openstack/keystone
Code Issues Proposed changes
keystone/keystone/cmd/cli.py
Stephen Finucane f174b4fa7c sql: Integrate alembic 
Switch to alembic for real by integrating it into the 'db sync' command
flow. From a user-facing perspective, things should remain pretty much
the same as before, with the key difference being that version
information (i.e. what's shown by 'keystone-manage db_sync --check' or
'keystone-manage db_version') will now take the form of a hash rather
than an integer. There are a few differences for contributors however.
The changes are described in the included release note and
documentation.

Note that there are a couple of important design decisions here that are
worth examining:

- We drop the idea of the 'data_migration' branch entirely and the
  'keystone-manage db_sync --migrate' command is now a no-op. Neutron
  doesn't do data migrations like we do and yet they manage just fine.
  Dropping this gets us closer to neutron's behavior, which is a good
  thing for users.

- We haven't re-added the ability to specify a version when doing
  'db_sync'. Neutron has this, but the logic needed to get this working
  is complex and of questionable value. We've managed without the
  ability to sync to a version since Newton and can continue to do so
  until someone asks for it (and does the work).

- sqlalchemy-migrate is not removed entirely. Instead, upon doing a
  'db_sync' we will apply all sqlalchemy-migrate migrations up to the
  final '079_expand_update_local_id_limit' migration and dummy apply the
  initial alembic migration, after which we will switch over to alembic.
  In a future release we can remove the sqlalchemy-migrate migrations
  and rely entirely on alembic. Until then, keeping this allows fast
  forward upgrades to continue as a thing.

- Related to the above, we always apply *all* sqlalchemy-migrate
  migrations when calling 'db_sync', even if this command is called with
  e.g. '--expand' (meaning only apply the expand branch). This is
  because there is at most one "real" migration to apply, the Xena-era
  '079_expand_update_local_id_limit' migration, which is an expand-only
  migration. There is no risk to applying the empty "data_migration" and
  "contract" parts of this migration, and applying everything in one go
  results in *much* simpler logic.

Future changes will update documentation and add developer tooling for
(auto-)generating new migrations, a la 'neutron-db-manage revision'.

Change-Id: Ia376cb87f5159a4e79e2cfbab8442b6bcead708f
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
2022-06-20 13:29:58 +01:00

1367 lines
53 KiB
Python
Raw Blame History

# Copyright 2012 OpenStack Foundation
#
# 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 argparse
import datetime
import os
import sys
import uuid
from oslo_config import cfg
from oslo_db import exception as db_exception
from oslo_log import log
from oslo_serialization import jsonutils
import pbr.version
from keystone.cmd import bootstrap
from keystone.cmd import doctor
from keystone.common import driver_hints
from keystone.common import fernet_utils
from keystone.common import jwt_utils
from keystone.common import sql
from keystone.common.sql import upgrades
from keystone.common import utils
import keystone.conf
from keystone.credential.providers import fernet as credential_fernet
from keystone import exception
from keystone.federation import idp
from keystone.federation import utils as mapping_engine
from keystone.i18n import _
from keystone.server import backends
CONF = keystone.conf.CONF
LOG = log.getLogger(__name__)
class BaseApp(object):
name = None
@classmethod
def add_argument_parser(cls, subparsers):
parser = subparsers.add_parser(cls.name, help=cls.__doc__)
parser.set_defaults(cmd_class=cls)
return parser
class BootStrap(BaseApp):
"""Perform the basic bootstrap process."""
name = "bootstrap"
def __init__(self):
self.bootstrapper = bootstrap.Bootstrapper()
@classmethod
def add_argument_parser(cls, subparsers):
parser = super(BootStrap, cls).add_argument_parser(subparsers)
parser.add_argument('--bootstrap-username', default='admin',
metavar='OS_BOOTSTRAP_USERNAME',
help=('The username of the initial keystone '
'user during bootstrap process.'))
# NOTE(morganfainberg): See below for ENV Variable that can be used
# in lieu of the command-line arguments.
parser.add_argument('--bootstrap-password', default=None,
metavar='OS_BOOTSTRAP_PASSWORD',
help='The bootstrap user password')
parser.add_argument('--bootstrap-project-name', default='admin',
metavar='OS_BOOTSTRAP_PROJECT_NAME',
help=('The initial project created during the '
'keystone bootstrap process.'))
parser.add_argument('--bootstrap-role-name', default='admin',
metavar='OS_BOOTSTRAP_ROLE_NAME',
help=('The initial role-name created during the '
'keystone bootstrap process.'))
parser.add_argument('--bootstrap-service-name', default='keystone',
metavar='OS_BOOTSTRAP_SERVICE_NAME',
help=('The initial name for the initial identity '
'service created during the keystone '
'bootstrap process.'))
parser.add_argument('--bootstrap-admin-url',
metavar='OS_BOOTSTRAP_ADMIN_URL',
help=('The initial identity admin url created '
'during the keystone bootstrap process. '
'e.g. http://127.0.0.1:5000/v3'))
parser.add_argument('--bootstrap-public-url',
metavar='OS_BOOTSTRAP_PUBLIC_URL',
help=('The initial identity public url created '
'during the keystone bootstrap process. '
'e.g. http://127.0.0.1:5000/v3'))
parser.add_argument('--bootstrap-internal-url',
metavar='OS_BOOTSTRAP_INTERNAL_URL',
help=('The initial identity internal url created '
'during the keystone bootstrap process. '
'e.g. http://127.0.0.1:5000/v3'))
parser.add_argument('--bootstrap-region-id',
metavar='OS_BOOTSTRAP_REGION_ID',
help=('The initial region_id endpoints will be '
'placed in during the keystone bootstrap '
'process.'))
parser.add_argument('--immutable-roles',
default=True,
action='store_true',
help=('Whether default roles (admin, member, and '
'reader) should be immutable. This is the '
'default.'))
parser.add_argument('--no-immutable-roles',
default=False,
action='store_true',
help=('Whether default roles (admin, member, and '
'reader) should be immutable. Immutable '
'default roles is the default, use this '
'flag to opt out of immutable default '
'roles.'))
return parser
def do_bootstrap(self):
"""Perform the bootstrap actions.
Create bootstrap user, project, and role so that CMS, humans, or
scripts can continue to perform initial setup (domains, projects,
services, endpoints, etc) of Keystone when standing up a new
deployment.
"""
self.username = (
os.environ.get('OS_BOOTSTRAP_USERNAME') or
CONF.command.bootstrap_username)
self.project_name = (
os.environ.get('OS_BOOTSTRAP_PROJECT_NAME') or
CONF.command.bootstrap_project_name)
self.role_name = (
os.environ.get('OS_BOOTSTRAP_ROLE_NAME') or
CONF.command.bootstrap_role_name)
self.password = (
os.environ.get('OS_BOOTSTRAP_PASSWORD') or
CONF.command.bootstrap_password)
self.service_name = (
os.environ.get('OS_BOOTSTRAP_SERVICE_NAME') or
CONF.command.bootstrap_service_name)
self.admin_url = (
os.environ.get('OS_BOOTSTRAP_ADMIN_URL') or
CONF.command.bootstrap_admin_url)
self.public_url = (
os.environ.get('OS_BOOTSTRAP_PUBLIC_URL') or
CONF.command.bootstrap_public_url)
self.internal_url = (
os.environ.get('OS_BOOTSTRAP_INTERNAL_URL') or
CONF.command.bootstrap_internal_url)
self.region_id = (
os.environ.get('OS_BOOTSTRAP_REGION_ID') or
CONF.command.bootstrap_region_id)
self.service_id = None
self.endpoints = None
if self.password is None:
print(_('ERROR: Either --bootstrap-password argument or '
'OS_BOOTSTRAP_PASSWORD must be set.'))
sys.exit(1)
self.bootstrapper.admin_password = self.password
self.bootstrapper.admin_username = self.username
self.bootstrapper.project_name = self.project_name
self.bootstrapper.admin_role_name = self.role_name
self.bootstrapper.service_name = self.service_name
self.bootstrapper.service_id = self.service_id
self.bootstrapper.admin_url = self.admin_url
self.bootstrapper.public_url = self.public_url
self.bootstrapper.internal_url = self.internal_url
self.bootstrapper.region_id = self.region_id
if CONF.command.no_immutable_roles:
self.bootstrapper.immutable_roles = False
else:
self.bootstrapper.immutable_roles = True
self.bootstrapper.bootstrap()
self.reader_role_id = self.bootstrapper.reader_role_id
self.member_role_id = self.bootstrapper.member_role_id
self.role_id = self.bootstrapper.admin_role_id
self.project_id = self.bootstrapper.project_id
@classmethod
def main(cls):
klass = cls()
klass.do_bootstrap()
class Doctor(BaseApp):
"""Diagnose common problems with keystone deployments."""
name = 'doctor'
@classmethod
def add_argument_parser(cls, subparsers):
parser = super(Doctor, cls).add_argument_parser(subparsers)
return parser
@staticmethod
def main():
# Return a non-zero exit code if we detect any symptoms.
raise SystemExit(doctor.diagnose())
class DbSync(BaseApp):
"""Sync the database."""
name = 'db_sync'
@classmethod
def add_argument_parser(cls, subparsers):
parser = super(DbSync, cls).add_argument_parser(subparsers)
parser.add_argument(
'version',
default=None,
nargs='?',
help=(
'Migrate the database up to a specified version. '
'If not provided, db_sync will migrate the database to the '
'latest known version. '
'Schema downgrades are not supported.'
),
)
group = parser.add_mutually_exclusive_group()
group.add_argument(
'--expand',
default=False,
action='store_true',
help=(
'Expand the database schema in preparation for data migration.'
),
)
group.add_argument(
'--migrate',
default=False,
action='store_true',
help=(
'Copy all data that needs to be migrated within the database '
'ahead of starting the first keystone node upgraded to the '
'new release. '
'This command should be run after the --expand command. '
'Once the --migrate command has completed, you can upgrade '
'all your keystone nodes to the new release and restart them.'
),
)
group.add_argument(
'--contract',
default=False,
action='store_true',
help=(
'Remove any database tables and columns that are no longer '
'required. This command should be run after all keystone '
'nodes are running the new release.'
),
)
group.add_argument(
'--check',
default=False,
action='store_true',
help=(
'Check for outstanding database actions that still need to be '
'executed. This command can be used to verify the condition '
'of the current database state.'
),
)
return parser
@classmethod
def check_db_sync_status(cls):
status = 0
try:
expand_version = upgrades.get_db_version(branch='expand')
except db_exception.DBMigrationError:
LOG.info(
'Your database is not currently under version '
'control or the database is already controlled. '
'Your first step is to run `keystone-manage db_sync --expand`.'
)
return 2
if isinstance(expand_version, int):
# we're still using sqlalchemy-migrate
LOG.info(
'Your database is currently using legacy version control. '
'Your first step is to run `keystone-manage db_sync --expand`.'
)
return 2
try:
contract_version = upgrades.get_db_version(branch='contract')
except db_exception.DBMigrationError:
contract_version = None
heads = upgrades.get_current_heads()
if (
upgrades.EXPAND_BRANCH not in heads or
heads[upgrades.EXPAND_BRANCH] != expand_version
):
LOG.info('Your database is not up to date. Your first step is '
'to run `keystone-manage db_sync --expand`.')
status = 2
elif (
upgrades.CONTRACT_BRANCH not in heads or
heads[upgrades.CONTRACT_BRANCH] != contract_version
):
LOG.info('Expand version is ahead of contract. Your next '
'step is to run `keystone-manage db_sync --contract`.')
status = 4
else:
LOG.info('All db_sync commands are upgraded to the same '
'version and up-to-date.')
LOG.info(
'Current repository versions:\n'
'Expand: %(expand)s (head: %(expand_head)s)\n'
'Contract: %(contract)s (head: %(contract_head)s)',
{
'expand': expand_version,
'expand_head': heads.get(upgrades.EXPAND_BRANCH),
'contract': contract_version,
'contract_head': heads.get(upgrades.CONTRACT_BRANCH),
},
)
return status
@staticmethod
def main():
if CONF.command.check:
sys.exit(DbSync.check_db_sync_status())
elif CONF.command.expand:
upgrades.expand_schema()
elif CONF.command.migrate:
upgrades.migrate_data()
elif CONF.command.contract:
upgrades.contract_schema()
else:
upgrades.offline_sync_database_to_version(
CONF.command.version)
class DbVersion(BaseApp):
"""Print the current migration version of the database."""
name = 'db_version'
@staticmethod
def main():
print(upgrades.get_db_version())
class BasePermissionsSetup(BaseApp):
"""Common user/group setup for file permissions."""
@classmethod
def add_argument_parser(cls, subparsers):
parser = super(BasePermissionsSetup,
cls).add_argument_parser(subparsers)
running_as_root = (os.geteuid() == 0)
parser.add_argument('--keystone-user', required=running_as_root)
parser.add_argument('--keystone-group', required=running_as_root)
return parser
@staticmethod
def get_user_group():
keystone_user_id = None
keystone_group_id = None
try:
a = CONF.command.keystone_user
if a:
keystone_user_id = utils.get_unix_user(a)[0]
except KeyError:
raise ValueError("Unknown user '%s' in --keystone-user" % a)
try:
a = CONF.command.keystone_group
if a:
keystone_group_id = utils.get_unix_group(a)[0]
except KeyError:
raise ValueError("Unknown group '%s' in --keystone-group" % a)
return keystone_user_id, keystone_group_id
@classmethod
def initialize_fernet_repository(
cls, keystone_user_id, keystone_group_id, config_group=None):
conf_group = getattr(CONF, config_group)
futils = fernet_utils.FernetUtils(
conf_group.key_repository,
conf_group.max_active_keys,
config_group
)
futils.create_key_directory(keystone_user_id, keystone_group_id)
if futils.validate_key_repository(requires_write=True):
futils.initialize_key_repository(
keystone_user_id, keystone_group_id)
@classmethod
def rotate_fernet_repository(
cls, keystone_user_id, keystone_group_id, config_group=None):
conf_group = getattr(CONF, config_group)
futils = fernet_utils.FernetUtils(
conf_group.key_repository,
conf_group.max_active_keys,
config_group
)
if futils.validate_key_repository(requires_write=True):
futils.rotate_keys(keystone_user_id, keystone_group_id)
class FernetSetup(BasePermissionsSetup):
"""Setup key repositories for Fernet tokens and auth receipts.
This also creates a primary key used for both creating and validating
Fernet tokens and auth receipts. To improve security, you should rotate
your keys (using keystone-manage fernet_rotate, for example).
"""
name = 'fernet_setup'
@classmethod
def main(cls):
keystone_user_id, keystone_group_id = cls.get_user_group()
cls.initialize_fernet_repository(
keystone_user_id, keystone_group_id, 'fernet_tokens')
if (os.path.abspath(CONF.fernet_tokens.key_repository) !=
os.path.abspath(CONF.fernet_receipts.key_repository)):
cls.initialize_fernet_repository(
keystone_user_id, keystone_group_id, 'fernet_receipts')
elif(CONF.fernet_tokens.max_active_keys !=
CONF.fernet_receipts.max_active_keys):
# WARNING(adriant): If the directories are the same,
# 'max_active_keys' is ignored from fernet_receipts in favor of
# fernet_tokens to avoid a potential mismatch. Only if the
# directories are different do we create a different one for
# receipts, and then respect 'max_active_keys' for receipts.
LOG.warning(
"Receipt and Token fernet key directories are the same "
"but `max_active_keys` is different. Receipt "
"`max_active_keys` will be ignored in favor of Token "
"`max_active_keys`."
)
class FernetRotate(BasePermissionsSetup):
"""Rotate Fernet encryption keys.
This assumes you have already run keystone-manage fernet_setup.
A new primary key is placed into rotation, which is used for new tokens.
The old primary key is demoted to secondary, which can then still be used
for validating tokens. Excess secondary keys (beyond [fernet_tokens]
max_active_keys) are revoked. Revoked keys are permanently deleted. A new
staged key will be created and used to validate tokens. The next time key
rotation takes place, the staged key will be put into rotation as the
primary key.
Rotating keys too frequently, or with [fernet_tokens] max_active_keys set
too low, will cause tokens to become invalid prior to their expiration.
"""
name = 'fernet_rotate'
@classmethod
def main(cls):
keystone_user_id, keystone_group_id = cls.get_user_group()
cls.rotate_fernet_repository(
keystone_user_id, keystone_group_id, 'fernet_tokens')
if (os.path.abspath(CONF.fernet_tokens.key_repository) !=
os.path.abspath(CONF.fernet_receipts.key_repository)):
cls.rotate_fernet_repository(
keystone_user_id, keystone_group_id, 'fernet_receipts')
class CreateJWSKeyPair(BasePermissionsSetup):
"""Create a key pair for signing and validating JWS tokens.
This command creates a public and private key pair to use for signing and
validating JWS token signatures. The key pair is written to the directory
where the command is invoked.
"""
name = 'create_jws_keypair'
@classmethod
def add_argument_parser(cls, subparsers):
parser = super(CreateJWSKeyPair, cls).add_argument_parser(subparsers)
parser.add_argument(
'--force', action='store_true',
help=('Forcibly overwrite keys if they already exist')
)
return parser
@classmethod
def main(cls):
current_directory = os.getcwd()
private_key_path = os.path.join(current_directory, 'private.pem')
public_key_path = os.path.join(current_directory, 'public.pem')
if os.path.isfile(private_key_path) and not CONF.command.force:
raise SystemExit(_('Private key %(path)s already exists')
% {'path': private_key_path})
if os.path.isfile(public_key_path) and not CONF.command.force:
raise SystemExit(_('Public key %(path)s already exists')
% {'path': public_key_path})
jwt_utils.create_jws_keypair(private_key_path, public_key_path)
class TokenSetup(BasePermissionsSetup):
"""Setup a key repository for tokens.
This also creates a primary key used for both creating and validating
tokens. To improve security, you should rotate your keys (using
keystone-manage token_rotate, for example).
"""
name = 'token_setup'
@classmethod
def main(cls):
keystone_user_id, keystone_group_id = cls.get_user_group()
cls.initialize_fernet_repository(
keystone_user_id, keystone_group_id, 'fernet_tokens')
class TokenRotate(BasePermissionsSetup):
"""Rotate token encryption keys.
This assumes you have already run keystone-manage token_setup.
A new primary key is placed into rotation, which is used for new tokens.
The old primary key is demoted to secondary, which can then still be used
for validating tokens. Excess secondary keys (beyond [token]
max_active_keys) are revoked. Revoked keys are permanently deleted. A new
staged key will be created and used to validate tokens. The next time key
rotation takes place, the staged key will be put into rotation as the
primary key.
Rotating keys too frequently, or with [token] max_active_keys set
too low, will cause tokens to become invalid prior to their expiration.
"""
name = 'token_rotate'
@classmethod
def main(cls):
keystone_user_id, keystone_group_id = cls.get_user_group()
cls.rotate_fernet_repository(
keystone_user_id, keystone_group_id, 'fernet_tokens')
class ReceiptSetup(BasePermissionsSetup):
"""Setup a key repository for auth receipts.
This also creates a primary key used for both creating and validating
receipts. To improve security, you should rotate your keys (using
keystone-manage receipt_rotate, for example).
"""
name = 'receipt_setup'
@classmethod
def main(cls):
keystone_user_id, keystone_group_id = cls.get_user_group()
cls.initialize_fernet_repository(
keystone_user_id, keystone_group_id, 'fernet_receipts')
class ReceiptRotate(BasePermissionsSetup):
"""Rotate auth receipts encryption keys.
This assumes you have already run keystone-manage receipt_setup.
A new primary key is placed into rotation, which is used for new receipts.
The old primary key is demoted to secondary, which can then still be used
for validating receipts. Excess secondary keys (beyond [receipt]
max_active_keys) are revoked. Revoked keys are permanently deleted. A new
staged key will be created and used to validate receipts. The next time key
rotation takes place, the staged key will be put into rotation as the
primary key.
Rotating keys too frequently, or with [receipt] max_active_keys set
too low, will cause receipts to become invalid prior to their expiration.
"""
name = 'receipt_rotate'
@classmethod
def main(cls):
keystone_user_id, keystone_group_id = cls.get_user_group()
cls.rotate_fernet_repository(
keystone_user_id, keystone_group_id, 'fernet_receipts')
class CredentialSetup(BasePermissionsSetup):
"""Setup a Fernet key repository for credential encryption.
The purpose of this command is very similar to `keystone-manage
fernet_setup` only the keys included in this repository are for encrypting
and decrypting credential secrets instead of token payloads. Keys can be
rotated using `keystone-manage credential_rotate`.
"""
name = 'credential_setup'
@classmethod
def main(cls):
futils = fernet_utils.FernetUtils(
CONF.credential.key_repository,
credential_fernet.MAX_ACTIVE_KEYS,
'credential'
)
keystone_user_id, keystone_group_id = cls.get_user_group()
futils.create_key_directory(keystone_user_id, keystone_group_id)
if futils.validate_key_repository(requires_write=True):
futils.initialize_key_repository(
keystone_user_id,
keystone_group_id
)
class CredentialRotate(BasePermissionsSetup):
"""Rotate Fernet encryption keys for credential encryption.
This assumes you have already run `keystone-manage credential_setup`.
A new primary key is placed into rotation only if all credentials are
encrypted with the current primary key. If any credentials are encrypted
with a secondary key the rotation will abort. This protects against
removing a key that is still required to decrypt credentials. Once a key is
removed from the repository, it is impossible to recover the original data
without restoring from a backup external to keystone (more on backups
below). To make sure all credentials are encrypted with the latest primary
key, please see the `keystone-manage credential_migrate` command. Since the
maximum number of keys in the credential repository is 3, once all
credentials are encrypted with the latest primary key we can safely
introduce a new primary key. All credentials will still be decryptable
since they are all encrypted with the only secondary key in the repository.
It is imperitive to understand the importance of backing up keys used to
encrypt credentials. In the event keys are overrotated, applying a key
repository from backup can help recover otherwise useless credentials.
Persisting snapshots of the key repository in secure and encrypted source
control, or a dedicated key management system are good examples of
encryption key backups.
The `keystone-manage credential_rotate` and `keystone-manage
credential_migrate` commands are intended to be done in sequence. After
performing a rotation, a migration must be done before performing another
rotation. This ensures we don't over-rotate encryption keys.
"""
name = 'credential_rotate'
def __init__(self):
drivers = backends.load_backends()
self.credential_provider_api = drivers['credential_provider_api']
self.credential_api = drivers['credential_api']
def validate_primary_key(self):
crypto, keys = credential_fernet.get_multi_fernet_keys()
primary_key_hash = credential_fernet.primary_key_hash(keys)
credentials = self.credential_api.driver.list_credentials(
driver_hints.Hints()
)
for credential in credentials:
if credential['key_hash'] != primary_key_hash:
msg = _('Unable to rotate credential keys because not all '
'credentials are encrypted with the primary key. '
'Please make sure all credentials have been encrypted '
'with the primary key using `keystone-manage '
'credential_migrate`.')
raise SystemExit(msg)
@classmethod
def main(cls):
futils = fernet_utils.FernetUtils(
CONF.credential.key_repository,
credential_fernet.MAX_ACTIVE_KEYS,
'credential'
)
keystone_user_id, keystone_group_id = cls.get_user_group()
if futils.validate_key_repository(requires_write=True):
klass = cls()
klass.validate_primary_key()
futils.rotate_keys(keystone_user_id, keystone_group_id)
class CredentialMigrate(BasePermissionsSetup):
"""Provides the ability to encrypt credentials using a new primary key.
This assumes that there is already a credential key repository in place and
that the database backend has been upgraded to at least the Newton schema.
If the credential repository doesn't exist yet, you can use
``keystone-manage credential_setup`` to create one.
"""
name = 'credential_migrate'
def __init__(self):
drivers = backends.load_backends()
self.credential_provider_api = drivers['credential_provider_api']
self.credential_api = drivers['credential_api']
def migrate_credentials(self):
crypto, keys = credential_fernet.get_multi_fernet_keys()
primary_key_hash = credential_fernet.primary_key_hash(keys)
# FIXME(lbragstad): We *should* be able to use Hints() to ask only for
# credentials that have a key_hash equal to a secondary key hash or
# None, but Hints() doesn't seem to honor None values. See
# https://bugs.launchpad.net/keystone/+bug/1614154. As a workaround -
# we have to ask for *all* credentials and filter them ourselves.
credentials = self.credential_api.driver.list_credentials(
driver_hints.Hints()
)
for credential in credentials:
if credential['key_hash'] != primary_key_hash:
# If the key_hash isn't None but doesn't match the
# primary_key_hash, then we know the credential was encrypted
# with a secondary key. Let's decrypt it, and send it through
# the update path to re-encrypt it with the new primary key.
decrypted_blob = self.credential_provider_api.decrypt(
credential['encrypted_blob']
)
cred = {'blob': decrypted_blob}
self.credential_api.update_credential(
credential['id'],
cred
)
@classmethod
def main(cls):
# Check to make sure we have a repository that works...
futils = fernet_utils.FernetUtils(
CONF.credential.key_repository,
credential_fernet.MAX_ACTIVE_KEYS,
'credential'
)
futils.validate_key_repository(requires_write=True)
klass = cls()
klass.migrate_credentials()
class TrustFlush(BaseApp):
"""Flush expired and non-expired soft deleted trusts from the backend."""
name = 'trust_flush'
@classmethod
def add_argument_parser(cls, subparsers):
parser = super(TrustFlush, cls).add_argument_parser(subparsers)
parser.add_argument('--project-id', default=None,
help=('The id of the project of which the '
'expired or non-expired soft-deleted '
'trusts is to be purged'))
parser.add_argument('--trustor-user-id', default=None,
help=('The id of the trustor of which the '
'expired or non-expired soft-deleted '
'trusts is to be purged'))
parser.add_argument('--trustee-user-id', default=None,
help=('The id of the trustee of which the '
'expired or non-expired soft-deleted '
'trusts is to be purged'))
parser.add_argument('--date', default=datetime.datetime.utcnow(),
help=('The date of which the expired or '
'non-expired soft-deleted trusts older '
'than that will be purged. The format of '
'the date to be "DD-MM-YYYY". If no date '
'is supplied keystone-manage will use the '
'system clock time at runtime'))
return parser
@classmethod
def main(cls):
drivers = backends.load_backends()
trust_manager = drivers['trust_api']
if CONF.command.date:
if not isinstance(CONF.command.date, datetime.datetime):
try:
CONF.command.date = datetime.datetime.strptime(
CONF.command.date, '%d-%m-%Y')
except KeyError:
raise ValueError("'%s'Invalid input for date, should be "
"DD-MM-YYYY", CONF.command.date)
else:
LOG.info("No date is supplied, keystone-manage will use the "
"system clock time at runtime ")
trust_manager.flush_expired_and_soft_deleted_trusts(
project_id=CONF.command.project_id,
trustor_user_id=CONF.command.trustor_user_id,
trustee_user_id=CONF.command.trustee_user_id,
date=CONF.command.date
)
class MappingPurge(BaseApp):
"""Purge the mapping table."""
name = 'mapping_purge'
@classmethod
def add_argument_parser(cls, subparsers):
parser = super(MappingPurge, cls).add_argument_parser(subparsers)
parser.add_argument('--all', default=False, action='store_true',
help=('Purge all mappings.'))
parser.add_argument('--domain-name', default=None,
help=('Purge any mappings for the domain '
'specified.'))
parser.add_argument('--public-id', default=None,
help=('Purge the mapping for the Public ID '
'specified.'))
parser.add_argument('--local-id', default=None,
help=('Purge the mappings for the Local ID '
'specified.'))
parser.add_argument('--type', default=None, choices=['user', 'group'],
help=('Purge any mappings for the type '
'specified.'))
return parser
@staticmethod
def main():
def validate_options():
# NOTE(henry-nash): It would be nice to use the argparse automated
# checking for this validation, but the only way I can see doing
# that is to make the default (i.e. if no optional parameters
# are specified) to purge all mappings - and that sounds too
# dangerous as a default. So we use it in a slightly
# unconventional way, where all parameters are optional, but you
# must specify at least one.
if (CONF.command.all is False and
CONF.command.domain_name is None and
CONF.command.public_id is None and
CONF.command.local_id is None and
CONF.command.type is None):
raise ValueError(_('At least one option must be provided'))
if (CONF.command.all is True and
(CONF.command.domain_name is not None or
CONF.command.public_id is not None or
CONF.command.local_id is not None or
CONF.command.type is not None)):
raise ValueError(_('--all option cannot be mixed with '
'other options'))
def get_domain_id(name):
try:
return resource_manager.get_domain_by_name(name)['id']
except KeyError:
raise ValueError(_("Unknown domain '%(name)s' specified by "
"--domain-name") % {'name': name})
validate_options()
drivers = backends.load_backends()
resource_manager = drivers['resource_api']
mapping_manager = drivers['id_mapping_api']
# Now that we have validated the options, we know that at least one
# option has been specified, and if it was the --all option then this
# was the only option specified.
#
# The mapping dict is used to filter which mappings are purged, so
# leaving it empty means purge them all
mapping = {}
if CONF.command.domain_name is not None:
mapping['domain_id'] = get_domain_id(CONF.command.domain_name)
if CONF.command.public_id is not None:
mapping['public_id'] = CONF.command.public_id
if CONF.command.local_id is not None:
mapping['local_id'] = CONF.command.local_id
if CONF.command.type is not None:
mapping['entity_type'] = CONF.command.type
mapping_manager.purge_mappings(mapping)
DOMAIN_CONF_FHEAD = 'keystone.'
DOMAIN_CONF_FTAIL = '.conf'
def _domain_config_finder(conf_dir):
"""Return a generator of all domain config files found in a directory.
Domain configs match the filename pattern of
'keystone.<domain_name>.conf'.
:returns: generator yielding (filename, domain_name) tuples
"""
LOG.info('Scanning %r for domain config files', conf_dir)
for r, d, f in os.walk(conf_dir):
for fname in f:
if (fname.startswith(DOMAIN_CONF_FHEAD) and
fname.endswith(DOMAIN_CONF_FTAIL)):
if fname.count('.') >= 2:
domain_name = fname[len(DOMAIN_CONF_FHEAD):
-len(DOMAIN_CONF_FTAIL)]
yield (os.path.join(r, fname), domain_name)
continue
LOG.warning('Ignoring file (%s) while scanning '
'domain config directory', fname)
class DomainConfigUploadFiles(object):
def __init__(self, domain_config_finder=_domain_config_finder):
super(DomainConfigUploadFiles, self).__init__()
self.load_backends()
self._domain_config_finder = domain_config_finder
def load_backends(self):
drivers = backends.load_backends()
self.resource_manager = drivers['resource_api']
self.domain_config_manager = drivers['domain_config_api']
def valid_options(self):
"""Validate the options, returning True if they are indeed valid.
It would be nice to use the argparse automated checking for this
validation, but the only way I can see doing that is to make the
default (i.e. if no optional parameters are specified) to upload
all configuration files - and that sounds too dangerous as a
default. So we use it in a slightly unconventional way, where all
parameters are optional, but you must specify at least one.
"""
if (CONF.command.all is False and
CONF.command.domain_name is None):
print(_('At least one option must be provided, use either '
'--all or --domain-name'))
return False
if (CONF.command.all is True and
CONF.command.domain_name is not None):
print(_('The --all option cannot be used with '
'the --domain-name option'))
return False
return True
def _upload_config_to_database(self, file_name, domain_name):
"""Upload a single config file to the database.
:param file_name: the file containing the config options
:param domain_name: the domain name
:returns: a boolean indicating if the upload succeeded
"""
try:
domain_ref = (
self.resource_manager.get_domain_by_name(domain_name))
except exception.DomainNotFound:
print(_('Invalid domain name: %(domain)s found in config file '
'name: %(file)s - ignoring this file.') % {
'domain': domain_name,
'file': file_name})
return False
if self.domain_config_manager.get_config_with_sensitive_info(
domain_ref['id']):
print(_('Domain: %(domain)s already has a configuration '
'defined - ignoring file: %(file)s.') % {
'domain': domain_name,
'file': file_name})
return False
sections = {}
try:
parser = cfg.ConfigParser(file_name, sections)
parser.parse()
except Exception:
# We explicitly don't try and differentiate the error cases, in
# order to keep the code in this tool more robust as oslo.config
# changes.
print(_('Error parsing configuration file for domain: %(domain)s, '
'file: %(file)s.') % {
'domain': domain_name,
'file': file_name})
return False
try:
for group in sections:
for option in sections[group]:
sections[group][option] = sections[group][option][0]
self.domain_config_manager.create_config(domain_ref['id'],
sections)
return True
except Exception as e:
msg = ('Error processing config file for domain: '
'%(domain_name)s, file: %(filename)s, error: %(error)s')
LOG.error(msg,
{'domain_name': domain_name,
'filename': file_name,
'error': e},
exc_info=True)
return False
def read_domain_configs_from_files(self):
"""Read configs from file(s) and load into database.
The command line parameters have already been parsed and the CONF
command option will have been set. It is either set to the name of an
explicit domain, or it's None to indicate that we want all domain
config files.
"""
domain_name = CONF.command.domain_name
conf_dir = CONF.identity.domain_config_dir
if not os.path.exists(conf_dir):
print(_('Unable to locate domain config directory: %s') % conf_dir)
raise ValueError
if domain_name:
# Request is to upload the configs for just one domain
fname = DOMAIN_CONF_FHEAD + domain_name + DOMAIN_CONF_FTAIL
if not self._upload_config_to_database(
os.path.join(conf_dir, fname), domain_name):
return False
return True
success_cnt = 0
failure_cnt = 0
for filename, domain_name in self._domain_config_finder(conf_dir):
if self._upload_config_to_database(filename, domain_name):
success_cnt += 1
LOG.info('Successfully uploaded domain config %r',
filename)
else:
failure_cnt += 1
if success_cnt == 0:
LOG.warning('No domain configs uploaded from %r', conf_dir)
if failure_cnt:
return False
return True
def run(self):
# First off, let's just check we can talk to the domain database
try:
self.resource_manager.list_domains(driver_hints.Hints())
except Exception:
# It is likely that there is some SQL or other backend error
# related to set up
print(_('Unable to access the keystone database, please check it '
'is configured correctly.'))
raise
if not self.valid_options():
return 1
if not self.read_domain_configs_from_files():
return 1
class DomainConfigUpload(BaseApp):
"""Upload the domain specific configuration files to the database."""
name = 'domain_config_upload'
@classmethod
def add_argument_parser(cls, subparsers):
parser = super(DomainConfigUpload, cls).add_argument_parser(subparsers)
parser.add_argument('--all', default=False, action='store_true',
help='Upload contents of all domain specific '
'configuration files. Either use this option '
'or use the --domain-name option to choose a '
'specific domain.')
parser.add_argument('--domain-name', default=None,
help='Upload contents of the specific '
'configuration file for the given domain. '
'Either use this option or use the --all '
'option to upload contents for all domains.')
return parser
@staticmethod
def main():
dcu = DomainConfigUploadFiles()
status = dcu.run()
if status is not None:
sys.exit(status)
class SamlIdentityProviderMetadata(BaseApp):
"""Generate Identity Provider metadata."""
name = 'saml_idp_metadata'
@staticmethod
def main():
metadata = idp.MetadataGenerator().generate_metadata()
print(metadata)
class MappingEngineTester(BaseApp):
"""Execute mapping engine locally."""
name = 'mapping_engine'
def __init__(self):
super(MappingEngineTester, self).__init__()
self.mapping_id = uuid.uuid4().hex
self.rules_pathname = None
self.rules = None
self.assertion_pathname = None
self.assertion = None
def read_rules(self, path):
self.rules_pathname = path
try:
with open(path, "rb") as file:
self.rules = jsonutils.load(file)
except ValueError as e:
raise SystemExit(_('Error while parsing rules '
'%(path)s: %(err)s') % {'path': path, 'err': e})
def read_assertion(self, path):
self.assertion_pathname = path
try:
with open(path) as file:
self.assertion = file.read().strip()
except IOError as e:
raise SystemExit(_("Error while opening file "
"%(path)s: %(err)s") % {'path': path, 'err': e})
def normalize_assertion(self):
def split(line, line_num):
try:
k, v = line.split(':', 1)
return k.strip(), v.strip()
except ValueError:
msg = _("assertion file %(pathname)s at line %(line_num)d "
"expected 'key: value' but found '%(line)s' "
"see help for file format")
raise SystemExit(msg % {'pathname': self.assertion_pathname,
'line_num': line_num,
'line': line})
assertion = self.assertion.splitlines()
assertion_dict = {}
prefix = CONF.command.prefix
for line_num, line in enumerate(assertion, 1):
line = line.strip()
if line == '':
continue
k, v = split(line, line_num)
if prefix:
if k.startswith(prefix):
assertion_dict[k] = v
else:
assertion_dict[k] = v
self.assertion = assertion_dict
def normalize_rules(self):
if isinstance(self.rules, list):
self.rules = {'rules': self.rules}
@classmethod
def main(cls):
if CONF.command.engine_debug:
mapping_engine.LOG.logger.setLevel('DEBUG')
else:
mapping_engine.LOG.logger.setLevel('WARN')
tester = cls()
tester.read_rules(CONF.command.rules)
tester.normalize_rules()
mapping_engine.validate_mapping_structure(tester.rules)
tester.read_assertion(CONF.command.input)
tester.normalize_assertion()
if CONF.command.engine_debug:
print("Using Rules:\n%s" % (
jsonutils.dumps(tester.rules, indent=2)))
print("Using Assertion:\n%s" % (
jsonutils.dumps(tester.assertion, indent=2)))
rp = mapping_engine.RuleProcessor(tester.mapping_id,
tester.rules['rules'])
mapped = rp.process(tester.assertion)
print(jsonutils.dumps(mapped, indent=2))
@classmethod
def add_argument_parser(cls, subparsers):
parser = super(MappingEngineTester,
cls).add_argument_parser(subparsers)
parser.formatter_class = argparse.RawTextHelpFormatter
parser.add_argument('--rules', default=None, required=True,
help=("Path to the file with "
"rules to be executed. "
"Content must be\na proper JSON structure, "
"with a top-level key 'rules' and\n"
"corresponding value being a list."))
parser.add_argument('--input', default=None, required=True,
help=("Path to the file with input attributes. "
"The content\nconsists of ':' separated "
"parameter names and their values.\nThere "
"is only one key-value pair per line. "
"A ';' in the\nvalue is a separator and "
"then a value is treated as a list.\n"
"Example:\n"
"\tEMAIL: me@example.com\n"
"\tLOGIN: me\n"
"\tGROUPS: group1;group2;group3"))
parser.add_argument('--prefix', default=None,
help=("A prefix used for each environment "
"variable in the\nassertion. For example, "
"all environment variables may have\nthe "
"prefix ASDF_."))
parser.add_argument('--engine-debug',
default=False, action="store_true",
help=("Enable debug messages from the mapping "
"engine."))
class MappingPopulate(BaseApp):
"""Pre-populate entries from domain-specific backends.
Running this command is not required. It should only be run right after
the LDAP was configured, when many new users were added, or when
"mapping_purge" is run.
This command will take a while to run. It is perfectly fine for it to run
more than several minutes.
"""
name = "mapping_populate"
@classmethod
def load_backends(cls):
drivers = backends.load_backends()
cls.identity_api = drivers['identity_api']
cls.resource_api = drivers['resource_api']
@classmethod
def add_argument_parser(cls, subparsers):
parser = super(MappingPopulate, cls).add_argument_parser(
subparsers)
parser.add_argument('--domain-name', default=None, required=True,
help=("Name of the domain configured to use "
"domain-specific backend"))
return parser
@classmethod
def main(cls):
"""Process entries for id_mapping_api."""
cls.load_backends()
domain_name = CONF.command.domain_name
try:
domain_id = cls.resource_api.get_domain_by_name(domain_name)['id']
except exception.DomainNotFound:
print(_('Invalid domain name: %(domain)s') % {
'domain': domain_name})
return False
# We don't actually need to tackle id_mapping_api in order to get
# entries there, because list_users does this anyway. That's why it
# will be enough to just make the call below.
cls.identity_api.list_users(domain_scope=domain_id)
CMDS = [
BootStrap,
CredentialMigrate,
CredentialRotate,
CredentialSetup,
DbSync,
DbVersion,
Doctor,
DomainConfigUpload,
FernetRotate,
FernetSetup,
CreateJWSKeyPair,
MappingPopulate,
MappingPurge,
MappingEngineTester,
ReceiptRotate,
ReceiptSetup,
SamlIdentityProviderMetadata,
TokenRotate,
TokenSetup,
TrustFlush
]
def add_command_parsers(subparsers):
for cmd in CMDS:
cmd.add_argument_parser(subparsers)
command_opt = cfg.SubCommandOpt('command',
title='Commands',
help='Available commands',
handler=add_command_parsers)
def main(argv=None, developer_config_file=None):
"""Main entry point into the keystone-manage CLI utility.
:param argv: Arguments supplied via the command line using the ``sys``
standard library.
:type argv: list
:param developer_config_file: The location of a configuration file normally
found in development environments.
:type developer_config_file: string
"""
CONF.register_cli_opt(command_opt)
keystone.conf.configure()
sql.initialize()
keystone.conf.set_default_for_default_log_levels()
user_supplied_config_file = False
if argv:
for argument in argv:
if argument == '--config-file':
user_supplied_config_file = True
if developer_config_file:
developer_config_file = [developer_config_file]
# NOTE(lbragstad): At this point in processing, the first element of argv
# is the binary location of keystone-manage, which oslo.config doesn't need
# and is keystone specific. Only pass a list of arguments so that
# oslo.config can determine configuration file locations based on user
# provided arguments, if present.
CONF(args=argv[1:],
project='keystone',
version=pbr.version.VersionInfo('keystone').version_string(),
usage='%(prog)s [' + '|'.join([cmd.name for cmd in CMDS]) + ']',
default_config_files=developer_config_file)
if not CONF.default_config_files and not user_supplied_config_file:
LOG.warning('Config file not found, using default configs.')
keystone.conf.setup_logging()
CONF.command.cmd_class.main()
View Git Blame Copy Permalink