deb-python-os-brick/os_brick/initiator/connectors/vmware.py
Gorka Eguileor 400ca5d6db Refactor iSCSI disconnect
This patch refactors iSCSI disconnect code changing the approach to one
that just uses `iscsiadm -m session` and sysfs to get all the required
information: devices from the connection, multipath system device name,
multipath name, the WWN for the block devices...

By doing so, not only do we fix a good number of bugs, but we also
improve the reliability and speed of the mechanism.

A good example of improvements and benefits achieved by this patch are:

- Common code for multipath and single path disconnects.

- No more querying iSCSI devices for their WWN (page 0x83) removing
  delays and issue on flaky connections.

- All devices are properly cleaned even if they are not part of the
  multipath.

- We wait for device removal and do it in parallel if there are
  multiple.

- Removed usage of `multipath -l` to find devices which is really slow
  with flaky connections and didn't work when called with a device from
  a path that is down.

- Prevent losing data when detaching, currently if the multipath flush
  fails for any other reason than "in use" we silently continue with the
  removal.  That is the case when all paths are momentarily down.

- Adds a new mechanism for the caller of the disconnect to specify that
  it's acceptable to lose data and that it's more important to leave a
  clean system.  That is the case if we are creating a volume from an
  image, since the volume will just be set to error, but we don't want
  leftovers.  Optionally we can tell os-brick to ignore errors and don't
  raise an exception if the flush fails.

- Add a warning when we could be leaving leftovers behind due to
  disconnect issues.

- Action retries (like multipath flush) will now only log the final
  exception instead of logging all the exceptions.

- Flushes of individual paths now use exponential backoff retries
  instead of random retries between 0.2 and 2 seconds (from oslo
  library).

- We no longer use symlinks from `/dev/disk/by-path`, `/dev/disk/by-id`,
  or `/dev/mapper` to find devices or multipaths, as they could be
  leftovers from previous runs.

- With high failure rates (above 30%) some CLI calls will enter into a
  weird state where they wait forever, so we add a timeout mechanism in
  our `execute` method and add it to those specific calls.

Closes-Bug: #1502534
Change-Id: I058ff0a0e5ad517507dc3cda39087c913558561d
2017-05-31 15:31:20 +02:00

278 lines
12 KiB
Python

# Copyright (c) 2016 VMware, Inc.
# All Rights Reserved.
#
# 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 os
import tempfile
from oslo_log import log as logging
from oslo_utils import fileutils
try:
from oslo_vmware import api
from oslo_vmware import exceptions as oslo_vmw_exceptions
from oslo_vmware import image_transfer
from oslo_vmware.objects import datastore
from oslo_vmware import rw_handles
from oslo_vmware import vim_util
except ImportError:
vim_util = None
import six
from os_brick import exception
from os_brick.i18n import _
from os_brick.initiator import initiator_connector
LOG = logging.getLogger(__name__)
class VmdkConnector(initiator_connector.InitiatorConnector):
"""Connector for volumes created by the VMDK driver.
This connector is only used for backup and restore of Cinder volumes.
"""
TMP_IMAGES_DATASTORE_FOLDER_PATH = "cinder_temp"
def __init__(self, *args, **kwargs):
# Check if oslo.vmware library is available.
if vim_util is None:
message = _("Missing oslo_vmware python module, ensure oslo.vmware"
" library is installed and available.")
raise exception.BrickException(message=message)
super(VmdkConnector, self).__init__(*args, **kwargs)
self._ip = None
self._port = None
self._username = None
self._password = None
self._api_retry_count = None
self._task_poll_interval = None
self._ca_file = None
self._insecure = None
self._tmp_dir = None
self._timeout = None
@staticmethod
def get_connector_properties(root_helper, *args, **kwargs):
return {}
def check_valid_device(self, path, run_as_root=True):
pass
def get_volume_paths(self, connection_properties):
return []
def get_search_path(self):
return None
def get_all_available_volumes(self, connection_properties=None):
pass
def _load_config(self, connection_properties):
config = connection_properties['config']
self._ip = config['vmware_host_ip']
self._port = config['vmware_host_port']
self._username = config['vmware_host_username']
self._password = config['vmware_host_password']
self._api_retry_count = config['vmware_api_retry_count']
self._task_poll_interval = config['vmware_task_poll_interval']
self._ca_file = config['vmware_ca_file']
self._insecure = config['vmware_insecure']
self._tmp_dir = config['vmware_tmp_dir']
self._timeout = config['vmware_image_transfer_timeout_secs']
def _create_session(self):
return api.VMwareAPISession(self._ip,
self._username,
self._password,
self._api_retry_count,
self._task_poll_interval,
port=self._port,
cacert=self._ca_file,
insecure=self._insecure)
def _create_temp_file(self, *args, **kwargs):
fileutils.ensure_tree(self._tmp_dir)
fd, tmp = tempfile.mkstemp(dir=self._tmp_dir, *args, **kwargs)
os.close(fd)
return tmp
def _download_vmdk(
self, tmp_file_path, session, backing, vmdk_path, vmdk_size):
with open(tmp_file_path, "wb") as tmp_file:
image_transfer.copy_stream_optimized_disk(
None,
self._timeout,
tmp_file,
session=session,
host=self._ip,
port=self._port,
vm=backing,
vmdk_file_path=vmdk_path,
vmdk_size=vmdk_size)
def connect_volume(self, connection_properties):
# Download the volume vmdk from vCenter server to a temporary file
# and return its path.
self._load_config(connection_properties)
session = self._create_session()
tmp_file_path = self._create_temp_file(
suffix=".vmdk", prefix=connection_properties['volume_id'])
backing = vim_util.get_moref(connection_properties['volume'],
"VirtualMachine")
vmdk_path = connection_properties['vmdk_path']
vmdk_size = connection_properties['vmdk_size']
try:
self._download_vmdk(
tmp_file_path, session, backing, vmdk_path, vmdk_size)
finally:
session.logout()
# Save the last modified time of the temporary so that we can decide
# whether to upload the file back to vCenter server during disconnect.
last_modified = os.path.getmtime(tmp_file_path)
return {'path': tmp_file_path, 'last_modified': last_modified}
def _snapshot_exists(self, session, backing):
snapshot = session.invoke_api(vim_util,
'get_object_property',
session.vim,
backing,
'snapshot')
if snapshot is None or snapshot.rootSnapshotList is None:
return False
return len(snapshot.rootSnapshotList) != 0
def _create_temp_ds_folder(self, session, ds_folder_path, dc_ref):
fileManager = session.vim.service_content.fileManager
try:
session.invoke_api(session.vim,
'MakeDirectory',
fileManager,
name=ds_folder_path,
datacenter=dc_ref)
except oslo_vmw_exceptions.FileAlreadyExistsException:
pass
# Note(vbala) remove this method when we implement it in oslo.vmware
def _upload_vmdk(
self, read_handle, host, port, dc_name, ds_name, cookies,
upload_file_path, file_size, cacerts, timeout_secs):
write_handle = rw_handles.FileWriteHandle(host,
port,
dc_name,
ds_name,
cookies,
upload_file_path,
file_size,
cacerts=cacerts)
image_transfer._start_transfer(read_handle, write_handle, timeout_secs)
def _disconnect(self, tmp_file_path, session, ds_ref, dc_ref, vmdk_path):
# The restored volume is in compressed (streamOptimized) format.
# So we upload it to a temporary location in vCenter datastore and copy
# the compressed vmdk to the volume vmdk. The copy operation
# decompresses the disk to a format suitable for attaching to Nova
# instances in vCenter.
dstore = datastore.get_datastore_by_ref(session, ds_ref)
ds_path = dstore.build_path(
VmdkConnector.TMP_IMAGES_DATASTORE_FOLDER_PATH,
os.path.basename(tmp_file_path))
self._create_temp_ds_folder(
session, six.text_type(ds_path.parent), dc_ref)
with open(tmp_file_path, "rb") as tmp_file:
dc_name = session.invoke_api(
vim_util, 'get_object_property', session.vim, dc_ref, 'name')
cookies = session.vim.client.options.transport.cookiejar
cacerts = self._ca_file if self._ca_file else not self._insecure
self._upload_vmdk(
tmp_file, self._ip, self._port, dc_name, dstore.name, cookies,
ds_path.rel_path, os.path.getsize(tmp_file_path), cacerts,
self._timeout)
# Delete the current volume vmdk because the copy operation does not
# overwrite.
LOG.debug("Deleting %s", vmdk_path)
disk_mgr = session.vim.service_content.virtualDiskManager
task = session.invoke_api(session.vim,
'DeleteVirtualDisk_Task',
disk_mgr,
name=vmdk_path,
datacenter=dc_ref)
session.wait_for_task(task)
src = six.text_type(ds_path)
LOG.debug("Copying %(src)s to %(dest)s", {'src': src,
'dest': vmdk_path})
task = session.invoke_api(session.vim,
'CopyVirtualDisk_Task',
disk_mgr,
sourceName=src,
sourceDatacenter=dc_ref,
destName=vmdk_path,
destDatacenter=dc_ref)
session.wait_for_task(task)
# Delete the compressed vmdk at the temporary location.
LOG.debug("Deleting %s", src)
file_mgr = session.vim.service_content.fileManager
task = session.invoke_api(session.vim,
'DeleteDatastoreFile_Task',
file_mgr,
name=src,
datacenter=dc_ref)
session.wait_for_task(task)
def disconnect_volume(self, connection_properties, device_info,
force=False, ignore_errors=False):
tmp_file_path = device_info['path']
if not os.path.exists(tmp_file_path):
msg = _("Vmdk: %s not found.") % tmp_file_path
raise exception.NotFound(message=msg)
session = None
try:
# We upload the temporary file to vCenter server only if it is
# modified after connect_volume.
if os.path.getmtime(tmp_file_path) > device_info['last_modified']:
self._load_config(connection_properties)
session = self._create_session()
backing = vim_util.get_moref(connection_properties['volume'],
"VirtualMachine")
# Currently there is no way we can restore the volume if it
# contains redo-log based snapshots (bug 1599026).
if self._snapshot_exists(session, backing):
msg = (_("Backing of volume: %s contains one or more "
"snapshots; cannot disconnect.") %
connection_properties['volume_id'])
raise exception.BrickException(message=msg)
ds_ref = vim_util.get_moref(
connection_properties['datastore'], "Datastore")
dc_ref = vim_util.get_moref(
connection_properties['datacenter'], "Datacenter")
vmdk_path = connection_properties['vmdk_path']
self._disconnect(
tmp_file_path, session, ds_ref, dc_ref, vmdk_path)
finally:
os.remove(tmp_file_path)
if session:
session.logout()
def extend_volume(self, connection_properties):
raise NotImplementedError