c5368a7394
The Dell SC driver is missing options for Data Reduction, Group QOS and Volume QOS. Added. Change-Id: I01951e41eea9985471830b5410892be37e13805a
1791 lines
80 KiB
Python
1791 lines
80 KiB
Python
# Copyright 2016 Dell 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.
|
|
|
|
import eventlet
|
|
from oslo_config import cfg
|
|
from oslo_config import types
|
|
from oslo_log import log as logging
|
|
from oslo_utils import excutils
|
|
import six
|
|
|
|
from cinder import exception
|
|
from cinder.i18n import _, _LE, _LI, _LW
|
|
from cinder.objects import fields
|
|
from cinder.volume import driver
|
|
from cinder.volume.drivers.dell import dell_storagecenter_api
|
|
from cinder.volume.drivers.san.san import san_opts
|
|
from cinder.volume import volume_types
|
|
|
|
|
|
common_opts = [
|
|
cfg.IntOpt('dell_sc_ssn',
|
|
default=64702,
|
|
help='Storage Center System Serial Number'),
|
|
cfg.PortOpt('dell_sc_api_port',
|
|
default=3033,
|
|
help='Dell API port'),
|
|
cfg.StrOpt('dell_sc_server_folder',
|
|
default='openstack',
|
|
help='Name of the server folder to use on the Storage Center'),
|
|
cfg.StrOpt('dell_sc_volume_folder',
|
|
default='openstack',
|
|
help='Name of the volume folder to use on the Storage Center'),
|
|
cfg.BoolOpt('dell_sc_verify_cert',
|
|
default=False,
|
|
help='Enable HTTPS SC certificate verification'),
|
|
cfg.StrOpt('secondary_san_ip',
|
|
default='',
|
|
help='IP address of secondary DSM controller'),
|
|
cfg.StrOpt('secondary_san_login',
|
|
default='Admin',
|
|
help='Secondary DSM user name'),
|
|
cfg.StrOpt('secondary_san_password',
|
|
default='',
|
|
help='Secondary DSM user password name',
|
|
secret=True),
|
|
cfg.PortOpt('secondary_sc_api_port',
|
|
default=3033,
|
|
help='Secondary Dell API port'),
|
|
cfg.MultiOpt('excluded_domain_ip',
|
|
item_type=types.IPAddress(),
|
|
default=None,
|
|
help='Domain IP to be excluded from iSCSI returns.'),
|
|
cfg.StrOpt('dell_server_os',
|
|
default='Red Hat Linux 6.x',
|
|
help='Server OS type to use when creating a new server on the '
|
|
'Storage Center.')
|
|
]
|
|
|
|
LOG = logging.getLogger(__name__)
|
|
|
|
CONF = cfg.CONF
|
|
CONF.register_opts(common_opts)
|
|
|
|
|
|
class DellCommonDriver(driver.ConsistencyGroupVD, driver.ManageableVD,
|
|
driver.ExtendVD, driver.ManageableSnapshotsVD,
|
|
driver.SnapshotVD, driver.BaseVD):
|
|
|
|
def __init__(self, *args, **kwargs):
|
|
super(DellCommonDriver, self).__init__(*args, **kwargs)
|
|
self.configuration.append_config_values(common_opts)
|
|
self.configuration.append_config_values(san_opts)
|
|
self.backend_name =\
|
|
self.configuration.safe_get('volume_backend_name') or 'Dell'
|
|
self.backends = self.configuration.safe_get('replication_device')
|
|
self.replication_enabled = True if self.backends else False
|
|
self.is_direct_connect = False
|
|
self.active_backend_id = kwargs.get('active_backend_id', None)
|
|
self.failed_over = True if self.active_backend_id else False
|
|
LOG.info(_LI('Loading %(name)s: Failover state is %(state)r'),
|
|
{'name': self.backend_name,
|
|
'state': self.failed_over})
|
|
self.storage_protocol = 'iSCSI'
|
|
self.failback_timeout = 30
|
|
|
|
def _bytes_to_gb(self, spacestring):
|
|
"""Space is returned in a string like ...
|
|
|
|
7.38197504E8 Bytes
|
|
Need to split that apart and convert to GB.
|
|
|
|
:returns: gbs in int form
|
|
"""
|
|
try:
|
|
n = spacestring.split(' ', 1)
|
|
fgbs = float(n[0]) / 1073741824.0
|
|
igbs = int(fgbs)
|
|
return igbs
|
|
except Exception:
|
|
# If any of that blew up it isn't in the format we
|
|
# thought so eat our error and return None
|
|
return None
|
|
|
|
def do_setup(self, context):
|
|
"""One time driver setup.
|
|
|
|
Called once by the manager after the driver is loaded.
|
|
Sets up clients, check licenses, sets up protocol
|
|
specific helpers.
|
|
"""
|
|
self._client = dell_storagecenter_api.StorageCenterApiHelper(
|
|
self.configuration, self.active_backend_id, self.storage_protocol)
|
|
|
|
def check_for_setup_error(self):
|
|
"""Validates the configuration information."""
|
|
with self._client.open_connection() as api:
|
|
api.find_sc()
|
|
self.is_direct_connect = api.is_direct_connect
|
|
if self.is_direct_connect and self.replication_enabled:
|
|
msg = _('Dell Cinder driver configuration error replication '
|
|
'not supported with direct connect.')
|
|
raise exception.InvalidHost(reason=msg)
|
|
|
|
# If we are a healthy replicated system make sure our backend
|
|
# is alive.
|
|
if self.replication_enabled and not self.failed_over:
|
|
# Check that our replication destinations are available.
|
|
for backend in self.backends:
|
|
replssn = backend['target_device_id']
|
|
try:
|
|
# Just do a find_sc on it. If it raises we catch
|
|
# that and raise with a correct exception.
|
|
api.find_sc(int(replssn))
|
|
except exception.VolumeBackendAPIException:
|
|
msg = _('Dell Cinder driver configuration error '
|
|
'replication_device %s not found') % replssn
|
|
raise exception.InvalidHost(reason=msg)
|
|
|
|
def _get_volume_extra_specs(self, obj):
|
|
"""Gets extra specs for the given object."""
|
|
type_id = obj.get('volume_type_id')
|
|
if type_id:
|
|
return volume_types.get_volume_type_extra_specs(type_id)
|
|
|
|
return {}
|
|
|
|
def _add_volume_to_consistency_group(self, api, scvolume, volume):
|
|
"""Just a helper to add a volume to a consistency group.
|
|
|
|
:param api: Dell SC API opbject.
|
|
:param scvolume: Dell SC Volume object.
|
|
:param volume: Cinder Volume object.
|
|
:returns: Nothing.
|
|
"""
|
|
if scvolume and volume.get('consistencygroup_id'):
|
|
profile = api.find_replay_profile(
|
|
volume.get('consistencygroup_id'))
|
|
if profile:
|
|
api.update_cg_volumes(profile, [volume])
|
|
|
|
def _get_replication_specs(self, obj):
|
|
"""Checks if we can do replication.
|
|
|
|
Need the extra spec set and we have to be talking to EM.
|
|
|
|
:param obj: Cinder Volume or snapshot object.
|
|
:return: rinfo dict.
|
|
"""
|
|
rinfo = {'enabled': False, 'sync': False,
|
|
'live': False, 'active': False,
|
|
'autofailover': False}
|
|
# Repl does not work with direct connect.
|
|
if not self.is_direct_connect:
|
|
specs = self._get_volume_extra_specs(obj)
|
|
if (not self.failed_over and
|
|
specs.get('replication_enabled') == '<is> True'):
|
|
rinfo['enabled'] = True
|
|
if specs.get('replication_type') == '<in> sync':
|
|
rinfo['sync'] = True
|
|
if specs.get('replication:livevolume') == '<is> True':
|
|
rinfo['live'] = True
|
|
if specs.get('replication:livevolume:autofailover') == '<is> True':
|
|
rinfo['autofailover'] = True
|
|
if specs.get('replication:activereplay') == '<is> True':
|
|
rinfo['active'] = True
|
|
|
|
# Some quick checks.
|
|
if rinfo['enabled']:
|
|
replication_target_count = len(self.backends)
|
|
msg = None
|
|
if replication_target_count == 0:
|
|
msg = _(
|
|
'Replication setup failure: replication has been '
|
|
'enabled but no replication target has been specified '
|
|
'for this backend.')
|
|
if rinfo['live'] and replication_target_count != 1:
|
|
msg = _('Replication setup failure: replication:livevolume'
|
|
' has been enabled but more than one replication '
|
|
'target has been specified for this backend.')
|
|
if msg:
|
|
LOG.debug(msg)
|
|
raise exception.ReplicationError(message=msg)
|
|
# Got this far. Life is good. Return our data.
|
|
return rinfo
|
|
|
|
def _is_live_vol(self, obj):
|
|
rspecs = self._get_replication_specs(obj)
|
|
return rspecs['enabled'] and rspecs['live']
|
|
|
|
def _create_replications(self, api, volume, scvolume):
|
|
"""Creates any appropriate replications for a given volume.
|
|
|
|
:param api: Dell REST API object.
|
|
:param volume: Cinder volume object.
|
|
:param scvolume: Dell Storage Center Volume object.
|
|
:return: model_update
|
|
"""
|
|
# Replication V2
|
|
# for now we assume we have an array named backends.
|
|
replication_driver_data = None
|
|
# Replicate if we are supposed to.
|
|
rspecs = self._get_replication_specs(volume)
|
|
if rspecs['enabled']:
|
|
for backend in self.backends:
|
|
targetdeviceid = backend['target_device_id']
|
|
primaryqos = backend.get('qosnode', 'cinderqos')
|
|
secondaryqos = backend.get('remoteqos', 'cinderqos')
|
|
diskfolder = backend.get('diskfolder', None)
|
|
obj = None
|
|
if rspecs['live']:
|
|
# We are rolling with a live volume.
|
|
obj = api.create_live_volume(scvolume, targetdeviceid,
|
|
rspecs['active'],
|
|
rspecs['sync'],
|
|
rspecs['autofailover'],
|
|
primaryqos, secondaryqos)
|
|
else:
|
|
# Else a regular replication.
|
|
obj = api.create_replication(scvolume, targetdeviceid,
|
|
primaryqos, rspecs['sync'],
|
|
diskfolder, rspecs['active'])
|
|
# This is either a ScReplication object or a ScLiveVolume
|
|
# object. So long as it isn't None we are fine.
|
|
if not obj:
|
|
# Create replication will have printed a better error.
|
|
msg = _('Replication %(name)s to %(ssn)s failed.') % {
|
|
'name': volume['id'],
|
|
'ssn': targetdeviceid}
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
if not replication_driver_data:
|
|
replication_driver_data = backend['target_device_id']
|
|
else:
|
|
replication_driver_data += ','
|
|
replication_driver_data += backend['target_device_id']
|
|
# If we did something return model update.
|
|
model_update = {}
|
|
if replication_driver_data:
|
|
model_update = {
|
|
'replication_status': fields.ReplicationStatus.ENABLED,
|
|
'replication_driver_data': replication_driver_data}
|
|
return model_update
|
|
|
|
@staticmethod
|
|
def _cleanup_failed_create_volume(api, volumename):
|
|
try:
|
|
api.delete_volume(volumename)
|
|
except exception.VolumeBackendAPIException as ex:
|
|
LOG.info(_LI('Non fatal cleanup error: %s.'), ex.msg)
|
|
|
|
def create_volume(self, volume):
|
|
"""Create a volume."""
|
|
model_update = {}
|
|
|
|
# We use id as our name as it is unique.
|
|
volume_name = volume.get('id')
|
|
# Look for our volume
|
|
volume_size = volume.get('size')
|
|
|
|
LOG.debug('Creating volume %(name)s of size %(size)s',
|
|
{'name': volume_name,
|
|
'size': volume_size})
|
|
scvolume = None
|
|
with self._client.open_connection() as api:
|
|
try:
|
|
# Get our extra specs.
|
|
specs = self._get_volume_extra_specs(volume)
|
|
scvolume = api.create_volume(
|
|
volume_name, volume_size,
|
|
specs.get('storagetype:storageprofile'),
|
|
specs.get('storagetype:replayprofiles'),
|
|
specs.get('storagetype:volumeqos'),
|
|
specs.get('storagetype:groupqos'),
|
|
specs.get('storagetype:datareductionprofile'))
|
|
if scvolume is None:
|
|
raise exception.VolumeBackendAPIException(
|
|
message=_('Unable to create volume %s') %
|
|
volume_name)
|
|
|
|
# Update Consistency Group
|
|
self._add_volume_to_consistency_group(api, scvolume, volume)
|
|
|
|
# Create replications. (Or not. It checks.)
|
|
model_update = self._create_replications(api, volume, scvolume)
|
|
|
|
# Save our provider_id.
|
|
model_update['provider_id'] = scvolume['instanceId']
|
|
|
|
except Exception:
|
|
# if we actually created a volume but failed elsewhere
|
|
# clean up the volume now.
|
|
self._cleanup_failed_create_volume(api, volume_name)
|
|
with excutils.save_and_reraise_exception():
|
|
LOG.error(_LE('Failed to create volume %s'),
|
|
volume_name)
|
|
if scvolume is None:
|
|
raise exception.VolumeBackendAPIException(
|
|
data=_('Unable to create volume. Backend down.'))
|
|
|
|
return model_update
|
|
|
|
def _split_driver_data(self, replication_driver_data):
|
|
"""Splits the replication_driver_data into an array of ssn strings.
|
|
|
|
:param replication_driver_data: A string of comma separated SSNs.
|
|
:returns: SSNs in an array of strings.
|
|
"""
|
|
ssnstrings = []
|
|
# We have any replication_driver_data.
|
|
if replication_driver_data:
|
|
# Split the array and wiffle through the entries.
|
|
for str in replication_driver_data.split(','):
|
|
# Strip any junk from the string.
|
|
ssnstring = str.strip()
|
|
# Anything left?
|
|
if ssnstring:
|
|
# Add it to our array.
|
|
ssnstrings.append(ssnstring)
|
|
return ssnstrings
|
|
|
|
def _delete_live_volume(self, api, volume):
|
|
"""Delete live volume associated with volume.
|
|
|
|
:param api:Dell REST API object.
|
|
:param volume: Cinder Volume object
|
|
:return: True if we actually deleted something. False for everything
|
|
else.
|
|
"""
|
|
# Live Volume was added after provider_id support. So just assume it is
|
|
# there.
|
|
replication_driver_data = volume.get('replication_driver_data')
|
|
# Do we have any replication driver data?
|
|
if replication_driver_data:
|
|
# Valid replication data?
|
|
ssnstrings = self._split_driver_data(replication_driver_data)
|
|
if ssnstrings:
|
|
ssn = int(ssnstrings[0])
|
|
sclivevolume = api.get_live_volume(volume.get('provider_id'),
|
|
volume.get('id'))
|
|
# Have we found the live volume?
|
|
if (sclivevolume and
|
|
sclivevolume.get('secondaryScSerialNumber') == ssn and
|
|
api.delete_live_volume(sclivevolume, True)):
|
|
LOG.info(_LI('%(vname)s\'s replication live volume has '
|
|
'been deleted from storage Center %(sc)s,'),
|
|
{'vname': volume.get('id'),
|
|
'sc': ssn})
|
|
return True
|
|
# If we are here either we do not have a live volume, we do not have
|
|
# one on our configured SC or we were not able to delete it.
|
|
# Either way, warn and leave.
|
|
LOG.warning(_LW('Unable to delete %s live volume.'),
|
|
volume.get('id'))
|
|
return False
|
|
|
|
def _delete_replications(self, api, volume):
|
|
"""Delete replications associated with a given volume.
|
|
|
|
We should be able to roll through the replication_driver_data list
|
|
of SSNs and delete replication objects between them and the source
|
|
volume.
|
|
|
|
:param api: Dell REST API object.
|
|
:param volume: Cinder Volume object
|
|
:return: None
|
|
"""
|
|
replication_driver_data = volume.get('replication_driver_data')
|
|
if replication_driver_data:
|
|
ssnstrings = self._split_driver_data(replication_driver_data)
|
|
volume_name = volume.get('id')
|
|
provider_id = volume.get('provider_id')
|
|
scvol = api.find_volume(volume_name, provider_id)
|
|
# This is just a string of ssns separated by commas.
|
|
# Trundle through these and delete them all.
|
|
for ssnstring in ssnstrings:
|
|
ssn = int(ssnstring)
|
|
# Are we a replication or a live volume?
|
|
if not api.delete_replication(scvol, ssn):
|
|
LOG.warning(_LW('Unable to delete replication of Volume '
|
|
'%(vname)s to Storage Center %(sc)s.'),
|
|
{'vname': volume_name,
|
|
'sc': ssnstring})
|
|
# If none of that worked or there was nothing to do doesn't matter.
|
|
# Just move on.
|
|
|
|
def delete_volume(self, volume):
|
|
deleted = False
|
|
# We use id as our name as it is unique.
|
|
volume_name = volume.get('id')
|
|
provider_id = volume.get('provider_id')
|
|
LOG.debug('Deleting volume %s', volume_name)
|
|
with self._client.open_connection() as api:
|
|
try:
|
|
rspecs = self._get_replication_specs(volume)
|
|
if rspecs['enabled']:
|
|
if rspecs['live']:
|
|
self._delete_live_volume(api, volume)
|
|
else:
|
|
self._delete_replications(api, volume)
|
|
deleted = api.delete_volume(volume_name, provider_id)
|
|
except Exception:
|
|
with excutils.save_and_reraise_exception():
|
|
LOG.error(_LE('Failed to delete volume %s'),
|
|
volume_name)
|
|
|
|
# if there was an error we will have raised an
|
|
# exception. If it failed to delete it is because
|
|
# the conditions to delete a volume were not met.
|
|
if deleted is False:
|
|
raise exception.VolumeIsBusy(volume_name=volume_name)
|
|
|
|
def create_snapshot(self, snapshot):
|
|
"""Create snapshot"""
|
|
# our volume name is the volume id
|
|
volume_name = snapshot.get('volume_id')
|
|
provider_id = snapshot.volume.get('provider_id')
|
|
snapshot_id = snapshot.get('id')
|
|
LOG.debug('Creating snapshot %(snap)s on volume %(vol)s',
|
|
{'snap': snapshot_id,
|
|
'vol': volume_name})
|
|
with self._client.open_connection() as api:
|
|
scvolume = api.find_volume(volume_name, provider_id,
|
|
self._is_live_vol(snapshot))
|
|
if scvolume is not None:
|
|
replay = api.create_replay(scvolume, snapshot_id, 0)
|
|
if replay:
|
|
return {'status': fields.SnapshotStatus.AVAILABLE,
|
|
'provider_id': scvolume['instanceId']}
|
|
else:
|
|
LOG.warning(_LW('Unable to locate volume:%s'),
|
|
volume_name)
|
|
|
|
snapshot['status'] = fields.SnapshotStatus.ERROR
|
|
msg = _('Failed to create snapshot %s') % snapshot_id
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
|
|
def create_volume_from_snapshot(self, volume, snapshot):
|
|
"""Create new volume from other volume's snapshot on appliance."""
|
|
model_update = {}
|
|
scvolume = None
|
|
volume_name = volume.get('id')
|
|
src_provider_id = snapshot.get('provider_id')
|
|
src_volume_name = snapshot.get('volume_id')
|
|
# This snapshot could have been created on its own or as part of a
|
|
# cgsnapshot. If it was a cgsnapshot it will be identified on the Dell
|
|
# backend under cgsnapshot_id. Given the volume ID and the
|
|
# cgsnapshot_id we can find the appropriate snapshot.
|
|
# So first we look for cgsnapshot_id. If that is blank then it must
|
|
# have been a normal snapshot which will be found under snapshot_id.
|
|
snapshot_id = snapshot.get('cgsnapshot_id')
|
|
if not snapshot_id:
|
|
snapshot_id = snapshot.get('id')
|
|
LOG.debug(
|
|
'Creating new volume %(vol)s from snapshot %(snap)s '
|
|
'from vol %(src)s',
|
|
{'vol': volume_name,
|
|
'snap': snapshot_id,
|
|
'src': src_volume_name})
|
|
with self._client.open_connection() as api:
|
|
try:
|
|
srcvol = api.find_volume(src_volume_name, src_provider_id)
|
|
if srcvol is not None:
|
|
replay = api.find_replay(srcvol, snapshot_id)
|
|
if replay is not None:
|
|
# See if we have any extra specs.
|
|
specs = self._get_volume_extra_specs(volume)
|
|
scvolume = api.create_view_volume(
|
|
volume_name, replay,
|
|
specs.get('storagetype:replayprofiles'),
|
|
specs.get('storagetype:volumeqos'),
|
|
specs.get('storagetype:groupqos'),
|
|
specs.get('storagetype:datareductionprofile'))
|
|
|
|
# Extend Volume
|
|
if scvolume and (volume['size'] >
|
|
snapshot["volume_size"]):
|
|
LOG.debug('Resize the new volume to %s.',
|
|
volume['size'])
|
|
scvolume = api.expand_volume(scvolume,
|
|
volume['size'])
|
|
if scvolume is None:
|
|
raise exception.VolumeBackendAPIException(
|
|
message=_('Unable to create volume '
|
|
'%(name)s from %(snap)s.') %
|
|
{'name': volume_name,
|
|
'snap': snapshot_id})
|
|
|
|
# Update Consistency Group
|
|
self._add_volume_to_consistency_group(api,
|
|
scvolume,
|
|
volume)
|
|
# Replicate if we are supposed to.
|
|
model_update = self._create_replications(api,
|
|
volume,
|
|
scvolume)
|
|
# Save our instanceid.
|
|
model_update['provider_id'] = (
|
|
scvolume['instanceId'])
|
|
|
|
except Exception:
|
|
# Clean up after ourselves.
|
|
self._cleanup_failed_create_volume(api, volume_name)
|
|
with excutils.save_and_reraise_exception():
|
|
LOG.error(_LE('Failed to create volume %s'),
|
|
volume_name)
|
|
if scvolume is not None:
|
|
LOG.debug('Volume %(vol)s created from %(snap)s',
|
|
{'vol': volume_name,
|
|
'snap': snapshot_id})
|
|
else:
|
|
msg = _('Failed to create volume %s') % volume_name
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
|
|
return model_update
|
|
|
|
def create_cloned_volume(self, volume, src_vref):
|
|
"""Creates a clone of the specified volume."""
|
|
model_update = {}
|
|
scvolume = None
|
|
src_volume_name = src_vref.get('id')
|
|
src_provider_id = src_vref.get('provider_id')
|
|
volume_name = volume.get('id')
|
|
LOG.debug('Creating cloned volume %(clone)s from volume %(vol)s',
|
|
{'clone': volume_name,
|
|
'vol': src_volume_name})
|
|
with self._client.open_connection() as api:
|
|
try:
|
|
srcvol = api.find_volume(src_volume_name, src_provider_id)
|
|
if srcvol is not None:
|
|
# Get our specs.
|
|
specs = self._get_volume_extra_specs(volume)
|
|
# Create our volume
|
|
scvolume = api.create_cloned_volume(
|
|
volume_name, srcvol,
|
|
specs.get('storagetype:replayprofiles'),
|
|
specs.get('storagetype:volumeqos'),
|
|
specs.get('storagetype:groupqos'),
|
|
specs.get('storagetype:datareductionprofile'))
|
|
|
|
# Extend Volume
|
|
if scvolume and volume['size'] > src_vref['size']:
|
|
LOG.debug('Resize the volume to %s.', volume['size'])
|
|
scvolume = api.expand_volume(scvolume, volume['size'])
|
|
|
|
# If either of those didn't work we bail.
|
|
if scvolume is None:
|
|
raise exception.VolumeBackendAPIException(
|
|
message=_('Unable to create volume '
|
|
'%(name)s from %(vol)s.') %
|
|
{'name': volume_name,
|
|
'vol': src_volume_name})
|
|
|
|
# Update Consistency Group
|
|
self._add_volume_to_consistency_group(api,
|
|
scvolume,
|
|
volume)
|
|
# Replicate if we are supposed to.
|
|
model_update = self._create_replications(api,
|
|
volume,
|
|
scvolume)
|
|
# Save our provider_id.
|
|
model_update['provider_id'] = scvolume['instanceId']
|
|
except Exception:
|
|
# Clean up after ourselves.
|
|
self._cleanup_failed_create_volume(api, volume_name)
|
|
with excutils.save_and_reraise_exception():
|
|
LOG.error(_LE('Failed to create volume %s'),
|
|
volume_name)
|
|
if scvolume is not None:
|
|
LOG.debug('Volume %(vol)s cloned from %(src)s',
|
|
{'vol': volume_name,
|
|
'src': src_volume_name})
|
|
else:
|
|
msg = _('Failed to create volume %s') % volume_name
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
return model_update
|
|
|
|
def delete_snapshot(self, snapshot):
|
|
"""delete_snapshot"""
|
|
volume_name = snapshot.get('volume_id')
|
|
snapshot_id = snapshot.get('id')
|
|
provider_id = snapshot.get('provider_id')
|
|
LOG.debug('Deleting snapshot %(snap)s from volume %(vol)s',
|
|
{'snap': snapshot_id,
|
|
'vol': volume_name})
|
|
with self._client.open_connection() as api:
|
|
scvolume = api.find_volume(volume_name, provider_id)
|
|
if scvolume and api.delete_replay(scvolume, snapshot_id):
|
|
return
|
|
# if we are here things went poorly.
|
|
snapshot['status'] = fields.SnapshotStatus.ERROR_DELETING
|
|
msg = _('Failed to delete snapshot %s') % snapshot_id
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
|
|
def create_export(self, context, volume, connector):
|
|
"""Create an export of a volume.
|
|
|
|
The volume exists on creation and will be visible on
|
|
initialize connection. So nothing to do here.
|
|
"""
|
|
pass
|
|
|
|
def ensure_export(self, context, volume):
|
|
"""Ensure an export of a volume.
|
|
|
|
Per the eqlx driver we just make sure that the volume actually
|
|
exists where we think it does.
|
|
"""
|
|
scvolume = None
|
|
volume_name = volume.get('id')
|
|
provider_id = volume.get('provider_id')
|
|
LOG.debug('Checking existence of volume %s', volume_name)
|
|
with self._client.open_connection() as api:
|
|
try:
|
|
scvolume = api.find_volume(volume_name, provider_id,
|
|
self._is_live_vol(volume))
|
|
except Exception:
|
|
with excutils.save_and_reraise_exception():
|
|
LOG.error(_LE('Failed to ensure export of volume %s'),
|
|
volume_name)
|
|
if scvolume is None:
|
|
msg = _('Unable to find volume %s') % volume_name
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
|
|
def remove_export(self, context, volume):
|
|
"""Remove an export of a volume.
|
|
|
|
We do nothing here to match the nothing we do in create export. Again
|
|
we do everything in initialize and terminate connection.
|
|
"""
|
|
pass
|
|
|
|
def extend_volume(self, volume, new_size):
|
|
"""Extend the size of the volume."""
|
|
volume_name = volume.get('id')
|
|
provider_id = volume.get('provider_id')
|
|
LOG.debug('Extending volume %(vol)s to %(size)s',
|
|
{'vol': volume_name,
|
|
'size': new_size})
|
|
if volume is not None:
|
|
with self._client.open_connection() as api:
|
|
scvolume = api.find_volume(volume_name, provider_id)
|
|
if api.expand_volume(scvolume, new_size) is not None:
|
|
return
|
|
# If we are here nothing good happened.
|
|
msg = _('Unable to extend volume %s') % volume_name
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
|
|
def get_volume_stats(self, refresh=False):
|
|
"""Get volume status.
|
|
|
|
If 'refresh' is True, run update the stats first.
|
|
"""
|
|
if refresh:
|
|
self._update_volume_stats()
|
|
|
|
# Take this opportunity to report our failover state.
|
|
if self.failed_over:
|
|
LOG.debug('%(source)s has been failed over to %(dest)s',
|
|
{'source': self.backend_name,
|
|
'dest': self.active_backend_id})
|
|
return self._stats
|
|
|
|
def _update_volume_stats(self):
|
|
"""Retrieve stats info from volume group."""
|
|
with self._client.open_connection() as api:
|
|
# Static stats.
|
|
data = {}
|
|
data['volume_backend_name'] = self.backend_name
|
|
data['vendor_name'] = 'Dell'
|
|
data['driver_version'] = self.VERSION
|
|
data['storage_protocol'] = self.storage_protocol
|
|
data['reserved_percentage'] = 0
|
|
data['consistencygroup_support'] = True
|
|
data['thin_provisioning_support'] = True
|
|
data['QoS_support'] = False
|
|
data['replication_enabled'] = self.replication_enabled
|
|
if self.replication_enabled:
|
|
data['replication_type'] = ['async', 'sync']
|
|
data['replication_count'] = len(self.backends)
|
|
replication_targets = []
|
|
# Trundle through our backends.
|
|
for backend in self.backends:
|
|
target_device_id = backend.get('target_device_id')
|
|
if target_device_id:
|
|
replication_targets.append(target_device_id)
|
|
data['replication_targets'] = replication_targets
|
|
|
|
# Get our capacity.
|
|
storageusage = api.get_storage_usage()
|
|
if storageusage:
|
|
# Get actual stats.
|
|
totalcapacity = storageusage.get('availableSpace')
|
|
totalcapacitygb = self._bytes_to_gb(totalcapacity)
|
|
data['total_capacity_gb'] = totalcapacitygb
|
|
freespace = storageusage.get('freeSpace')
|
|
freespacegb = self._bytes_to_gb(freespace)
|
|
data['free_capacity_gb'] = freespacegb
|
|
else:
|
|
# Soldier on. Just return 0 for this iteration.
|
|
LOG.error(_LE('Unable to retrieve volume stats.'))
|
|
data['total_capacity_gb'] = 0
|
|
data['free_capacity_gb'] = 0
|
|
|
|
self._stats = data
|
|
LOG.debug('Total cap %(total)s Free cap %(free)s',
|
|
{'total': data['total_capacity_gb'],
|
|
'free': data['free_capacity_gb']})
|
|
|
|
def update_migrated_volume(self, ctxt, volume, new_volume,
|
|
original_volume_status):
|
|
"""Return model update for migrated volume.
|
|
|
|
:param volume: The original volume that was migrated to this backend
|
|
:param new_volume: The migration volume object that was created on
|
|
this backend as part of the migration process
|
|
:param original_volume_status: The status of the original volume
|
|
:returns: model_update to update DB with any needed changes
|
|
"""
|
|
# We use id as our volume name so we need to rename the backend
|
|
# volume to the original volume name.
|
|
original_volume_name = volume.get('id')
|
|
current_name = new_volume.get('id')
|
|
# We should have this. If we don't we'll set it below.
|
|
provider_id = new_volume.get('provider_id')
|
|
LOG.debug('update_migrated_volume: %(current)s to %(original)s',
|
|
{'current': current_name,
|
|
'original': original_volume_name})
|
|
if original_volume_name:
|
|
with self._client.open_connection() as api:
|
|
scvolume = api.find_volume(current_name, provider_id)
|
|
if (scvolume and
|
|
api.rename_volume(scvolume, original_volume_name)):
|
|
# Replicate if we are supposed to.
|
|
model_update = self._create_replications(api,
|
|
new_volume,
|
|
scvolume)
|
|
model_update['_name_id'] = None
|
|
model_update['provider_id'] = scvolume['instanceId']
|
|
|
|
return model_update
|
|
# The world was horrible to us so we should error and leave.
|
|
LOG.error(_LE('Unable to rename the logical volume for volume: %s'),
|
|
original_volume_name)
|
|
|
|
return {'_name_id': new_volume['_name_id'] or new_volume['id']}
|
|
|
|
def create_consistencygroup(self, context, group):
|
|
"""This creates a replay profile on the storage backend.
|
|
|
|
:param context: the context of the caller.
|
|
:param group: the dictionary of the consistency group to be created.
|
|
:returns: Nothing on success.
|
|
:raises: VolumeBackendAPIException
|
|
"""
|
|
gid = group['id']
|
|
with self._client.open_connection() as api:
|
|
cgroup = api.create_replay_profile(gid)
|
|
if cgroup:
|
|
LOG.info(_LI('Created Consistency Group %s'), gid)
|
|
return
|
|
msg = _('Unable to create consistency group %s') % gid
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
|
|
def delete_consistencygroup(self, context, group, volumes):
|
|
"""Delete the Dell SC profile associated with this consistency group.
|
|
|
|
:param context: the context of the caller.
|
|
:param group: the dictionary of the consistency group to be created.
|
|
:returns: Updated model_update, volumes.
|
|
"""
|
|
gid = group['id']
|
|
with self._client.open_connection() as api:
|
|
profile = api.find_replay_profile(gid)
|
|
if profile:
|
|
api.delete_replay_profile(profile)
|
|
# If we are here because we found no profile that should be fine
|
|
# as we are trying to delete it anyway.
|
|
|
|
# Trundle through the list deleting the volumes.
|
|
for volume in volumes:
|
|
self.delete_volume(volume)
|
|
volume['status'] = 'deleted'
|
|
|
|
model_update = {'status': group['status']}
|
|
|
|
return model_update, volumes
|
|
|
|
def update_consistencygroup(self, context, group,
|
|
add_volumes=None, remove_volumes=None):
|
|
"""Updates a consistency group.
|
|
|
|
:param context: the context of the caller.
|
|
:param group: the dictionary of the consistency group to be updated.
|
|
:param add_volumes: a list of volume dictionaries to be added.
|
|
:param remove_volumes: a list of volume dictionaries to be removed.
|
|
:returns: model_update, add_volumes_update, remove_volumes_update
|
|
|
|
model_update is a dictionary that the driver wants the manager
|
|
to update upon a successful return. If None is returned, the manager
|
|
will set the status to 'available'.
|
|
|
|
add_volumes_update and remove_volumes_update are lists of dictionaries
|
|
that the driver wants the manager to update upon a successful return.
|
|
Note that each entry requires a {'id': xxx} so that the correct
|
|
volume entry can be updated. If None is returned, the volume will
|
|
remain its original status. Also note that you cannot directly
|
|
assign add_volumes to add_volumes_update as add_volumes is a list of
|
|
cinder.db.sqlalchemy.models.Volume objects and cannot be used for
|
|
db update directly. Same with remove_volumes.
|
|
|
|
If the driver throws an exception, the status of the group as well as
|
|
those of the volumes to be added/removed will be set to 'error'.
|
|
"""
|
|
gid = group['id']
|
|
with self._client.open_connection() as api:
|
|
profile = api.find_replay_profile(gid)
|
|
if not profile:
|
|
LOG.error(_LE('Cannot find Consistency Group %s'), gid)
|
|
elif api.update_cg_volumes(profile,
|
|
add_volumes,
|
|
remove_volumes):
|
|
LOG.info(_LI('Updated Consistency Group %s'), gid)
|
|
# we need nothing updated above us so just return None.
|
|
return None, None, None
|
|
# Things did not go well so throw.
|
|
msg = _('Unable to update consistency group %s') % gid
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
|
|
def create_cgsnapshot(self, context, cgsnapshot, snapshots):
|
|
"""Takes a snapshot of the consistency group.
|
|
|
|
:param context: the context of the caller.
|
|
:param cgsnapshot: Information about the snapshot to take.
|
|
:param snapshots: List of snapshots for this cgsnapshot.
|
|
:returns: Updated model_update, snapshots.
|
|
:raises: VolumeBackendAPIException.
|
|
"""
|
|
cgid = cgsnapshot['consistencygroup_id']
|
|
snapshotid = cgsnapshot['id']
|
|
|
|
with self._client.open_connection() as api:
|
|
profile = api.find_replay_profile(cgid)
|
|
if profile:
|
|
LOG.debug('profile %s replayid %s', profile, snapshotid)
|
|
if api.snap_cg_replay(profile, snapshotid, 0):
|
|
snapshot_updates = []
|
|
for snapshot in snapshots:
|
|
snapshot_updates.append({
|
|
'id': snapshot.id,
|
|
'status': fields.SnapshotStatus.AVAILABLE
|
|
})
|
|
|
|
model_update = {'status': fields.SnapshotStatus.AVAILABLE}
|
|
|
|
return model_update, snapshot_updates
|
|
|
|
# That didn't go well. Tell them why. Then bomb out.
|
|
LOG.error(_LE('Failed to snap Consistency Group %s'), cgid)
|
|
else:
|
|
LOG.error(_LE('Cannot find Consistency Group %s'), cgid)
|
|
|
|
msg = _('Unable to snap Consistency Group %s') % cgid
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
|
|
def delete_cgsnapshot(self, context, cgsnapshot, snapshots):
|
|
"""Deletes a cgsnapshot.
|
|
|
|
If profile isn't found return success. If failed to delete the
|
|
replay (the snapshot) then raise an exception.
|
|
|
|
:param context: the context of the caller.
|
|
:param cgsnapshot: Information about the snapshot to delete.
|
|
:returns: Updated model_update, snapshots.
|
|
:raises: VolumeBackendAPIException.
|
|
"""
|
|
cgid = cgsnapshot['consistencygroup_id']
|
|
snapshotid = cgsnapshot['id']
|
|
|
|
with self._client.open_connection() as api:
|
|
profile = api.find_replay_profile(cgid)
|
|
if profile:
|
|
LOG.info(_LI('Deleting snapshot %(ss)s from %(pro)s'),
|
|
{'ss': snapshotid,
|
|
'pro': profile})
|
|
if not api.delete_cg_replay(profile, snapshotid):
|
|
msg = (_('Unable to delete Consistency Group snapshot %s')
|
|
% snapshotid)
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
|
|
for snapshot in snapshots:
|
|
snapshot.status = fields.SnapshotStatus.DELETED
|
|
|
|
model_update = {'status': fields.SnapshotStatus.DELETED}
|
|
|
|
return model_update, snapshots
|
|
|
|
def manage_existing(self, volume, existing_ref):
|
|
"""Brings an existing backend storage object under Cinder management.
|
|
|
|
existing_ref is passed straight through from the API request's
|
|
manage_existing_ref value, and it is up to the driver how this should
|
|
be interpreted. It should be sufficient to identify a storage object
|
|
that the driver should somehow associate with the newly-created cinder
|
|
volume structure.
|
|
|
|
There are two ways to do this:
|
|
|
|
1. Rename the backend storage object so that it matches the,
|
|
volume['name'] which is how drivers traditionally map between a
|
|
cinder volume and the associated backend storage object.
|
|
|
|
2. Place some metadata on the volume, or somewhere in the backend, that
|
|
allows other driver requests (e.g. delete, clone, attach, detach...)
|
|
to locate the backend storage object when required.
|
|
|
|
If the existing_ref doesn't make sense, or doesn't refer to an existing
|
|
backend storage object, raise a ManageExistingInvalidReference
|
|
exception.
|
|
|
|
The volume may have a volume_type, and the driver can inspect that and
|
|
compare against the properties of the referenced backend storage
|
|
object. If they are incompatible, raise a
|
|
ManageExistingVolumeTypeMismatch, specifying a reason for the failure.
|
|
|
|
:param volume: Cinder volume to manage
|
|
:param existing_ref: Driver-specific information used to identify a
|
|
volume
|
|
"""
|
|
if existing_ref.get('source-name') or existing_ref.get('source-id'):
|
|
with self._client.open_connection() as api:
|
|
api.manage_existing(volume['id'], existing_ref)
|
|
# Replicate if we are supposed to.
|
|
volume_name = volume.get('id')
|
|
provider_id = volume.get('provider_id')
|
|
scvolume = api.find_volume(volume_name, provider_id)
|
|
model_update = self._create_replications(api, volume, scvolume)
|
|
if model_update:
|
|
return model_update
|
|
else:
|
|
msg = _('Must specify source-name or source-id.')
|
|
raise exception.ManageExistingInvalidReference(
|
|
existing_ref=existing_ref, reason=msg)
|
|
# Only return a model_update if we have replication info to add.
|
|
return None
|
|
|
|
def manage_existing_get_size(self, volume, existing_ref):
|
|
"""Return size of volume to be managed by manage_existing.
|
|
|
|
When calculating the size, round up to the next GB.
|
|
|
|
:param volume: Cinder volume to manage
|
|
:param existing_ref: Driver-specific information used to identify a
|
|
volume
|
|
"""
|
|
if existing_ref.get('source-name') or existing_ref.get('source-id'):
|
|
with self._client.open_connection() as api:
|
|
return api.get_unmanaged_volume_size(existing_ref)
|
|
else:
|
|
msg = _('Must specify source-name or source-id.')
|
|
raise exception.ManageExistingInvalidReference(
|
|
existing_ref=existing_ref, reason=msg)
|
|
|
|
def unmanage(self, volume):
|
|
"""Removes the specified volume from Cinder management.
|
|
|
|
Does not delete the underlying backend storage object.
|
|
|
|
For most drivers, this will not need to do anything. However, some
|
|
drivers might use this call as an opportunity to clean up any
|
|
Cinder-specific configuration that they have associated with the
|
|
backend storage object.
|
|
|
|
:param volume: Cinder volume to unmanage
|
|
"""
|
|
with self._client.open_connection() as api:
|
|
volume_name = volume.get('id')
|
|
provider_id = volume.get('provider_id')
|
|
scvolume = api.find_volume(volume_name, provider_id)
|
|
if scvolume:
|
|
api.unmanage(scvolume)
|
|
|
|
def _get_retype_spec(self, diff, volume_name, specname, spectype):
|
|
"""Helper function to get current and requested spec.
|
|
|
|
:param diff: A difference dictionary.
|
|
:param volume_name: The volume name we are working with.
|
|
:param specname: The pretty name of the parameter.
|
|
:param spectype: The actual spec string.
|
|
:return: current, requested spec.
|
|
:raises: VolumeBackendAPIException
|
|
"""
|
|
spec = (diff['extra_specs'].get(spectype))
|
|
if spec:
|
|
if len(spec) != 2:
|
|
msg = _('Unable to retype %(specname)s, expected to receive '
|
|
'current and requested %(spectype)s values. Value '
|
|
'received: %(spec)s') % {'specname': specname,
|
|
'spectype': spectype,
|
|
'spec': spec}
|
|
LOG.error(msg)
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
|
|
current = spec[0]
|
|
requested = spec[1]
|
|
|
|
if current != requested:
|
|
LOG.debug('Retyping volume %(vol)s to use %(specname)s '
|
|
'%(spec)s.',
|
|
{'vol': volume_name,
|
|
'specname': specname,
|
|
'spec': requested})
|
|
return current, requested
|
|
else:
|
|
LOG.info(_LI('Retype was to same Storage Profile.'))
|
|
return None, None
|
|
|
|
def retype(self, ctxt, volume, new_type, diff, host):
|
|
"""Convert the volume to be of the new type.
|
|
|
|
Returns a boolean indicating whether the retype occurred.
|
|
|
|
:param ctxt: Context
|
|
:param volume: A dictionary describing the volume to migrate
|
|
:param new_type: A dictionary describing the volume type to convert to
|
|
:param diff: A dictionary with the difference between the two types
|
|
:param host: A dictionary describing the host to migrate to, where
|
|
host['host'] is its name, and host['capabilities'] is a
|
|
dictionary of its reported capabilities (Not Used).
|
|
"""
|
|
LOG.info(_LI('retype: volume_name: %(name)s new_type: %(newtype)s '
|
|
'diff: %(diff)s host: %(host)s'),
|
|
{'name': volume.get('id'), 'newtype': new_type,
|
|
'diff': diff, 'host': host})
|
|
model_update = None
|
|
# Any spec changes?
|
|
if diff['extra_specs']:
|
|
volume_name = volume.get('id')
|
|
provider_id = volume.get('provider_id')
|
|
with self._client.open_connection() as api:
|
|
try:
|
|
# Get our volume
|
|
scvolume = api.find_volume(volume_name, provider_id)
|
|
if scvolume is None:
|
|
LOG.error(_LE('Retype unable to find volume %s.'),
|
|
volume_name)
|
|
return False
|
|
# Check our specs.
|
|
# Storage profiles.
|
|
current, requested = (
|
|
self._get_retype_spec(diff, volume_name,
|
|
'Storage Profile',
|
|
'storagetype:storageprofile'))
|
|
# if there is a change and it didn't work fast fail.
|
|
if (current != requested and not
|
|
api.update_storage_profile(scvolume, requested)):
|
|
LOG.error(_LE('Failed to update storage profile'))
|
|
return False
|
|
|
|
# Replay profiles.
|
|
current, requested = (
|
|
self._get_retype_spec(diff, volume_name,
|
|
'Replay Profiles',
|
|
'storagetype:replayprofiles'))
|
|
# if there is a change and it didn't work fast fail.
|
|
if requested and not api.update_replay_profiles(scvolume,
|
|
requested):
|
|
LOG.error(_LE('Failed to update replay profiles'))
|
|
return False
|
|
|
|
# Volume QOS profiles.
|
|
current, requested = (
|
|
self._get_retype_spec(diff, volume_name,
|
|
'Volume QOS Profile',
|
|
'storagetype:volumeqos'))
|
|
if current != requested:
|
|
if not api.update_qos_profile(scvolume, requested):
|
|
LOG.error(_LE('Failed to update volume '
|
|
'qos profile'))
|
|
|
|
# Group QOS profiles.
|
|
current, requested = (
|
|
self._get_retype_spec(diff, volume_name,
|
|
'Group QOS Profile',
|
|
'storagetype:groupqos'))
|
|
if current != requested:
|
|
if not api.update_qos_profile(scvolume, requested,
|
|
True):
|
|
LOG.error(_LE('Failed to update group '
|
|
'qos profile'))
|
|
return False
|
|
|
|
# Data reduction profiles.
|
|
current, requested = (
|
|
self._get_retype_spec(
|
|
diff, volume_name, 'Data Reduction Profile',
|
|
'storagetype:datareductionprofile'))
|
|
if current != requested:
|
|
if not api.update_datareduction_profile(scvolume,
|
|
requested):
|
|
LOG.error(_LE('Failed to update data reduction '
|
|
'profile'))
|
|
return False
|
|
|
|
# Replication_enabled.
|
|
current, requested = (
|
|
self._get_retype_spec(diff,
|
|
volume_name,
|
|
'replication_enabled',
|
|
'replication_enabled'))
|
|
# if there is a change and it didn't work fast fail.
|
|
if current != requested:
|
|
if requested == '<is> True':
|
|
model_update = self._create_replications(api,
|
|
volume,
|
|
scvolume)
|
|
elif current == '<is> True':
|
|
self._delete_replications(api, volume)
|
|
model_update = {'replication_status':
|
|
fields.ReplicationStatus.DISABLED,
|
|
'replication_driver_data': ''}
|
|
|
|
# Active Replay
|
|
current, requested = (
|
|
self._get_retype_spec(diff, volume_name,
|
|
'Replicate Active Replay',
|
|
'replication:activereplay'))
|
|
if current != requested and not (
|
|
api.update_replicate_active_replay(
|
|
scvolume, requested == '<is> True')):
|
|
LOG.error(_LE('Failed to apply '
|
|
'replication:activereplay setting'))
|
|
return False
|
|
|
|
except exception.VolumeBackendAPIException:
|
|
# We do nothing with this. We simply return failure.
|
|
return False
|
|
# If we have something to send down...
|
|
if model_update:
|
|
return model_update
|
|
return True
|
|
|
|
def _parse_secondary(self, api, secondary):
|
|
"""Find the replication destination associated with secondary.
|
|
|
|
:param api: Dell StorageCenterApi
|
|
:param secondary: String indicating the secondary to failover to.
|
|
:return: Destination SSN for the given secondary.
|
|
"""
|
|
LOG.debug('_parse_secondary. Looking for %s.', secondary)
|
|
destssn = None
|
|
# Trundle through these looking for our secondary.
|
|
for backend in self.backends:
|
|
ssnstring = backend['target_device_id']
|
|
# If they list a secondary it has to match.
|
|
# If they do not list a secondary we return the first
|
|
# replication on a working system.
|
|
if not secondary or secondary == ssnstring:
|
|
# Is a string. Need an int.
|
|
ssn = int(ssnstring)
|
|
# Without the source being up we have no good
|
|
# way to pick a destination to failover to. So just
|
|
# look for one that is just up.
|
|
try:
|
|
# If the SC ssn exists use it.
|
|
if api.find_sc(ssn):
|
|
destssn = ssn
|
|
break
|
|
except exception.VolumeBackendAPIException:
|
|
LOG.warning(_LW('SSN %s appears to be down.'), ssn)
|
|
LOG.info(_LI('replication failover secondary is %(ssn)s'),
|
|
{'ssn': destssn})
|
|
return destssn
|
|
|
|
def _update_backend(self, active_backend_id):
|
|
# Mark for failover or undo failover.
|
|
LOG.debug('active_backend_id: %s', active_backend_id)
|
|
if active_backend_id:
|
|
self.active_backend_id = six.text_type(active_backend_id)
|
|
self.failed_over = True
|
|
else:
|
|
self.active_backend_id = None
|
|
self.failed_over = False
|
|
|
|
self._client.active_backend_id = self.active_backend_id
|
|
|
|
def _get_qos(self, targetssn):
|
|
# Find our QOS.
|
|
qosnode = None
|
|
for backend in self.backends:
|
|
if int(backend['target_device_id']) == targetssn:
|
|
qosnode = backend.get('qosnode', 'cinderqos')
|
|
return qosnode
|
|
|
|
def _parse_extraspecs(self, volume):
|
|
# Digest our extra specs for replication.
|
|
extraspecs = {}
|
|
specs = self._get_volume_extra_specs(volume)
|
|
if specs.get('replication_type') == '<in> sync':
|
|
extraspecs['replicationtype'] = 'Synchronous'
|
|
else:
|
|
extraspecs['replicationtype'] = 'Asynchronous'
|
|
if specs.get('replication:activereplay') == '<is> True':
|
|
extraspecs['activereplay'] = True
|
|
else:
|
|
extraspecs['activereplay'] = False
|
|
extraspecs['storage_profile'] = specs.get('storagetype:storageprofile')
|
|
extraspecs['replay_profile_string'] = (
|
|
specs.get('storagetype:replayprofiles'))
|
|
return extraspecs
|
|
|
|
def _wait_for_replication(self, api, items):
|
|
# Wait for our replications to resync with their original volumes.
|
|
# We wait for completion, errors or timeout.
|
|
deadcount = 5
|
|
lastremain = 0.0
|
|
# The big wait loop.
|
|
while True:
|
|
# We run until all volumes are synced or in error.
|
|
done = True
|
|
currentremain = 0.0
|
|
# Run the list.
|
|
for item in items:
|
|
# If we have one cooking.
|
|
if item['status'] == 'inprogress':
|
|
# Is it done?
|
|
synced, remain = api.replication_progress(item['screpl'])
|
|
currentremain += remain
|
|
if synced:
|
|
# It is! Get our volumes.
|
|
cvol = api.get_volume(item['cvol'])
|
|
nvol = api.get_volume(item['nvol'])
|
|
|
|
# Flip replication.
|
|
if (cvol and nvol and api.flip_replication(
|
|
cvol, nvol, item['volume']['id'],
|
|
item['specs']['replicationtype'],
|
|
item['qosnode'],
|
|
item['specs']['activereplay'])):
|
|
# rename the original. Doesn't matter if it
|
|
# succeeded as we should have the provider_id
|
|
# of the new volume.
|
|
ovol = api.get_volume(item['ovol'])
|
|
if not ovol or not api.rename_volume(
|
|
ovol, 'org:' + ovol['name']):
|
|
# Not a reason to fail but will possibly
|
|
# cause confusion so warn.
|
|
LOG.warning(_LW('Unable to locate and rename '
|
|
'original volume: %s'),
|
|
item['ovol'])
|
|
item['status'] = 'synced'
|
|
else:
|
|
item['status'] = 'error'
|
|
elif synced is None:
|
|
# Couldn't get info on this one. Call it baked.
|
|
item['status'] = 'error'
|
|
else:
|
|
# Miles to go before we're done.
|
|
done = False
|
|
# done? then leave.
|
|
if done:
|
|
break
|
|
|
|
# Confirm we are or are not still making progress.
|
|
if lastremain == currentremain:
|
|
# One chance down. Warn user.
|
|
deadcount -= 1
|
|
LOG.warning(_LW('Waiting for replications to complete. '
|
|
'No progress for 30 seconds. deadcount = %d'),
|
|
deadcount)
|
|
else:
|
|
# Reset
|
|
lastremain = currentremain
|
|
deadcount = 5
|
|
|
|
# If we've used up our 5 chances we error and log..
|
|
if deadcount == 0:
|
|
LOG.error(_LE('Replication progress has stopped.'))
|
|
for item in items:
|
|
if item['status'] == 'inprogress':
|
|
LOG.error(_LE('Failback failed for volume: %s. '
|
|
'Timeout waiting for replication to '
|
|
'sync with original volume.'),
|
|
item['volume']['id'])
|
|
item['status'] = 'error'
|
|
break
|
|
# This is part of an async call so we should be good sleeping here.
|
|
# Have to balance hammering the backend for no good reason with
|
|
# the max timeout for the unit tests. Yeah, silly.
|
|
eventlet.sleep(self.failback_timeout)
|
|
|
|
def _reattach_remaining_replications(self, api, items):
|
|
# Wiffle through our backends and reattach any remaining replication
|
|
# targets.
|
|
for item in items:
|
|
if item['status'] == 'synced':
|
|
svol = api.get_volume(item['nvol'])
|
|
# assume it went well. Will error out if not.
|
|
item['status'] = 'reattached'
|
|
# wiffle through our backends and kick off replications.
|
|
for backend in self.backends:
|
|
rssn = int(backend['target_device_id'])
|
|
if rssn != api.ssn:
|
|
rvol = api.find_repl_volume(item['volume']['id'],
|
|
rssn, None)
|
|
# if there is an old replication whack it.
|
|
api.delete_replication(svol, rssn, False)
|
|
if api.start_replication(
|
|
svol, rvol,
|
|
item['specs']['replicationtype'],
|
|
self._get_qos(rssn),
|
|
item['specs']['activereplay']):
|
|
# Save our replication_driver_data.
|
|
item['rdd'] += ','
|
|
item['rdd'] += backend['target_device_id']
|
|
else:
|
|
# No joy. Bail
|
|
item['status'] = 'error'
|
|
|
|
def _fixup_types(self, api, items):
|
|
# Update our replay profiles.
|
|
for item in items:
|
|
if item['status'] == 'reattached':
|
|
# Re-apply any appropriate replay profiles.
|
|
item['status'] = 'available'
|
|
rps = item['specs']['replay_profile_string']
|
|
if rps:
|
|
svol = api.get_volume(item['nvol'])
|
|
if not api.update_replay_profiles(svol, rps):
|
|
item['status'] = 'error'
|
|
|
|
def _volume_updates(self, items):
|
|
# Update our volume updates.
|
|
volume_updates = []
|
|
for item in items:
|
|
# Set our status for our replicated volumes
|
|
model_update = {'provider_id': item['nvol'],
|
|
'replication_driver_data': item['rdd']}
|
|
# These are simple. If the volume reaches available then,
|
|
# since we were replicating it, replication status must
|
|
# be good. Else error/error.
|
|
if item['status'] == 'available':
|
|
model_update['status'] = 'available'
|
|
model_update['replication_status'] = (
|
|
fields.ReplicationStatus.ENABLED)
|
|
else:
|
|
model_update['status'] = 'error'
|
|
model_update['replication_status'] = (
|
|
fields.ReplicationStatus.ERROR)
|
|
volume_updates.append({'volume_id': item['volume']['id'],
|
|
'updates': model_update})
|
|
return volume_updates
|
|
|
|
def _failback_replication(self, api, volume, qosnode):
|
|
"""Sets up the replication failback.
|
|
|
|
:param api: Dell SC API.
|
|
:param volume: Cinder Volume
|
|
:param qosnode: Dell QOS node object.
|
|
:return: replitem dict.
|
|
"""
|
|
LOG.info(_LI('failback_volumes: replicated volume'))
|
|
# Get our current volume.
|
|
cvol = api.find_volume(volume['id'], volume['provider_id'])
|
|
# Original volume on the primary.
|
|
ovol = api.find_repl_volume(volume['id'], api.primaryssn,
|
|
None, True, False)
|
|
# Delete our current mappings.
|
|
api.remove_mappings(cvol)
|
|
# If there is a replication to delete do so.
|
|
api.delete_replication(ovol, api.ssn, False)
|
|
# Replicate to a common replay.
|
|
screpl = api.replicate_to_common(cvol, ovol, 'tempqos')
|
|
# We made it this far. Update our status.
|
|
screplid = None
|
|
status = ''
|
|
if screpl:
|
|
screplid = screpl['instanceId']
|
|
nvolid = screpl['destinationVolume']['instanceId']
|
|
status = 'inprogress'
|
|
else:
|
|
LOG.error(_LE('Unable to restore %s'), volume['id'])
|
|
screplid = None
|
|
nvolid = None
|
|
status = 'error'
|
|
|
|
# Save some information for the next step.
|
|
# nvol is the new volume created by replicate_to_common.
|
|
# We also grab our extra specs here.
|
|
replitem = {
|
|
'volume': volume,
|
|
'specs': self._parse_extraspecs(volume),
|
|
'qosnode': qosnode,
|
|
'screpl': screplid,
|
|
'cvol': cvol['instanceId'],
|
|
'ovol': ovol['instanceId'],
|
|
'nvol': nvolid,
|
|
'rdd': six.text_type(api.ssn),
|
|
'status': status}
|
|
|
|
return replitem
|
|
|
|
def _failback_live_volume(self, api, id, provider_id):
|
|
"""failback the live volume to its original
|
|
|
|
:param api: Dell SC API
|
|
:param id: Volume ID
|
|
:param provider_id: Dell Instance ID
|
|
:return: model_update dict
|
|
"""
|
|
model_update = {}
|
|
# We do not search by name. Only failback if we have a complete
|
|
# LV object.
|
|
sclivevolume = api.get_live_volume(provider_id)
|
|
# TODO(tswanson): Check swapped state first.
|
|
if sclivevolume and api.swap_roles_live_volume(sclivevolume):
|
|
LOG.info(_LI('Success swapping sclivevolume roles %s'), id)
|
|
model_update = {
|
|
'status': 'available',
|
|
'replication_status': fields.ReplicationStatus.ENABLED,
|
|
'provider_id':
|
|
sclivevolume['secondaryVolume']['instanceId']}
|
|
else:
|
|
LOG.info(_LI('Failure swapping roles %s'), id)
|
|
model_update = {'status': 'error'}
|
|
|
|
return model_update
|
|
|
|
def _finish_failback(self, api, replitems):
|
|
# Wait for replication to complete.
|
|
# This will also flip replication.
|
|
self._wait_for_replication(api, replitems)
|
|
# Replications are done. Attach to any additional replication
|
|
# backends.
|
|
self._reattach_remaining_replications(api, replitems)
|
|
self._fixup_types(api, replitems)
|
|
return self._volume_updates(replitems)
|
|
|
|
def failback_volumes(self, volumes):
|
|
"""This is a generic volume failback.
|
|
|
|
:param volumes: List of volumes that need to be failed back.
|
|
:return: volume_updates for the list of volumes.
|
|
"""
|
|
LOG.info(_LI('failback_volumes'))
|
|
with self._client.open_connection() as api:
|
|
# Get our qosnode. This is a good way to make sure the backend
|
|
# is still setup so that we can do this.
|
|
qosnode = self._get_qos(api.ssn)
|
|
if not qosnode:
|
|
raise exception.VolumeBackendAPIException(
|
|
message=_('Unable to failback. Backend is misconfigured.'))
|
|
|
|
volume_updates = []
|
|
replitems = []
|
|
|
|
# Trundle through the volumes. Update non replicated to alive again
|
|
# and reverse the replications for the remaining volumes.
|
|
for volume in volumes:
|
|
LOG.info(_LI('failback_volumes: starting volume: %s'), volume)
|
|
model_update = {}
|
|
if volume.get('replication_driver_data'):
|
|
rspecs = self._get_replication_specs(volume)
|
|
if rspecs['live']:
|
|
model_update = self._failback_live_volume(
|
|
api, volume['id'], volume['provider_id'])
|
|
else:
|
|
replitem = self._failback_replication(api, volume,
|
|
qosnode)
|
|
|
|
# Save some information for the next step.
|
|
# nvol is the new volume created by
|
|
# replicate_to_common. We also grab our
|
|
# extra specs here.
|
|
replitems.append(replitem)
|
|
else:
|
|
# Not replicated. Just set it to available.
|
|
model_update = {'status': 'available'}
|
|
|
|
# Save our update
|
|
if model_update:
|
|
volume_updates.append({'volume_id': volume['id'],
|
|
'updates': model_update})
|
|
# Let's do up to 5 replications at once.
|
|
if len(replitems) == 5:
|
|
volume_updates += self._finish_failback(api, replitems)
|
|
replitems = []
|
|
|
|
# Finish any leftover items
|
|
if replitems:
|
|
volume_updates += self._finish_failback(api, replitems)
|
|
|
|
# Set us back to a happy state.
|
|
# The only way this doesn't happen is if the primary is down.
|
|
self._update_backend(None)
|
|
return volume_updates
|
|
|
|
def _failover_replication(self, api, id, provider_id, destssn):
|
|
rvol = api.break_replication(id, provider_id, destssn)
|
|
model_update = {}
|
|
if rvol:
|
|
LOG.info(_LI('Success failing over volume %s'), id)
|
|
model_update = {'replication_status':
|
|
fields.ReplicationStatus.FAILED_OVER,
|
|
'provider_id': rvol['instanceId']}
|
|
else:
|
|
LOG.info(_LI('Failed failing over volume %s'), id)
|
|
model_update = {'status': 'error'}
|
|
|
|
return model_update
|
|
|
|
def _failover_live_volume(self, api, id, provider_id):
|
|
model_update = {}
|
|
# Search for volume by id if we have to.
|
|
sclivevolume = api.get_live_volume(provider_id, id)
|
|
if sclivevolume:
|
|
swapped = api.is_swapped(provider_id, sclivevolume)
|
|
# If we aren't swapped try it. If fail error out.
|
|
if not swapped and not api.swap_roles_live_volume(sclivevolume):
|
|
LOG.info(_LI('Failure swapping roles %s'), id)
|
|
model_update = {'status': 'error'}
|
|
return model_update
|
|
|
|
LOG.info(_LI('Success swapping sclivevolume roles %s'), id)
|
|
sclivevolume = api.get_live_volume(provider_id)
|
|
model_update = {
|
|
'replication_status':
|
|
fields.ReplicationStatus.FAILED_OVER,
|
|
'provider_id':
|
|
sclivevolume['primaryVolume']['instanceId']}
|
|
|
|
# Error and leave.
|
|
return model_update
|
|
|
|
def failover_host(self, context, volumes, secondary_id=None):
|
|
"""Failover to secondary.
|
|
|
|
:param context: security context
|
|
:param secondary_id: Specifies rep target to fail over to
|
|
:param volumes: List of volumes serviced by this backend.
|
|
:returns: destssn, volume_updates data structure
|
|
|
|
Example volume_updates data structure:
|
|
|
|
.. code-block:: json
|
|
|
|
[{'volume_id': <cinder-uuid>,
|
|
'updates': {'provider_id': 8,
|
|
'replication_status': 'failed-over',
|
|
'replication_extended_status': 'whatever',...}},]
|
|
"""
|
|
LOG.debug('failover-host')
|
|
LOG.debug(self.failed_over)
|
|
LOG.debug(self.active_backend_id)
|
|
LOG.debug(self.replication_enabled)
|
|
if self.failed_over:
|
|
if secondary_id == 'default':
|
|
LOG.debug('failing back')
|
|
return 'default', self.failback_volumes(volumes)
|
|
raise exception.VolumeBackendAPIException(
|
|
message='Already failed over.')
|
|
|
|
LOG.info(_LI('Failing backend to %s'), secondary_id)
|
|
# basic check
|
|
if self.replication_enabled:
|
|
with self._client.open_connection() as api:
|
|
# Look for the specified secondary.
|
|
destssn = self._parse_secondary(api, secondary_id)
|
|
if destssn:
|
|
# We roll through trying to break replications.
|
|
# Is failing here a complete failure of failover?
|
|
volume_updates = []
|
|
for volume in volumes:
|
|
model_update = {}
|
|
if volume.get('replication_driver_data'):
|
|
rspecs = self._get_replication_specs(volume)
|
|
if rspecs['live']:
|
|
model_update = self._failover_live_volume(
|
|
api, volume['id'],
|
|
volume.get('provider_id'))
|
|
else:
|
|
model_update = self._failover_replication(
|
|
api, volume['id'],
|
|
volume.get('provider_id'), destssn)
|
|
else:
|
|
# Not a replicated volume. Try to unmap it.
|
|
scvolume = api.find_volume(
|
|
volume['id'], volume.get('provider_id'))
|
|
api.remove_mappings(scvolume)
|
|
model_update = {'status': 'error'}
|
|
# Either we are failed over or our status is now error.
|
|
volume_updates.append({'volume_id': volume['id'],
|
|
'updates': model_update})
|
|
|
|
# this is it.
|
|
self._update_backend(destssn)
|
|
LOG.debug('after update backend')
|
|
LOG.debug(self.failed_over)
|
|
LOG.debug(self.active_backend_id)
|
|
LOG.debug(self.replication_enabled)
|
|
return destssn, volume_updates
|
|
else:
|
|
raise exception.InvalidInput(message=(
|
|
_('replication_failover failed. %s not found.') %
|
|
secondary_id))
|
|
# I don't think we should ever get here.
|
|
raise exception.VolumeBackendAPIException(message=(
|
|
_('replication_failover failed. '
|
|
'Backend not configured for failover')))
|
|
|
|
def _get_unmanaged_replay(self, api, volume_name, provider_id,
|
|
existing_ref):
|
|
replay_name = None
|
|
if existing_ref:
|
|
replay_name = existing_ref.get('source-name')
|
|
if not replay_name:
|
|
msg = _('_get_unmanaged_replay: Must specify source-name.')
|
|
LOG.error(msg)
|
|
raise exception.ManageExistingInvalidReference(
|
|
existing_ref=existing_ref, reason=msg)
|
|
# Find our volume.
|
|
scvolume = api.find_volume(volume_name, provider_id)
|
|
if not scvolume:
|
|
# Didn't find it.
|
|
msg = (_('_get_unmanaged_replay: Cannot find volume id %s')
|
|
% volume_name)
|
|
LOG.error(msg)
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
# Find our replay.
|
|
screplay = api.find_replay(scvolume, replay_name)
|
|
if not screplay:
|
|
# Didn't find it. Reference must be invalid.
|
|
msg = (_('_get_unmanaged_replay: Cannot '
|
|
'find snapshot named %s') % replay_name)
|
|
LOG.error(msg)
|
|
raise exception.ManageExistingInvalidReference(
|
|
existing_ref=existing_ref, reason=msg)
|
|
return screplay
|
|
|
|
def manage_existing_snapshot(self, snapshot, existing_ref):
|
|
"""Brings an existing backend storage object under Cinder management.
|
|
|
|
existing_ref is passed straight through from the API request's
|
|
manage_existing_ref value, and it is up to the driver how this should
|
|
be interpreted. It should be sufficient to identify a storage object
|
|
that the driver should somehow associate with the newly-created cinder
|
|
snapshot structure.
|
|
|
|
There are two ways to do this:
|
|
|
|
1. Rename the backend storage object so that it matches the
|
|
snapshot['name'] which is how drivers traditionally map between a
|
|
cinder snapshot and the associated backend storage object.
|
|
|
|
2. Place some metadata on the snapshot, or somewhere in the backend,
|
|
that allows other driver requests (e.g. delete) to locate the
|
|
backend storage object when required.
|
|
|
|
If the existing_ref doesn't make sense, or doesn't refer to an existing
|
|
backend storage object, raise a ManageExistingInvalidReference
|
|
exception.
|
|
"""
|
|
with self._client.open_connection() as api:
|
|
# Find our unmanaged snapshot. This will raise on error.
|
|
volume_name = snapshot.get('volume_id')
|
|
provider_id = snapshot.get('provider_id')
|
|
snapshot_id = snapshot.get('id')
|
|
screplay = self._get_unmanaged_replay(api, volume_name,
|
|
provider_id, existing_ref)
|
|
# Manage means update description and update expiration.
|
|
if not api.manage_replay(screplay, snapshot_id):
|
|
# That didn't work. Error.
|
|
msg = (_('manage_existing_snapshot: Error managing '
|
|
'existing replay %(ss)s on volume %(vol)s') %
|
|
{'ss': screplay.get('description'),
|
|
'vol': volume_name})
|
|
LOG.error(msg)
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
|
|
# Life is good. Let the world know what we've done.
|
|
LOG.info(_LI('manage_existing_snapshot: snapshot %(exist)s on '
|
|
'volume %(volume)s has been renamed to %(id)s and is '
|
|
'now managed by Cinder.'),
|
|
{'exist': screplay.get('description'),
|
|
'volume': volume_name,
|
|
'id': snapshot_id})
|
|
return {'provider_id': screplay['createVolume']['instanceId']}
|
|
|
|
# NOTE: Can't use abstractmethod before all drivers implement it
|
|
def manage_existing_snapshot_get_size(self, snapshot, existing_ref):
|
|
"""Return size of snapshot to be managed by manage_existing.
|
|
|
|
When calculating the size, round up to the next GB.
|
|
"""
|
|
volume_name = snapshot.get('volume_id')
|
|
provider_id = snapshot.get('provider_id')
|
|
with self._client.open_connection() as api:
|
|
screplay = self._get_unmanaged_replay(api, volume_name,
|
|
provider_id, existing_ref)
|
|
sz, rem = dell_storagecenter_api.StorageCenterApi.size_to_gb(
|
|
screplay['size'])
|
|
if rem > 0:
|
|
raise exception.VolumeBackendAPIException(
|
|
data=_('Volume size must be a multiple of 1 GB.'))
|
|
return sz
|
|
|
|
# NOTE: Can't use abstractmethod before all drivers implement it
|
|
def unmanage_snapshot(self, snapshot):
|
|
"""Removes the specified snapshot from Cinder management.
|
|
|
|
Does not delete the underlying backend storage object.
|
|
|
|
NOTE: We do set the expire countdown to 1 day. Once a snapshot is
|
|
unmanaged it will expire 24 hours later.
|
|
"""
|
|
with self._client.open_connection() as api:
|
|
snapshot_id = snapshot.get('id')
|
|
# provider_id is the snapshot's parent volume's instanceId.
|
|
provider_id = snapshot.get('provider_id')
|
|
volume_name = snapshot.get('volume_id')
|
|
# Find our volume.
|
|
scvolume = api.find_volume(volume_name, provider_id)
|
|
if not scvolume:
|
|
# Didn't find it.
|
|
msg = (_('unmanage_snapshot: Cannot find volume id %s')
|
|
% volume_name)
|
|
LOG.error(msg)
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
# Find our replay.
|
|
screplay = api.find_replay(scvolume, snapshot_id)
|
|
if not screplay:
|
|
# Didn't find it. Reference must be invalid.
|
|
msg = (_('unmanage_snapshot: Cannot find snapshot named %s')
|
|
% snapshot_id)
|
|
LOG.error(msg)
|
|
raise exception.VolumeBackendAPIException(data=msg)
|
|
# Free our snapshot.
|
|
api.unmanage_replay(screplay)
|
|
# Do not check our result.
|
|
|
|
def thaw_backend(self, context):
|
|
"""Notify the backend that it's unfrozen/thawed.
|
|
|
|
This is a gate. We do not allow the backend to be thawed if
|
|
it is still failed over.
|
|
|
|
:param context: security context
|
|
:response: True on success
|
|
:raises Invalid: if it cannot be thawed.
|
|
"""
|
|
# We shouldn't be called if we are not failed over.
|
|
if self.failed_over:
|
|
msg = _('The Dell SC array does not support thawing a failed over'
|
|
' replication. Please migrate volumes to an operational '
|
|
'back-end or resolve primary system issues and '
|
|
'fail back to reenable full functionality.')
|
|
LOG.error(msg)
|
|
raise exception.Invalid(reason=msg)
|
|
|
|
return True
|