storage on the list of block devices provided
:param block_devices: list: List of whitelisted block devices to detect
and use if found
:param overwrite: bool: Scrub any existing block data if block device is
not already in-use
:param remove_missing: bool: Remove missing physical volumes from volume
group if logical volume not allocated on them
:param remove_missing_force: bool: Remove missing physical volumes from
volume group even if logical volumes are allocated
on them. Overrides 'remove_missing' if set.
'''
log_lvm_info()
devices = []
for block_device in block_devices:
(block_device, size) = _parse_block_device(block_device)
if not is_device_mounted(block_device):
if size == 0 and is_block_device(block_device):
devices.append(block_device)
elif size > 0:
devices.append(ensure_loopback_device(block_device, str(size)))
# NOTE(jamespage)
# might need todo an initial one-time scrub on install if need be
vg_found = False
new_devices = []
for device in devices:
if not is_lvm_physical_volume(device):
# Unused device
if overwrite is True or not has_partition_table(device):
prepare_volume(device)
new_devices.append(device)
elif (is_lvm_physical_volume(device) and
list_lvm_volume_group(device) != volume_group):
# Existing LVM but not part of required VG or new device
if overwrite is True:
prepare_volume(device)
new_devices.append(device)
elif (is_lvm_physical_volume(device) and
list_lvm_volume_group(device) == volume_group):
# Mark vg as found
vg_found = True
log_lvm_info()
if vg_found is False and len(new_devices) > 0:
if overwrite:
ensure_lvm_volume_group_non_existent(volume_group)
# Create new volume group from first device
create_lvm_volume_group(volume_group, new_devices[0])
new_devices.remove(new_devices[0])
# Remove missing physical volumes from volume group
try:
if remove_missing_force:
reduce_lvm_volume_group_missing(volume_group,
extra_args=['--force'])
elif remove_missing:
reduce_lvm_volume_group_missing(volume_group)
except subprocess.CalledProcessError as e:
juju_log("reduce_lvm_volume_group_missing() didn't complete."
" LVM may not be fully configured yet. Error was: '{}'."
.format(str(e)))
if len(new_devices) > 0:
# Extend the volume group as required
for new_device in new_devices:
extend_lvm_volume_group(volume_group, new_device)
thin_pools = list_thin_logical_volume_pools(path_mode=True)
if len(thin_pools) == 0:
juju_log("No thin pools found")
elif len(thin_pools) == 1:
juju_log("Thin pool {} found, extending with {}".format(
thin_pools[0],
new_device))
extend_logical_volume_by_device(thin_pools[0], new_device)
else:
juju_log("Multiple thin pools ({}) found, "
"skipping auto extending with {}".format(
','.join(thin_pools),
new_device))
log_lvm_info()
def prepare_volume(device):
juju_log("prepare_volume: {}".format(device))
clean_storage(device)
create_lvm_physical_volume(device)
juju_log("prepared volume: {}".format(device))
def has_partition_table(block_device):
out = subprocess.check_output(['fdisk', '-l', block_device],
stderr=subprocess.STDOUT).decode('UTF-8')
return "doesn't contain a valid partition" not in out
def clean_storage(block_device):
'''Ensures a block device is clean. That is:
- unmounted
- any lvm volume groups are deactivated
- any lvm physical device signatures removed
- partition table wiped
:param block_device: str: Full path to block device to clean.
'''
for mp, d in mounts():
if d == block_device:
juju_log('clean_storage(): Found %s mounted @ %s, unmounting.' %
(d, mp))
umount(mp, persist=True)
if is_lvm_physical_volume(block_device):
deactivate_lvm_volume_group(block_device)
remove_lvm_physical_volume(block_device)
zap_disk(block_device)
def _parse_block_device(block_device):
''' Parse a block device string and return either the full path
to the block device, or the path to a loopback device and its size
:param: block_device: str: Block device as provided in configuration
:returns: (str, int): Full path to block device and 0 OR
Full path to loopback device and required size
'''
_none = ['None', 'none', None]
if block_device in _none:
return (None, 0)
if block_device.startswith('/dev/'):
return (block_device, 0)
elif block_device.startswith('/'):
_bd = block_device.split('|')
if len(_bd) == 2:
bdev, size = _bd
else:
bdev = block_device
size = DEFAULT_LOOPBACK_SIZE
return (bdev, bytes_from_string(str(size)))
else:
return ('/dev/{}'.format(block_device), 0)
def is_db_initialised(cluster_rid=None):
"""
Check whether a db intialisation has been performed by any peer unit.
We base our decision on whether we or any of our peers has previously
sent or echoed an initialisation notification.
@param cluster_rid: current relation id. If none provided, all cluster
relation ids will be checked.
@return: True if there has been a db initialisation otherwise False.
"""
if cluster_rid:
rids = [cluster_rid]
else:
rids = relation_ids('cluster')
for c_rid in rids:
units = related_units(relid=c_rid) + [local_unit()]
for unit in units:
settings = relation_get(unit=unit, rid=c_rid) or {}
for key in [CINDER_DB_INIT_RKEY, CINDER_DB_INIT_ECHO_RKEY]:
if settings.get(key):
return True
return False
def is_new_dbinit_notification(init_id, echoed_init_id):
"""Returns True if we have a received a new db initialisation notification
from a peer unit and we have not previously echoed it to indicate that we
have already performed the necessary actions as result.
@param init_db: received initialisation notification.
@param echoed_init_db: value currently set for the echo key.
@return: True if new notification and False if not.
"""
return (init_id and (local_unit() not in init_id) and
(echoed_init_id != init_id))
def check_local_db_actions_complete():
"""Check if we have received db init'd notification and restart services
if we have not already.
NOTE: this must only be called from peer relation context.
"""
if not is_db_initialised():
return
settings = relation_get() or {}
if settings:
init_id = settings.get(CINDER_DB_INIT_RKEY)
echoed_init_id = relation_get(unit=local_unit(),
attribute=CINDER_DB_INIT_ECHO_RKEY)
# If we have received an init notification from a peer unit
# (assumed to be the leader) then restart cinder-* and echo the
# notification and don't restart again unless we receive a new
# (different) notification.
if is_new_dbinit_notification(init_id, echoed_init_id):
if not is_unit_paused_set():
log("Restarting cinder services following db "
"initialisation", level=DEBUG)
for svc in enabled_services():
service_restart(svc)
# Echo notification
data = {CINDER_DB_INIT_ECHO_RKEY: init_id}
# BUG: #1928383 - clear CINDER_DB_INIT_RKEY if not the leader
if not is_elected_leader(CLUSTER_RES):
data[CINDER_DB_INIT_RKEY] = None
relation_set(**data)
# NOTE(jamespage): Retry deals with sync issues during one-shot HA deploys.
# mysql might be restarting or suchlike.
@retry_on_exception(5, base_delay=3, exc_type=subprocess.CalledProcessError)
def migrate_database(upgrade=False):
"""Initialise cinder database if not already done so.
Runs cinder-manage to initialize a new database or migrate existing and
restarts services to ensure that the changes are picked up. The first
(leader) unit to perform this action should have broadcast this information
to its peers so first we check whether this has already occurred.
"""
if not upgrade and is_db_initialised():
log("Database is already initialised.", level=DEBUG)
return
cmd = ['cinder-manage', 'db', 'sync']
try:
subprocess.check_output(cmd, universal_newlines=True)
except subprocess.CalledProcessError as e:
# LP #1908037 - Due to a bug in cinder, the sync can fail if the
# volume types have not been migrated to __DEFAULT__ volume type.
# Check the output message from the logs to see if online data
# migrations are necessary and run that first.
if not upgrade:
raise
need_online_migration_msg = (
"Please run `cinder-manage db online_data_migrations`")
if need_online_migration_msg not in e.output:
raise
# Re-run online_data_migrations until the exit code is 0 or 2.
# See https://docs.openstack.org/cinder/latest/admin/upgrades.html
num_runs = 1
cmd_online_data_migrations = ['cinder-manage', 'db',
'online_data_migrations',
'--max_count', MAX_OBJECTS_COUNT]
exit_code = 1
while exit_code == 1:
log("Running online_data_migrations (%d iteration(s))" % num_runs,
level=INFO)
result = subprocess.run(cmd_online_data_migrations, check=False)
exit_code = result.returncode
# log as ERROR when the exit code 2 which is a fatal failure.
level = ERROR if exit_code == 2 else DEBUG
log("%s: exit code %s" % (" ".join(cmd_online_data_migrations),
result.returncode),
level=level)
log("stdout: %s" % result.stdout, level=level)
log("stderr: %s" % result.stderr, level=level)
if exit_code == 2:
# If no further migrations are possible, the exit status will
# be 2 if some migrations are still generating errors, which
# requires intervention to resolve.
raise subprocess.CalledProcessError(result.returncode,
result.args,
result.stdout,
result.stderr)
num_runs += 1
log("Re-running database migration", level=INFO)
subprocess.check_call(cmd)
# Notify peers so that services get restarted
log("Notifying peer(s) that db is initialised and restarting services",
level=DEBUG)
for r_id in relation_ids('cluster'):
if not is_unit_paused_set():
for svc in enabled_services():
service_restart(svc)
id = "%s-%s" % (local_unit(), uuid.uuid4())
relation_set(relation_id=r_id, **{CINDER_DB_INIT_RKEY: id})
def do_openstack_upgrade(configs=None):
"""Perform an uprade of cinder. Takes care of upgrading
packages, rewriting configs + database migration and
potentially any other post-upgrade actions.
:param configs: The charms main OSConfigRenderer object.
"""
new_src = config('openstack-origin')
new_os_rel = get_os_codename_install_source(new_src)
juju_log('Performing OpenStack upgrade to %s.' % (new_os_rel))
configure_installation_source(new_src)
dpkg_opts = [
'--option', 'Dpkg::Options::=--force-confnew',
'--option', 'Dpkg::Options::=--force-confdef',
]
apt_update()
apt_upgrade(options=dpkg_opts, fatal=True, dist=True)
reset_os_release()
apt_install(determine_packages(), fatal=True)
remove_old_packages()
# NOTE(hopem): must do this after packages have been upgraded so that
# we ensure that correct configs are selected for the target release.
# See LP 1726527.
configs = register_configs()
# set CONFIGS to load templates from new release and regenerate config
configs.set_release(openstack_release=new_os_rel)
configs.write_all()
if service_enabled('api') and run_in_apache():
disable_package_apache_site()
# Stop/start services and migrate DB if leader
[service_stop(s) for s in services()]
if is_elected_leader(CLUSTER_RES):
migrate_database(upgrade=True)
if not is_unit_paused_set():
[service_start(s) for s in services()]
def setup_ipv6():
ubuntu_rel = lsb_release()['DISTRIB_CODENAME'].lower()
if CompareHostReleases(ubuntu_rel) < "trusty":
raise Exception("IPv6 is not supported in the charms for Ubuntu "
"versions less than Trusty 14.04")
# Need haproxy >= 1.5.3 for ipv6 so for Trusty if we are <= Kilo we need to
# use trusty-backports otherwise we can use the UCA.
if (ubuntu_rel == 'trusty' and
CompareOpenStackReleases(os_release('cinder')) < 'liberty'):
add_source('deb http://archive.ubuntu.com/ubuntu trusty-backports '
'main')
apt_update()
apt_install('haproxy/trusty-backports', fatal=True)
def filesystem_mounted(fs):
return subprocess.call(['grep', '-wqs', fs, '/proc/mounts']) == 0
def get_optional_interfaces():
"""Return the optional interfaces that should be checked if the relavent
relations have appeared.
:returns: {general_interface: [specific_int1, specific_int2, ...], ...}
"""
optional_interfaces = {}
if relation_ids('ha'):
optional_interfaces['ha'] = ['cluster']
if relation_ids('storage-backend') or relation_ids('ceph'):
optional_interfaces['storage-backend'] = ['storage-backend', 'ceph']
if relation_ids('image-service'):
optional_interfaces['image'] = ['image-service']
if service_enabled('volume') and not service_enabled('api'):
optional_interfaces['identity-credentials'] = ['identity-credentials']
return optional_interfaces
def check_optional_relations(configs):
"""Check that if we have a relation_id for high availability that we can
get the hacluster config. If we can't then we are blocked. This function
is called from assess_status/set_os_workload_status as the charm_func and
needs to return either "unknown", "" if there is no problem or the status,
message if there is a problem.
:param configs: an OSConfigRender() instance.
:return 2-tuple: (string, string) = (status, message)
"""
if relation_ids('ha'):
try:
get_hacluster_config()
except Exception:
return ('blocked',
'hacluster missing configuration: '
'vip, vip_iface, vip_cidr')
# return 'unknown' as the lowest priority to not clobber an existing
# status.
return 'unknown', ''
def assess_status(configs):
"""Assess status of current unit
Decides what the state of the unit should be based on the current
configuration.
SIDE EFFECT: calls set_os_workload_status(...) which sets the workload
status of the unit.
Also calls status_set(...) directly if paused state isn't complete.
@param configs: a templating.OSConfigRenderer() object
@returns None - this function is executed for its side-effect
"""
assess_status_func(configs)()
os_application_version_set(VERSION_PACKAGE)
def assess_status_func(configs):
"""Helper function to create the function that will assess_status() for
the unit.
Uses charmhelpers.contrib.openstack.utils.make_assess_status_func() to
create the appropriate status function and then returns it.
Used directly by assess_status() and also for pausing and resuming
the unit.
NOTE: REQUIRED_INTERFACES is augmented with the optional interfaces
depending on the current config before being passed to the
make_assess_status_func() function.
NOTE(ajkavanagh) ports are not checked due to race hazards with services
that don't behave sychronously w.r.t their service scripts. e.g.
apache2.
@param configs: a templating.OSConfigRenderer() object
@return f() -> None : a function that assesses the unit's workload status
"""
interfaces = required_interfaces()
interfaces.update(get_optional_interfaces())
_services, _ = get_managed_services_and_ports(services(), [])
return make_assess_status_func(
configs, interfaces,
charm_func=check_optional_relations,
services=_services, ports=None)
def pause_unit_helper(configs):
"""Helper function to pause a unit, and then call assess_status(...) in
effect, so that the status is correctly updated.
Uses charmhelpers.contrib.openstack.utils.pause_unit() to do the work.
@param configs: a templating.OSConfigRenderer() object
@returns None - this function is executed for its side-effect
"""
_pause_resume_helper(pause_unit, configs)
def resume_unit_helper(configs):
"""Helper function to resume a unit, and then call assess_status(...) in
effect, so that the status is correctly updated.
Uses charmhelpers.contrib.openstack.utils.resume_unit() to do the work.
@param configs: a templating.OSConfigRenderer() object
@returns None - this function is executed for its side-effect
"""
_pause_resume_helper(resume_unit, configs)
def _pause_resume_helper(f, configs):
"""Helper function that uses the make_assess_status_func(...) from
charmhelpers.contrib.openstack.utils to create an assess_status(...)
function that can be used with the pause/resume of the unit
@param f: the function to be used with the assess_status(...) function
@returns None - this function is executed for its side-effect
"""
# TODO(ajkavanagh) - ports= has been left off because of the race hazard
# that exists due to service_start()
_services, _ = get_managed_services_and_ports(services(), [])
f(assess_status_func(configs),
services=services(),
ports=None)
def run_in_apache():
"""Return true if cinder API is run under apache2 with mod_wsgi in
this release.
"""
return CompareOpenStackReleases(os_release('cinder-common')) >= 'ocata'
def disable_package_apache_site():
"""Ensure that the package-provided apache configuration is disabled to
prevent it from conflicting with the charm-provided version.
"""
if os.path.exists(PACKAGE_CINDER_API_CONF):
subprocess.check_call(['a2disconf', 'cinder-wsgi'])
def scrub_old_style_ceph():
"""Purge any legacy ceph configuration from install"""
# NOTE: purge old override file - no longer needed
if os.path.exists('/etc/init/cinder-volume.override'):
os.remove('/etc/init/cinder-volume.override')
# NOTE: purge any CEPH_ARGS data from /etc/environment
env_file = '/etc/environment'
ceph_match = re.compile("^CEPH_ARGS.*").search
with open(env_file, 'rt') as input_file:
with NamedTemporaryFile(mode='wt',
delete=False,
dir=os.path.dirname(env_file)) as outfile:
for line in input_file:
if not ceph_match(line):
print(line, end='', file=outfile)
os.rename(outfile.name, input_file.name)
View Git Blame Copy Permalink