openstack
/
oslo.messaging
Code Issues Proposed changes

Replace retrying with tenacity 

This patch replaces the legacy retrying library with the newer
and more convenient tenacity one, taking into account that:
1) retrying uses milliseconds for wait times, but tenacity uses seconds;
2) retrying has a lot of numeric arguments for specifying behaviour
of decorated functions, while tenacity has a few of them, which are
specialized objects, thus making the retry-decorator more flexible.

Change-Id: Ib6ecffe5d1cf292badbb9eb6db6260f17460f343
Closes-Bug: #1635399
This commit is contained in:
Gevorg Davoian 2016-10-21 17:18:55 +03:00
parent b1b3677a75
commit 90125aa885
6 changed files with 74 additions and 62 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

33
oslo_messaging/_drivers/impl_pika.py
View File

@ -16,7 +16,7 @@ from oslo_config import cfg
from oslo_log import log as logging
from oslo_utils import timeutils
import pika_pool
import retrying
import tenacity
from oslo_messaging._drivers import base
from oslo_messaging._drivers.pika_driver import (pika_connection_factory as
@ -201,14 +201,15 @@ class PikaDriver(base.BaseDriver):
else:
return False
retrier = (
None if retry == 0 else
retrying.retry(
stop_max_attempt_number=(None if retry == -1 else retry),
retry_on_exception=on_exception,
wait_fixed=self._pika_engine.rpc_retry_delay * 1000,
if retry:
retrier = tenacity.retry(
stop=(tenacity.stop_never if retry == -1 else
tenacity.stop_after_attempt(retry)),
retry=tenacity.retry_if_exception(on_exception),
wait=tenacity.wait_fixed(self._pika_engine.rpc_retry_delay)
)
)
else:
retrier = None
if target.fanout:
return self.cast_all_workers(
@ -312,11 +313,17 @@ class PikaDriver(base.BaseDriver):
else:
return False
retrier = retrying.retry(
stop_max_attempt_number=(None if retry == -1 else retry),
retry_on_exception=on_exception,
wait_fixed=self._pika_engine.notification_retry_delay * 1000,
)
if retry:
retrier = tenacity.retry(
stop=(tenacity.stop_never if retry == -1 else
tenacity.stop_after_attempt(retry)),
retry=tenacity.retry_if_exception(on_exception),
wait=tenacity.wait_fixed(
self._pika_engine.notification_retry_delay
)
)
else:
retrier = None
msg = pika_drv_msg.PikaOutgoingMessage(self._pika_engine, message,
ctxt)

42
oslo_messaging/_drivers/pika_driver/pika_message.py
View File

@ -25,8 +25,8 @@ from oslo_utils import timeutils
from pika import exceptions as pika_exceptions
from pika import spec as pika_spec
import pika_pool
import retrying
import six
import tenacity
import oslo_messaging
@ -201,14 +201,22 @@ class RpcPikaIncomingMessage(PikaIncomingMessage, base.RpcIncomingMessage):
else:
return False
retrier = retrying.retry(
stop_max_attempt_number=(
None if self._pika_engine.rpc_reply_retry_attempts == -1
else self._pika_engine.rpc_reply_retry_attempts
),
retry_on_exception=on_exception,
wait_fixed=self._pika_engine.rpc_reply_retry_delay * 1000,
) if self._pika_engine.rpc_reply_retry_attempts else None
if self._pika_engine.rpc_reply_retry_attempts:
retrier = tenacity.retry(
stop=(
tenacity.stop_never
if self._pika_engine.rpc_reply_retry_attempts == -1 else
tenacity.stop_after_attempt(
self._pika_engine.rpc_reply_retry_attempts
)
),
retry=tenacity.retry_if_exception(on_exception),
wait=tenacity.wait_fixed(
self._pika_engine.rpc_reply_retry_delay
)
)
else:
retrier = None
try:
timeout = (None if self.expiration_time is None else
@ -438,8 +446,8 @@ class PikaOutgoingMessage(object):
for routing into durable queues
:param stopwatch: StopWatch, stopwatch object for calculating
allowed timeouts
:param retrier: retrying.Retrier, configured retrier object for sending
message, if None no retrying is performed
:param retrier: tenacity.Retrying, configured retrier object for
sending message, if None no retrying is performed
"""
msg_props.delivery_mode = 2 if persistent else 1
@ -475,8 +483,8 @@ class PikaOutgoingMessage(object):
for routing into durable queues
:param stopwatch: StopWatch, stopwatch object for calculating
allowed timeouts
:param retrier: retrying.Retrier, configured retrier object for sending
message, if None no retrying is performed
:param retrier: tenacity.Retrying, configured retrier object for
sending message, if None no retrying is performed
"""
msg_dict, msg_props = self._prepare_message_to_send()
@ -506,8 +514,8 @@ class RpcPikaOutgoingMessage(PikaOutgoingMessage):
reply. If None - return immediately without reply waiting
:param stopwatch: StopWatch, stopwatch object for calculating
allowed timeouts
:param retrier: retrying.Retrier, configured retrier object for sending
message, if None no retrying is performed
:param retrier: tenacity.Retrying, configured retrier object for
sending message, if None no retrying is performed
"""
msg_dict, msg_props = self._prepare_message_to_send()
@ -595,8 +603,8 @@ class RpcReplyPikaOutgoingMessage(PikaOutgoingMessage):
:param reply_q: String, queue name for sending reply
:param stopwatch: StopWatch, stopwatch object for calculating
allowed timeouts
:param retrier: retrying.Retrier, configured retrier object for sending
message, if None no retrying is performed
:param retrier: tenacity.Retrying, configured retrier object for
sending message, if None no retrying is performed
"""
msg_dict, msg_props = self._prepare_message_to_send()

12
oslo_messaging/_drivers/zmq_driver/client/zmq_publisher_manager.py
View File

@ -16,8 +16,8 @@ import abc
import contextlib
import logging
import retrying
import six
import tenacity
from oslo_messaging._drivers.zmq_driver.matchmaker import zmq_matchmaker_base
from oslo_messaging._drivers.zmq_driver import zmq_async
@ -30,9 +30,9 @@ zmq = zmq_async.import_zmq()
def _drop_message_warn(request):
LOG.warning(_LW("Matchmaker contains no records for specified "
"target %(target)s. Dropping message %(msg_id)s.")
% {"target": request.target,
"msg_id": request.message_id})
"target %(target)s. Dropping message %(msg_id)s."),
{"target": request.target,
"msg_id": request.message_id})
def target_not_found_warn(func):
@ -40,7 +40,7 @@ def target_not_found_warn(func):
try:
return func(self, request, *args, **kwargs)
except (zmq_matchmaker_base.MatchmakerUnavailable,
retrying.RetryError):
tenacity.RetryError):
_drop_message_warn(request)
return _target_not_found_warn
@ -50,7 +50,7 @@ def target_not_found_timeout(func):
try:
return func(self, request, *args, **kwargs)
except (zmq_matchmaker_base.MatchmakerUnavailable,
retrying.RetryError):
tenacity.RetryError):
_drop_message_warn(request)
self.publisher._raise_timeout(request)
return _target_not_found_timeout

32
oslo_messaging/_drivers/zmq_driver/matchmaker/zmq_matchmaker_redis.py
View File

@ -20,8 +20,8 @@ import time
from oslo_config import cfg
from oslo_utils import importutils
from retrying import retry
import six
import tenacity
from oslo_messaging._drivers.zmq_driver.matchmaker import zmq_matchmaker_base
from oslo_messaging._drivers.zmq_driver import zmq_address
@ -30,6 +30,7 @@ from oslo_messaging._i18n import _LE, _LI, _LW
redis = importutils.try_import('redis')
redis_sentinel = importutils.try_import('redis.sentinel')
LOG = logging.getLogger(__name__)
@ -54,8 +55,8 @@ matchmaker_redis_opts = [
default=[],
deprecated_for_removal=True,
deprecated_reason="Replaced by [DEFAULT]/transport_url",
help='List of Redis Sentinel hosts (fault tolerance mode) e.g.\
[host:port, host1:port ... ]'),
help='List of Redis Sentinel hosts (fault tolerance mode), '
'e.g., [host:port, host1:port ... ]'),
cfg.StrOpt('sentinel_group_name',
default='oslo-messaging-zeromq',
help='Redis replica set name.'),
@ -67,7 +68,7 @@ matchmaker_redis_opts = [
help='Time in ms to wait before the transaction is killed.'),
cfg.IntOpt('socket_timeout',
default=10000,
help='Timeout in ms on blocking socket operations'),
help='Timeout in ms on blocking socket operations.'),
]
_PUBLISHERS_KEY = "PUBLISHERS"
@ -132,11 +133,7 @@ def empty_list_on_error(func):
return func_wrapper
def retry_if_connection_error(ex):
return isinstance(ex, zmq_matchmaker_base.MatchmakerUnavailable)
def retry_if_empty(hosts):
def is_empty(hosts):
return not hosts
@ -239,12 +236,15 @@ class MatchmakerRedisBase(zmq_matchmaker_base.MatchmakerBase):
return self._retry_method(target, listener_type, self.get_hosts_fanout)
def _retry_method(self, target, listener_type, method):
@retry(retry_on_result=retry_if_empty,
wrap_exception=True,
wait_fixed=self.conf.matchmaker_redis.wait_timeout,
stop_max_delay=self.conf.matchmaker_redis.check_timeout)
wait_timeout = self.conf.matchmaker_redis.wait_timeout / 1000.
check_timeout = self.conf.matchmaker_redis.check_timeout / 1000.
@tenacity.retry(retry=tenacity.retry_if_result(is_empty),
wait=tenacity.wait_fixed(wait_timeout),
stop=tenacity.stop_after_delay(check_timeout))
def _get_hosts_retry(target, listener_type):
return method(target, listener_type)
return _get_hosts_retry(target, listener_type)
@ -362,15 +362,15 @@ class MatchmakerSentinel(MatchmakerRedisBase):
def __init__(self, conf, *args, **kwargs):
super(MatchmakerSentinel, self).__init__(conf, *args, **kwargs)
socket_timeout = self.conf.matchmaker_redis.socket_timeout / 1000.
self._sentinel_hosts, password, master_group = \
self._extract_sentinel_hosts()
self._sentinel = redis_sentinel.Sentinel(
sentinels=self._sentinel_hosts,
socket_timeout=socket_timeout,
password=password)
socket_timeout=self.conf.matchmaker_redis.socket_timeout / 1000.,
password=password
)
self._redis_master = self._sentinel.master_for(master_group)
self._redis_slave = self._sentinel.slave_for(master_group)

11
oslo_messaging/tests/drivers/pika/test_message.py
View File

@ -11,6 +11,7 @@
# 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 functools
import unittest
@ -172,7 +173,7 @@ class RpcPikaIncomingMessageTestCase(unittest.TestCase):
@patch("oslo_messaging._drivers.pika_driver.pika_message."
"RpcReplyPikaOutgoingMessage")
@patch("retrying.retry")
@patch("tenacity.retry")
def test_positive_reply_for_call_message(self,
retry_mock,
outgoing_message_mock):
@ -202,13 +203,12 @@ class RpcPikaIncomingMessageTestCase(unittest.TestCase):
reply_q='reply_queue', stopwatch=mock.ANY, retrier=mock.ANY
)
retry_mock.assert_called_once_with(
retry_on_exception=mock.ANY, stop_max_attempt_number=3,
wait_fixed=250.0
stop=mock.ANY, retry=mock.ANY, wait=mock.ANY
)
@patch("oslo_messaging._drivers.pika_driver.pika_message."
"RpcReplyPikaOutgoingMessage")
@patch("retrying.retry")
@patch("tenacity.retry")
def test_negative_reply_for_call_message(self,
retry_mock,
outgoing_message_mock):
@ -241,8 +241,7 @@ class RpcPikaIncomingMessageTestCase(unittest.TestCase):
reply_q='reply_queue', stopwatch=mock.ANY, retrier=mock.ANY
)
retry_mock.assert_called_once_with(
retry_on_exception=mock.ANY, stop_max_attempt_number=3,
wait_fixed=250.0
stop=mock.ANY, retry=mock.ANY, wait=mock.ANY
)

6
requirements.txt
View File

@ -20,9 +20,7 @@ monotonic>=0.6 # Apache-2.0
six>=1.9.0 # MIT
cachetools>=1.1.0 # MIT License
# FIXME(markmc): remove this when the drivers no longer
# import eventlet
# FIXME(markmc): remove this when the drivers no longer import eventlet
eventlet!=0.18.3,>=0.18.2 # MIT
greenlet>=0.3.2 # MIT
@ -41,7 +39,7 @@ pika-pool>=0.1.3 # BSD
# used by pika and zmq drivers
futures>=3.0;python_version=='2.7' or python_version=='2.6' # BSD
retrying!=1.3.0,>=1.2.3 # Apache-2.0
tenacity>=3.2.1 # Apache-2.0
# middleware
oslo.middleware>=3.0.0 # Apache-2.0