openstack/cinder
Code Issues Proposed changes

Merge "3PAR: Implement v2 replication (managed)"

Browse Source
This commit is contained in:
Jenkins 2015-12-09 02:25:13 +00:00 committed by Gerrit Code Review
commit e2feef3274
4 changed files with 1662 additions and 47 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

882
cinder/tests/unit/test_hpe3par.py
View File

File diff suppressed because it is too large Load Diff

697
cinder/volume/drivers/hpe/hpe_3par_common.py
View File

@ -71,6 +71,7 @@ from taskflow.patterns import linear_flow
LOG = logging.getLogger(__name__)
MIN_CLIENT_VERSION = '4.0.0'
MIN_REP_CLIENT_VERSION = '4.0.2'
DEDUP_API_VERSION = 30201120
FLASH_CACHE_API_VERSION = 30201200
SRSTATLD_API_VERSION = 30201200
@ -214,10 +215,11 @@ class HPE3PARCommon(object):
3.0.1 - Fixed find_existing_vluns bug #1515033
3.0.2 - Python 3 support
3.0.3 - Remove db access for consistency groups
3.0.4 - Adds v2 managed replication support
"""
VERSION = "3.0.3"
VERSION = "3.0.4"
stats = {}
@ -234,6 +236,12 @@ class HPE3PARCommon(object):
CONVERT_TO_FULL = 2
CONVERT_TO_DEDUP = 3
# v2 replication constants
SYNC = 1
PERIODIC = 2
EXTRA_SPEC_REP_MODE = "replication:mode"
EXTRA_SPEC_REP_SYNC_PERIOD = "replication:sync_period"
# Valid values for volume type extra specs
# The first value in the list is the default value
valid_prov_values = ['thin', 'full', 'dedup']
@ -258,6 +266,8 @@ class HPE3PARCommon(object):
self.config = config
self.client = None
self.uuid = uuid.uuid4()
self._replication_targets = []
self._replication_enabled = False
def get_version(self):
return self.VERSION
@ -269,8 +279,14 @@ class HPE3PARCommon(object):
LOG.error(msg)
raise exception.InvalidInput(reason=msg)
def _create_client(self):
cl = client.HPE3ParClient(self.config.hpe3par_api_url)
def _create_client(self, timeout=None):
# Timeout is only supported in version 4.0.2 and greater of the
# python-3parclient.
if hpe3parclient.version >= MIN_REP_CLIENT_VERSION:
cl = client.HPE3ParClient(self.config.hpe3par_api_url,
timeout=timeout)
else:
cl = client.HPE3ParClient(self.config.hpe3par_api_url)
client_version = hpe3parclient.version
if client_version < MIN_CLIENT_VERSION:
@ -314,17 +330,53 @@ class HPE3PARCommon(object):
LOG.debug("Disconnect from 3PAR REST and SSH %s", self.uuid)
self.client.logout()
def do_setup(self, context):
def _create_replication_client(self, remote_array):
try:
cl = client.HPE3ParClient(remote_array['hpe3par_api_url'])
cl.login(remote_array['hpe3par_username'],
remote_array['hpe3par_password'])
except hpeexceptions.HTTPUnauthorized as ex:
msg = (_("Failed to Login to 3PAR (%(url)s) because %(err)s") %
{'url': remote_array['hpe3par_api_url'], 'err': ex})
LOG.error(msg)
raise exception.InvalidInput(reason=msg)
known_hosts_file = CONF.ssh_hosts_key_file
policy = "AutoAddPolicy"
if CONF.strict_ssh_host_key_policy:
policy = "RejectPolicy"
cl.setSSHOptions(
remote_array['san_ip'],
remote_array['san_login'],
remote_array['san_password'],
port=remote_array['san_ssh_port'],
conn_timeout=remote_array['ssh_conn_timeout'],
privatekey=remote_array['san_private_key'],
missing_key_policy=policy,
known_hosts_file=known_hosts_file)
return cl
def _destroy_replication_client(self, client):
client.logout()
def do_setup(self, context, timeout=None):
if hpe3parclient is None:
msg = _('You must install hpe3parclient before using 3PAR'
' drivers. Run "pip install python-3parclient" to'
' install the hpe3parclient.')
raise exception.VolumeBackendAPIException(data=msg)
try:
self.client = self._create_client()
self.client = self._create_client(timeout=timeout)
wsapi_version = self.client.getWsApiVersion()
self.API_VERSION = wsapi_version['build']
except hpeexceptions.UnsupportedVersion as ex:
# In the event we cannot contact the configured primary array,
# we want to allow a failover if replication is enabled.
if hpe3parclient.version >= MIN_REP_CLIENT_VERSION:
self._do_replication_setup()
if self._replication_enabled:
self.client = None
raise exception.InvalidInput(ex)
if context:
@ -354,16 +406,32 @@ class HPE3PARCommon(object):
LOG.error(msg)
raise exception.InvalidInput(message=msg)
def check_for_setup_error(self):
self.client_login()
# get the client ID for provider_location
try:
cpg_names = self.config.hpe3par_cpg
for cpg_name in cpg_names:
self.validate_cpg(cpg_name)
self.client_login()
info = self.client.getStorageSystemInfo()
self.client.id = six.text_type(info['id'])
except Exception:
self.client.id = 0
finally:
self.client_logout()
# v2 replication setup
if not self._replication_enabled and (
hpe3parclient.version >= MIN_REP_CLIENT_VERSION):
self._do_replication_setup()
def check_for_setup_error(self):
if self.client:
self.client_login()
try:
cpg_names = self.config.hpe3par_cpg
for cpg_name in cpg_names:
self.validate_cpg(cpg_name)
finally:
self.client_logout()
def validate_cpg(self, cpg_name):
try:
self.client.getCPG(cpg_name)
@ -742,12 +810,25 @@ class HPE3PARCommon(object):
def _extend_volume(self, volume, volume_name, growth_size_mib,
_convert_to_base=False):
model_update = None
rcg_name = self._get_3par_rcg_name(volume['id'])
is_volume_replicated = self._volume_of_replicated_type(volume)
try:
if _convert_to_base:
LOG.debug("Converting to base volume prior to growing.")
model_update = self._convert_to_base_volume(volume)
# If the volume is replicated and we are not failed over,
# remote copy has to be stopped before the volume can be extended.
failed_over = volume.get("replication_status", None)
is_failed_over = failed_over == "failed-over"
if is_volume_replicated and not is_failed_over:
self.client.stopRemoteCopy(rcg_name)
self.client.growVolume(volume_name, growth_size_mib)
if is_volume_replicated and not is_failed_over:
self.client.startRemoteCopy(rcg_name)
except Exception as ex:
# If the extend fails, we must restart remote copy.
if is_volume_replicated:
self.client.startRemoteCopy(rcg_name)
with excutils.save_and_reraise_exception() as ex_ctxt:
if (not _convert_to_base and
isinstance(ex, hpeexceptions.HTTPForbidden) and
@ -799,6 +880,16 @@ class HPE3PARCommon(object):
unm_name = self._encode_name(volume_id)
return "unm-%s" % unm_name
# v2 replication conversion
def _get_3par_rcg_name(self, volume_id):
rcg_name = self._encode_name(volume_id)
rcg = "rcg-%s" % rcg_name
return rcg[:22]
def _get_3par_remote_rcg_name(self, volume_id, provider_location):
return self._get_3par_rcg_name(volume_id) + ".r" + (
six.text_type(provider_location))
def _encode_name(self, name):
uuid_str = name.replace("-", "")
vol_uuid = uuid.UUID('urn:uuid:%s' % uuid_str)
@ -1012,6 +1103,11 @@ class HPE3PARCommon(object):
'consistencygroup_support': True,
}
if hpe3parclient.version >= MIN_REP_CLIENT_VERSION:
pool['replication_enabled'] = self._replication_enabled
pool['replication_type'] = ['sync', 'periodic']
pool['replication_count'] = len(self._replication_targets)
pools.append(pool)
self.stats = {'driver_version': '3.0',
@ -1521,6 +1617,13 @@ class HPE3PARCommon(object):
self.client.deleteVolume(volume_name)
LOG.error(_LE("Exception: %s"), ex)
raise exception.CinderException(ex)
# v2 replication check
replication_flag = False
if self._volume_of_replicated_type(volume) and (
self._do_volume_replication_setup(volume)):
replication_flag = True
except hpeexceptions.HTTPConflict:
msg = _("Volume (%s) already exists on array") % volume_name
LOG.error(msg)
@ -1538,7 +1641,9 @@ class HPE3PARCommon(object):
LOG.error(_LE("Exception: %s"), ex)
raise exception.CinderException(ex)
return self._get_model_update(volume['host'], cpg)
return self._get_model_update(volume['host'], cpg,
replication=replication_flag,
provider_location=self.client.id)
def _copy_volume(self, src_name, dest_name, cpg, snap_cpg=None,
tpvv=True, tdvv=False):
@ -1571,7 +1676,8 @@ class HPE3PARCommon(object):
return comment_dict[key]
return None
def _get_model_update(self, volume_host, cpg):
def _get_model_update(self, volume_host, cpg, replication=False,
provider_location=None):
"""Get model_update dict to use when we select a pool.
The pools implementation uses a volume['host'] suffix of :poolname.
@ -1589,12 +1695,18 @@ class HPE3PARCommon(object):
:param cpg: The actual pool (cpg) used, for example from the type.
:return: dict Model update if we need to update volume host, else None
"""
model_update = None
model_update = {}
host = volume_utils.extract_host(volume_host, 'backend')
host_and_pool = volume_utils.append_host(host, cpg)
if volume_host != host_and_pool:
# Since we selected a pool based on type, update the model.
model_update = {'host': host_and_pool}
model_update['host'] = host_and_pool
if replication:
model_update['replication_status'] = 'enabled'
if replication and provider_location:
model_update['provider_location'] = provider_location
if not model_update:
model_update = None
return model_update
def create_cloned_volume(self, volume, src_vref):
@ -1612,7 +1724,15 @@ class HPE3PARCommon(object):
tpvv=type_info['tpvv'],
tdvv=type_info['tdvv'])
return self._get_model_update(volume['host'], cpg)
# v2 replication check
replication_flag = False
if self._volume_of_replicated_type(volume) and (
self._do_volume_replication_setup(volume)):
replication_flag = True
return self._get_model_update(volume['host'], cpg,
replication=replication_flag,
provider_location=self.client.id)
except hpeexceptions.HTTPForbidden:
raise exception.NotAuthorized()
@ -1623,6 +1743,17 @@ class HPE3PARCommon(object):
raise exception.CinderException(ex)
def delete_volume(self, volume):
# v2 replication check
# If the volume type is replication enabled, we want to call our own
# method of deconstructing the volume and its dependencies
if self._volume_of_replicated_type(volume):
replication_status = volume.get('replication_status', None)
if replication_status and replication_status == "failed-over":
self._delete_replicated_failed_over_volume(volume)
else:
self._do_volume_replication_destroy(volume)
return
try:
volume_name = self._get_3par_vol_name(volume['id'])
# Try and delete the volume, it might fail here because
@ -1702,7 +1833,7 @@ class HPE3PARCommon(object):
{'vol_name': pprint.pformat(volume['display_name']),
'ss_name': pprint.pformat(snapshot['display_name'])})
model_update = None
model_update = {}
if volume['size'] < snapshot['volume_size']:
err = ("You cannot reduce size of the volume. It must "
"be greater than or equal to the snapshot.")
@ -1771,6 +1902,13 @@ class HPE3PARCommon(object):
self.client.deleteVolume(volume_name)
LOG.error(_LE("Exception: %s"), ex)
raise exception.CinderException(ex)
# v2 replication check
if self._volume_of_replicated_type(volume) and (
self._do_volume_replication_setup(volume)):
model_update['replication_status'] = 'enabled'
model_update['provider_location'] = self.client.id
except hpeexceptions.HTTPForbidden as ex:
LOG.error(_LE("Exception: %s"), ex)
raise exception.NotAuthorized()
@ -2444,6 +2582,531 @@ class HPE3PARCommon(object):
pass
return existing_vluns
# v2 replication methods
def get_replication_updates(self, context):
# TODO(aorourke): the manager does not do anything with these updates.
# When that is chanaged, I will modify this as well.
errors = []
return errors
def replication_enable(self, context, volume):
"""Enable replication on a replication capable volume."""
if not self._volume_of_replicated_type(volume):
msg = _("Unable to enable volume replication because volume is "
"not of replicated type.")
LOG.error(msg)
raise exception.VolumeBackendAPIException(data=msg)
model_update = {"provider_location": self.client.id}
# If replication is not enabled and the volume is of replicated type,
# we treat this as an error.
if not self._replication_enabled:
msg = _LE("Enabling replication failed because replication is "
"not properly configured.")
LOG.error(msg)
model_update['replication_status'] = "error"
else:
if self._do_volume_replication_setup(volume):
model_update['replication_status'] = "enabled"
else:
model_update['replication_status'] = "error"
return model_update
def replication_disable(self, context, volume):
"""Disable replication on the specified volume."""
if not self._volume_of_replicated_type(volume):
msg = _("Unable to disable volume replication because volume is "
"not of replicated type.")
LOG.error(msg)
raise exception.VolumeBackendAPIException(data=msg)
model_update = {}
# If replication is not enabled and the volume is of replicated type,
# we treat this as an error.
if self._replication_enabled:
model_update['replication_status'] = 'disabled'
rcg_name = self._get_3par_rcg_name(volume['id'])
vol_name = self._get_3par_vol_name(volume['id'])
try:
self.client.stopRemoteCopy(rcg_name)
except Exception as ex:
msg = (_LE("There was a problem disabling replication on "
"volume '%(name)s': %(error)s") %
{'name': vol_name,
'error': six.text_type(ex)})
LOG.error(msg)
model_update['replication_status'] = 'disable_failed'
else:
msg = _LE("Disabling replication failed because replication is "
"not properly configured.")
LOG.error(msg)
model_update['replication_status'] = 'error'
return model_update
def replication_failover(self, context, volume, secondary):
"""Force failover to a secondary replication target."""
if not self._volume_of_replicated_type(volume):
msg = _("Unable to failover because volume is not of "
"replicated type.")
LOG.error(msg)
raise exception.VolumeBackendAPIException(data=msg)
# If replication is not enabled and the volume is of replicated type,
# we treat this as an error.
if not self._replication_enabled:
msg = _LE("Issuing a fail-over failed because replication is "
"not properly configured.")
LOG.error(msg)
model_update = {"replication_status": "error"}
return model_update
failover_target = None
for target in self._replication_targets:
if target['target_device_id'] == secondary:
failover_target = target
break
if not failover_target:
msg = _("A valid secondary target MUST be specified in order "
"to failover.")
LOG.error(msg)
raise exception.VolumeBackendAPIException(data=msg)
if self.client is not None and failover_target['id'] == self.client.id:
msg = _("The failover array cannot be the primary array.")
LOG.error(msg)
raise exception.VolumeBackendAPIException(data=msg)
try:
# Try and stop remote-copy on main array.
rcg_name = self._get_3par_rcg_name(volume['id'])
self.client.stopRemoteCopy(rcg_name)
except Exception:
pass
try:
# Failover to secondary array.
remote_rcg_name = self._get_3par_remote_rcg_name(
volume['id'], volume['provider_location'])
cl = self._create_replication_client(failover_target)
cl.recoverRemoteCopyGroupFromDisaster(
remote_rcg_name, self.client.RC_ACTION_CHANGE_TO_PRIMARY)
new_location = volume['provider_location'] + ":" + (
failover_target['id'])
model_update = {"provider_location": new_location,
"replication_status": "inactive"}
if failover_target['managed_backend_name']:
# We want to update the volumes host if our target is managed.
model_update['host'] = failover_target['managed_backend_name']
except Exception as ex:
msg = _("There was a problem with the failover (%s) and it was "
"unsuccessful.") % six.text_type(ex)
LOG.error(msg)
raise exception.VolumeBackendAPIException(data=msg)
finally:
self._destroy_replication_client(cl)
return model_update
def list_replication_targets(self, context, volume):
"""Provides a means to obtain replication targets for a volume.
This will query all enabled targets on a 3PAR backend and cross
reference them with all entries in cinder.conf. It will return
only those that appear on both, aka enabled replication targets.
"""
if not self._volume_of_replicated_type(volume):
return []
allowed_names = []
# If the primary target is offline we can not ask it what targets are
# available. Our only option is to list all cinder.conf entries.
try:
rcg_name = self._get_3par_rcg_name(volume['id'])
rcg = self.client.getRemoteCopyGroup(rcg_name)
rcg_targets = rcg['targets']
for target in rcg_targets:
allowed_names.append(target['targetName'])
except Exception:
LOG.warning(_LW("The primary array is currently unreachable. All "
"targets returned from list_replication_targets "
"are pulled directly from cinder.conf and are not "
"guarenteed to be available because they could "
"not be verified with the primary array."))
replication_targets = []
for target in self._replication_targets:
if not allowed_names or (
target['target_device_id'] in allowed_names):
list_vals = {'target_device_id': target['target_device_id']}
replication_targets.append(list_vals)
return {'volume_id': volume['id'],
'targets': replication_targets}
def _do_replication_setup(self):
replication_devices = self.config.replication_device
if replication_devices:
for dev in replication_devices:
remote_array = {}
is_managed = dev.get('managed_backend_name')
if not is_managed:
msg = _("Unmanaged replication is not supported at this "
"time. Please configure cinder.conf for managed "
"replication.")
LOG.error(msg)
raise exception.VolumeBackendAPIException(data=msg)
remote_array['managed_backend_name'] = is_managed
remote_array['replication_mode'] = (
self._get_remote_copy_mode_num(
dev.get('replication_mode')))
remote_array['target_device_id'] = (
dev.get('target_device_id'))
remote_array['cpg_map'] = (
dev.get('cpg_map'))
remote_array['hpe3par_api_url'] = (
dev.get('hpe3par_api_url'))
remote_array['hpe3par_username'] = (
dev.get('hpe3par_username'))
remote_array['hpe3par_password'] = (
dev.get('hpe3par_password'))
remote_array['san_ip'] = (
dev.get('san_ip'))
remote_array['san_login'] = (
dev.get('san_login'))
remote_array['san_password'] = (
dev.get('san_password'))
remote_array['san_ssh_port'] = (
dev.get('san_ssh_port', self.config.san_ssh_port))
remote_array['ssh_conn_timeout'] = (
dev.get('ssh_conn_timeout', self.config.ssh_conn_timeout))
remote_array['san_private_key'] = (
dev.get('san_private_key', self.config.san_private_key))
array_name = remote_array['target_device_id']
# Make sure we can log into the client, that it has been
# correctly configured, and it its version matches the
# primary arrarys version.
try:
cl = self._create_replication_client(remote_array)
array_id = six.text_type(cl.getStorageSystemInfo()['id'])
remote_array['id'] = array_id
wsapi_version = cl.getWsApiVersion()['build']
if self.client is not None and (
wsapi_version != self.API_VERSION):
msg = (_LW("The target array and all of its secondary "
"arrays must be on the same API version. "
"Array '%(target)s' is on %(target_ver)s "
"while the primary array is on "
"%(primary_ver)s, therefore it will not "
"be added as a valid replication target.") %
{'target': array_name,
'target_ver': wsapi_version,
'primary_ver': self.API_VERSION})
LOG.warning(msg)
elif not self._is_valid_replication_array(remote_array):
msg = (_LW("'%s' is not a valid replication array. "
"In order to be valid, target_device_id, "
"replication_mode, "
"hpe3par_api_url, hpe3par_username, "
"hpe3par_password, cpg_map, and "
"must be specified. If the target is "
"managed, managed_backend_name must be set "
"as well.") % array_name)
LOG.warning(msg)
else:
self._replication_targets.append(remote_array)
except Exception:
msg = (_LE("Could not log in to 3PAR array (%s) with the "
"provided credentials.") % array_name)
LOG.error(msg)
finally:
self._destroy_replication_client(cl)
if self._is_replication_configured_correct():
self._replication_enabled = True
def _is_valid_replication_array(self, target):
for k, v in target.items():
if v is None:
return False
return True
def _is_replication_configured_correct(self):
rep_flag = True
# Make sure there is at least one replication target.
if len(self._replication_targets) < 1:
LOG.error(_LE("There must be at least one valid replication "
"device configured."))
rep_flag = False
return rep_flag
def _is_replication_mode_correct(self, mode, sync_num):
rep_flag = True
# Make sure replication_mode is set to either sync|periodic.
mode = self._get_remote_copy_mode_num(mode)
if not mode:
LOG.error(_LE("Extra spec replication:mode must be set and must "
"be either 'sync' or 'periodic'."))
rep_flag = False
else:
# If replication:mode is periodic, replication_sync_period must be
# set between 300 - 31622400 seconds.
if mode == self.PERIODIC and (
sync_num < 300 or sync_num > 31622400):
LOG.error(_LE("Extra spec replication:sync_period must be "
"greater than 299 and less than 31622401 "
"seconds."))
rep_flag = False
return rep_flag
def _volume_of_replicated_type(self, volume):
replicated_type = False
volume_type_id = volume.get('volume_type_id')
if volume_type_id:
volume_type = self._get_volume_type(volume_type_id)
extra_specs = volume_type.get('extra_specs')
if extra_specs and 'replication_enabled' in extra_specs:
rep_val = extra_specs['replication_enabled']
replicated_type = (rep_val == "<is> True")
return replicated_type
def _is_volume_in_remote_copy_group(self, volume):
rcg_name = self._get_3par_rcg_name(volume['id'])
try:
self.client.getRemoteCopyGroup(rcg_name)
return True
except hpeexceptions.HTTPNotFound:
return False
def _get_remote_copy_mode_num(self, mode):
ret_mode = None
if mode == "sync":
ret_mode = self.SYNC
if mode == "periodic":
ret_mode = self.PERIODIC
return ret_mode
def _get_cpg_from_cpg_map(self, cpg_map, target_cpg):
ret_target_cpg = None
cpg_pairs = cpg_map.split(' ')
for cpg_pair in cpg_pairs:
cpgs = cpg_pair.split(':')
cpg = cpgs[0]
dest_cpg = cpgs[1]
if cpg == target_cpg:
ret_target_cpg = dest_cpg
return ret_target_cpg
def _do_volume_replication_setup(self, volume):
"""This function will do or ensure the following:
-Create volume on main array (already done in create_volume)
-Create Remote Copy Group on main array
-Add volume to Remote Copy Group on main array
-Start remote copy
If anything here fails, we will need to clean everything up in
reverse order, including the original volume.
"""
rcg_name = self._get_3par_rcg_name(volume['id'])
# If the volume is already in a remote copy group, return True
# after starting remote copy. If remote copy is already started,
# issuing this command again will be fine.
if self._is_volume_in_remote_copy_group(volume):
try:
self.client.startRemoteCopy(rcg_name)
except Exception:
pass
return True
try:
# Grab the extra_spec entries for replication and make sure they
# are set correctly.
volume_type = self._get_volume_type(volume["volume_type_id"])
extra_specs = volume_type.get("extra_specs")
replication_mode = extra_specs.get(self.EXTRA_SPEC_REP_MODE)
replication_mode_num = self._get_remote_copy_mode_num(
replication_mode)
replication_sync_period = extra_specs.get(
self.EXTRA_SPEC_REP_SYNC_PERIOD)
if replication_sync_period:
replication_sync_period = int(replication_sync_period)
if not self._is_replication_mode_correct(replication_mode,
replication_sync_period):
msg = _("The replication mode was not configured correctly "
"in the volume type extra_specs. If replication:mode "
"is periodic, replication:sync_period must also be "
"specified and be between 300 and 31622400 seconds.")
LOG.error(msg)
raise exception.VolumeBackendAPIException(data=msg)
vol_settings = self.get_volume_settings_from_type(volume)
local_cpg = vol_settings['cpg']
vol_name = self._get_3par_vol_name(volume['id'])
# Create remote copy group on main array.
rcg_targets = []
sync_targets = []
for target in self._replication_targets:
# Only add targets that match the volumes replication mode.
if target['replication_mode'] == replication_mode_num:
cpg = self._get_cpg_from_cpg_map(target['cpg_map'],
local_cpg)
rcg_target = {'targetName': target['target_device_id'],
'mode': replication_mode_num,
'snapCPG': cpg,
'userCPG': cpg}
rcg_targets.append(rcg_target)
sync_target = {'targetName': target['target_device_id'],
'syncPeriod': replication_sync_period}
sync_targets.append(sync_target)
optional = {'localSnapCPG': vol_settings['snap_cpg'],
'localUserCPG': local_cpg}
pool = volume_utils.extract_host(volume['host'], level='pool')
domain = self.get_domain(pool)
if domain:
optional["domain"] = domain
try:
self.client.createRemoteCopyGroup(rcg_name, rcg_targets,
optional)
except Exception as ex:
msg = (_("There was an error creating the remote copy "
"group: %s.") %
six.text_type(ex))
LOG.error(msg)
raise exception.VolumeBackendAPIException(data=msg)
# Add volume to remote copy group.
rcg_targets = []
for target in self._replication_targets:
# Only add targets that match the volumes replication mode.
if target['replication_mode'] == replication_mode_num:
rcg_target = {'targetName': target['target_device_id'],
'secVolumeName': vol_name}
rcg_targets.append(rcg_target)
optional = {'volumeAutoCreation': True}
try:
self.client.addVolumeToRemoteCopyGroup(rcg_name, vol_name,
rcg_targets,
optional=optional)
except Exception as ex:
msg = (_("There was an error adding the volume to the remote "
"copy group: %s.") %
six.text_type(ex))
LOG.error(msg)
raise exception.VolumeBackendAPIException(data=msg)
# Check and see if we are in periodic mode. If we are, update
# Remote Copy Group to have a sync period.
if replication_sync_period and (
replication_mode_num == self.PERIODIC):
opt = {'targets': sync_targets}
try:
self.client.modifyRemoteCopyGroup(rcg_name, opt)
except Exception as ex:
msg = (_("There was an error setting the sync period for "
"the remote copy group: %s.") %
six.text_type(ex))
LOG.error(msg)
raise exception.VolumeBackendAPIException(data=msg)
# Start the remote copy.
try:
self.client.startRemoteCopy(rcg_name)
except Exception as ex:
msg = (_("There was an error starting remote copy: %s.") %
six.text_type(ex))
LOG.error(msg)
raise exception.VolumeBackendAPIException(data=msg)
return True
except Exception as ex:
self._do_volume_replication_destroy(volume)
msg = (_("There was an error setting up a remote copy group "
"on the 3PAR arrays: ('%s'). The volume will not be "
"recognized as replication type.") %
six.text_type(ex))
LOG.error(msg)
raise exception.VolumeBackendAPIException(data=msg)
def _do_volume_replication_destroy(self, volume, rcg_name=None):
"""This will completely remove all traces of a remote copy group.
It should be used when deleting a replication enabled volume
or if setting up a remote copy group fails. It will try and do the
following:
-Stop remote copy
-Remove volume from Remote Copy Group on main array
-Delete Remote Copy Group from main array
-Delete volume from main array
"""
if not rcg_name:
rcg_name = self._get_3par_rcg_name(volume['id'])
vol_name = self._get_3par_vol_name(volume['id'])
# Stop remote copy.
try:
self.client.stopRemoteCopy(rcg_name)
except Exception:
pass
# Delete volume from remote copy group on main array.
try:
self.client.removeVolumeFromRemoteCopyGroup(
rcg_name, vol_name, removeFromTarget=True)
except Exception:
pass
# Delete remote copy group on main array.
try:
self.client.removeRemoteCopyGroup(rcg_name)
except Exception:
pass
# Delete volume on the main array.
try:
self.client.deleteVolume(vol_name)
except Exception:
pass
def _delete_replicated_failed_over_volume(self, volume):
old_location, new_location = volume['provider_location'].split(':')
rcg_name = self._get_3par_remote_rcg_name(volume['id'], old_location)
targets = self.client.getRemoteCopyGroup(rcg_name)['targets']
# When failed over, we want to temporarily disable config mirroring
# in order to be allowed to delete the volume and remote copy group
for target in targets:
target_name = target['targetName']
self.client.toggleRemoteCopyConfigMirror(target_name,
mirror_config=False)
# Do regular volume replication destroy now config mirroring is off
try:
self._do_volume_replication_destroy(volume, rcg_name)
except Exception:
msg = (_("The failed-over volume could not be deleted."))
LOG.error(msg)
raise exception.VolumeIsBusy(message=msg)
finally:
# Turn config mirroring back on
for target in targets:
target_name = target['targetName']
self.client.toggleRemoteCopyConfigMirror(target_name,
mirror_config=True)
class TaskWaiter(object):
"""TaskWaiter waits for task to be not active and returns status."""

66
cinder/volume/drivers/hpe/hpe_3par_fc.py
View File

@ -37,7 +37,7 @@ except ImportError:
from oslo_log import log as logging
from cinder import exception
from cinder.i18n import _, _LI
from cinder.i18n import _, _LI, _LW
from cinder.volume import driver
from cinder.volume.drivers.hpe import hpe_3par_common as hpecommon
from cinder.volume.drivers.san import san
@ -91,10 +91,11 @@ class HPE3PARFCDriver(driver.TransferVD,
2.0.21 - Added update_migrated_volume. bug # 1492023
3.0.0 - Rebranded HP to HPE.
3.0.1 - Remove db access for consistency groups
3.0.2 - Adds v2 managed replication support
"""
VERSION = "3.0.1"
VERSION = "3.0.2"
def __init__(self, *args, **kwargs):
super(HPE3PARFCDriver, self).__init__(*args, **kwargs)
@ -105,13 +106,29 @@ class HPE3PARFCDriver(driver.TransferVD,
def _init_common(self):
return hpecommon.HPE3PARCommon(self.configuration)
def _login(self):
def _login(self, timeout=None):
common = self._init_common()
common.do_setup(None)
common.client_login()
# If replication is enabled and we cannot login, we do not want to
# raise an exception so a failover can still be executed.
try:
common.do_setup(None, timeout=timeout)
common.client_login()
except Exception:
if common._replication_enabled:
LOG.warning(_LW("The primary array is not reachable at this "
"time. Since replication is enabled, "
"listing replication targets and failing over "
"a volume can still be performed."))
pass
else:
raise
return common
def _logout(self, common):
# If replication is enabled and we do not have a client ID, we did not
# login, but can still failover. There is no need to logout.
if common.client is None and common._replication_enabled:
return
common.client_logout()
def _check_flags(self, common):
@ -564,3 +581,42 @@ class HPE3PARFCDriver(driver.TransferVD,
raise exception.InvalidVolume(reason)
finally:
self._logout(common)
def get_replication_updates(self, context):
common = self._login()
try:
return common.get_replication_updates(context)
finally:
self._logout(common)
def replication_enable(self, context, volume):
"""Enable replication on a replication capable volume."""
common = self._login()
try:
return common.replication_enable(context, volume)
finally:
self._logout(common)
def replication_disable(self, context, volume):
"""Disable replication on the specified volume."""
common = self._login()
try:
return common.replication_disable(context, volume)
finally:
self._logout(common)
def replication_failover(self, context, volume, secondary):
"""Force failover to a secondary replication target."""
common = self._login(timeout=30)
try:
return common.replication_failover(context, volume, secondary)
finally:
self._logout(common)
def list_replication_targets(self, context, volume):
"""Provides a means to obtain replication targets for a volume."""
common = self._login(timeout=30)
try:
return common.list_replication_targets(context, volume)
finally:
self._logout(common)

64
cinder/volume/drivers/hpe/hpe_3par_iscsi.py
View File

@ -103,10 +103,11 @@ class HPE3PARISCSIDriver(driver.TransferVD,
3.0.1 - Python 3 support
3.0.2 - Remove db access for consistency groups
3.0.3 - Fix multipath dictionary key error. bug #1522062
3.0.4 - Adds v2 managed replication support
"""
VERSION = "3.0.3"
VERSION = "3.0.4"
def __init__(self, *args, **kwargs):
super(HPE3PARISCSIDriver, self).__init__(*args, **kwargs)
@ -116,13 +117,29 @@ class HPE3PARISCSIDriver(driver.TransferVD,
def _init_common(self):
return hpecommon.HPE3PARCommon(self.configuration)
def _login(self):
def _login(self, timeout=None):
common = self._init_common()
common.do_setup(None)
common.client_login()
common.do_setup(None, timeout=timeout)
# If replication is enabled and we cannot login, we do not want to
# raise an exception so a failover can still be executed.
try:
common.client_login()
except Exception:
if common._replication_enabled:
LOG.warning(_LW("The primary array is not reachable at this "
"time. Since replication is enabled, "
"listing replication targets and failing over "
"a volume can still be performed."))
pass
else:
raise
return common
def _logout(self, common):
# If replication is enabled and we do not have a client ID, we did not
# login, but can still failover. There is no need to logout.
if common.client is None and common._replication_enabled:
return
common.client_logout()
def _check_flags(self, common):
@ -863,3 +880,42 @@ class HPE3PARISCSIDriver(driver.TransferVD,
raise exception.InvalidVolume(reason)
finally:
self._logout(common)
def get_replication_updates(self, context):
common = self._login()
try:
return common.get_replication_updates(context)
finally:
self._logout(common)
def replication_enable(self, context, volume):
"""Enable replication on a replication capable volume."""
common = self._login()
try:
return common.replication_enable(context, volume)
finally:
self._logout(common)
def replication_disable(self, context, volume):
"""Disable replication on the specified volume."""
common = self._login()
try:
return common.replication_disable(context, volume)
finally:
self._logout(common)
def replication_failover(self, context, volume, secondary):
"""Force failover to a secondary replication target."""
common = self._login(timeout=30)
try:
return common.replication_failover(context, volume, secondary)
finally:
self._logout(common)
def list_replication_targets(self, context, volume):
"""Provides a means to obtain replication targets for a volume."""
common = self._login(timeout=30)
try:
return common.list_replication_targets(context, volume)
finally:
self._logout(common)