karbor/karbor/services/protection/protection_plugins/volume/cinder_protection_plugin.py

374 lines
15 KiB
Python

# 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 functools import partial
import six
from cinderclient import exceptions as cinder_exc
from karbor.common import constants
from karbor import exception
from karbor.services.protection.client_factory import ClientFactory
from karbor.services.protection import protection_plugin
from karbor.services.protection.protection_plugins.volume \
import volume_plugin_cinder_schemas as cinder_schemas
from karbor.services.protection.restore_heat import HeatResource
from oslo_config import cfg
from oslo_log import log as logging
from oslo_service import loopingcall
from oslo_utils import uuidutils
LOG = logging.getLogger(__name__)
cinder_backup_opts = [
cfg.IntOpt(
'poll_interval', default=15,
help='Poll interval for Cinder backup status'
),
cfg.BoolOpt(
'backup_from_snapshot', default=True,
help='First take a snapshot of the volume, and backup from '
'it. Minimizes the time the volume is unavailable.'
),
]
def status_poll(get_status_func, interval, success_statuses=set(),
failure_statuses=set(), ignore_statuses=set(),
ignore_unexpected=False):
def _poll():
status = get_status_func()
if status in success_statuses:
raise loopingcall.LoopingCallDone(retvalue=True)
if status in failure_statuses:
raise loopingcall.LoopingCallDone(retvalue=False)
if status in ignore_statuses:
return
if ignore_unexpected is False:
raise loopingcall.LoopingCallDone(retvalue=False)
loop = loopingcall.FixedIntervalLoopingCall(_poll)
return loop.start(interval=interval, initial_delay=interval).wait()
def get_backup_status(cinder_client, backup_id):
return get_resource_status(cinder_client.backups, backup_id, 'backup')
def get_volume_status(cinder_client, volume_id):
return get_resource_status(cinder_client.volumes, volume_id, 'volume')
def get_snapshot_status(cinder_client, snapshot_id):
return get_resource_status(cinder_client.volume_snapshots, snapshot_id,
'snapshot')
def get_resource_status(resource_manager, resource_id, resource_type):
LOG.debug('Polling %(resource_type)s (id: %(resource_id)s)', {
'resource_type': resource_type,
'resource_id': resource_id,
})
try:
resource = resource_manager.get(resource_id)
status = resource.status
except cinder_exc.NotFound:
status = 'not-found'
LOG.debug(
'Polled %(resource_type)s (id: %(resource_id)s) status: %(status)s',
{
'resource_type': resource_type,
'resource_id': resource_id,
'status': status
}
)
return status
class ProtectOperation(protection_plugin.Operation):
def __init__(self, poll_interval, backup_from_snapshot):
super(ProtectOperation, self).__init__()
self._interval = poll_interval
self._backup_from_snapshot = backup_from_snapshot
self.snapshot_id = None
def _create_snapshot(self, cinder_client, volume_id):
snapshot = cinder_client.volume_snapshots.create(volume_id, force=True)
snapshot_id = snapshot.id
is_success = status_poll(
partial(get_snapshot_status, cinder_client, snapshot_id),
interval=self._interval,
success_statuses={'available', },
failure_statuses={'error', 'error_deleting', 'deleting',
'not-found'},
ignore_statuses={'creating', },
)
if not is_success:
raise Exception
return snapshot_id
def _delete_snapshot(self, cinder_client, snapshot_id):
LOG.info('Cleaning up snapshot (snapshot_id: %s)', snapshot_id)
cinder_client.volume_snapshots.delete(snapshot_id)
return status_poll(
partial(get_snapshot_status, cinder_client, snapshot_id),
interval=self._interval,
success_statuses={'not-found', },
failure_statuses={'error', 'error_deleting', 'creating'},
ignore_statuses={'deleting', },
)
def _create_backup(self, cinder_client, volume_id, backup_name,
description, snapshot_id=None, incremental=False,
container=None, force=False):
backup = cinder_client.backups.create(
volume_id=volume_id,
name=backup_name,
description=description,
force=force,
snapshot_id=snapshot_id,
incremental=incremental,
container=container
)
backup_id = backup.id
is_success = status_poll(
partial(get_backup_status, cinder_client, backup_id),
interval=self._interval,
success_statuses={'available'},
failure_statuses={'error'},
ignore_statuses={'creating'},
)
if not is_success:
try:
backup = cinder_client.backups.get(backup_id)
except Exception:
reason = 'Unable to find backup'
else:
reason = backup.fail_reason
raise Exception(reason)
return backup_id
def on_prepare_finish(self, checkpoint, resource, context, parameters,
**kwargs):
volume_id = resource.id
if not self._backup_from_snapshot:
LOG.info('Skipping taking snapshot of volume %s - backing up '
'directly', volume_id)
return
LOG.info('Taking snapshot of volume %s', volume_id)
bank_section = checkpoint.get_resource_bank_section(volume_id)
bank_section.update_object('status',
constants.RESOURCE_STATUS_PROTECTING)
cinder_client = ClientFactory.create_client('cinder', context)
try:
self.snapshot_id = self._create_snapshot(cinder_client, volume_id)
except Exception:
bank_section.update_object('status',
constants.RESOURCE_STATUS_ERROR)
raise exception.CreateBackupFailed(
reason='Error creating snapshot for volume',
resource_id=volume_id,
resource_type=constants.VOLUME_RESOURCE_TYPE,
)
def on_main(self, checkpoint, resource, context, parameters, **kwargs):
volume_id = resource.id
bank_section = checkpoint.get_resource_bank_section(volume_id)
cinder_client = ClientFactory.create_client('cinder', context)
LOG.info('creating volume backup, volume_id: %s', volume_id)
bank_section.update_object('status',
constants.RESOURCE_STATUS_PROTECTING)
resource_metadata = {
'volume_id': volume_id,
}
is_success = status_poll(
partial(get_volume_status, cinder_client, volume_id),
interval=self._interval,
success_statuses={'available', 'in-use', 'error_extending',
'error_restoring'},
failure_statuses={'error', 'error_deleting', 'deleting',
'not-found'},
ignore_statuses={'attaching', 'creating', 'backing-up',
'restoring-backup'},
)
if not is_success:
bank_section.update_object('status',
constants.RESOURCE_STATUS_ERROR)
raise exception.CreateBackupFailed(
reason='Volume is in errorneous state',
resource_id=volume_id,
resource_type=constants.VOLUME_RESOURCE_TYPE,
)
backup_name = parameters.get('backup_name', None)
description = parameters.get('description', None)
backup_mode = parameters.get('backup_mode', "full")
container = parameters.get('container', None)
force = parameters.get('force', False)
incremental = False
if backup_mode == "incremental":
incremental = True
elif backup_mode == "full":
incremental = False
try:
backup_id = self._create_backup(cinder_client, volume_id,
backup_name, description,
self.snapshot_id,
incremental, container, force)
except Exception as e:
LOG.error('Error creating backup (volume_id: %(volume_id)s '
'snapshot_id: %(snapshot_id)s): %(reason)s',
{'volume_id': volume_id,
'snapshot_id': self.snapshot_id,
'reason': e}
)
bank_section.update_object('status',
constants.RESOURCE_STATUS_ERROR)
raise exception.CreateBackupFailed(
reason=e,
resource_id=volume_id,
resource_type=constants.VOLUME_RESOURCE_TYPE,
)
resource_metadata['backup_id'] = backup_id
bank_section.update_object('metadata', resource_metadata)
bank_section.update_object('status',
constants.RESOURCE_STATUS_AVAILABLE)
LOG.info('Backed up volume (volume_id: %(volume_id)s snapshot_id: '
'%(snapshot_id)s backup_id: %(backup_id)s) successfully',
{'backup_id': backup_id,
'snapshot_id': self.snapshot_id,
'volume_id': volume_id}
)
if self.snapshot_id:
try:
self._delete_snapshot(cinder_client, self.snapshot_id)
except Exception as e:
LOG.warn('Failed deleting snapshot: %(snapshot_id)s. '
'Reason: %(reason)s',
{'snapshot_id': self.snapshot_id, 'reason': e}
)
class RestoreOperation(protection_plugin.Operation):
def on_main(self, checkpoint, resource, context, parameters, heat_template,
**kwargs):
resource_id = resource.id
bank_section = checkpoint.get_resource_bank_section(resource_id)
resource_metadata = bank_section.get_object('metadata')
name = parameters.get('restore_name',
'%s@%s' % (checkpoint.id, resource_id))
heat_resource_id = uuidutils.generate_uuid()
heat_resource = HeatResource(heat_resource_id,
constants.VOLUME_RESOURCE_TYPE)
heat_resource.set_property('name', name)
if 'restore_description' in parameters:
heat_resource.set_property('description',
parameters['restore_description'])
heat_resource.set_property('backup_id', resource_metadata['backup_id'])
heat_template.put_resource(resource_id, heat_resource)
class DeleteOperation(protection_plugin.Operation):
def __init__(self, poll_interval):
super(DeleteOperation, self).__init__()
self._interval = poll_interval
def on_main(self, checkpoint, resource, context, parameters, **kwargs):
resource_id = resource.id
bank_section = checkpoint.get_resource_bank_section(resource_id)
backup_id = None
try:
bank_section.update_object('status',
constants.RESOURCE_STATUS_DELETING)
resource_metadata = bank_section.get_object('metadata')
backup_id = resource_metadata['backup_id']
cinder_client = ClientFactory.create_client('cinder', context)
try:
backup = cinder_client.backups.get(backup_id)
cinder_client.backups.delete(backup)
except cinder_exc.NotFound:
LOG.info('Backup id: %s not found. Assuming deleted',
backup_id)
is_success = status_poll(
partial(get_backup_status, cinder_client, backup_id),
interval=self._interval,
success_statuses={'deleted', 'not-found'},
failure_statuses={'error', 'error_deleting'},
ignore_statuses={'deleting'},
)
if not is_success:
raise exception.NotFound()
bank_section.delete_object('metadata')
bank_section.update_object('status',
constants.RESOURCE_STATUS_DELETED)
except Exception as e:
LOG.error('delete volume backup failed, backup_id: %s', backup_id)
bank_section.update_object('status',
constants.RESOURCE_STATUS_ERROR)
raise exception.DeleteBackupFailed(
reason=six.text_type(e),
resource_id=resource_id,
resource_type=constants.VOLUME_RESOURCE_TYPE
)
class CinderBackupProtectionPlugin(protection_plugin.ProtectionPlugin):
_SUPPORT_RESOURCE_TYPES = [constants.VOLUME_RESOURCE_TYPE]
def __init__(self, config=None):
super(CinderBackupProtectionPlugin, self).__init__(config)
self._config.register_opts(cinder_backup_opts,
'cinder_backup_protection_plugin')
self._plugin_config = self._config.cinder_backup_protection_plugin
self._poll_interval = self._plugin_config.poll_interval
self._backup_from_snapshot = self._plugin_config.backup_from_snapshot
@classmethod
def get_supported_resources_types(cls):
return cls._SUPPORT_RESOURCE_TYPES
@classmethod
def get_options_schema(cls, resources_type):
return cinder_schemas.OPTIONS_SCHEMA
@classmethod
def get_restore_schema(cls, resources_type):
return cinder_schemas.RESTORE_SCHEMA
@classmethod
def get_saved_info_schema(cls, resources_type):
return cinder_schemas.SAVED_INFO_SCHEMA
@classmethod
def get_saved_info(cls, metadata_store, resource):
# TODO(hurong)
pass
def get_protect_operation(self, resource):
return ProtectOperation(self._poll_interval,
self._backup_from_snapshot)
def get_restore_operation(self, resource):
return RestoreOperation()
def get_delete_operation(self, resource):
return DeleteOperation(self._poll_interval)