Load Balancing as a Service (LBaaS) for OpenStack
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 

600 lines
26 KiB

# Copyright 2015 Hewlett-Packard Development Company, L.P.
#
# 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.
#
from cryptography import fernet
from oslo_config import cfg
from oslo_log import log as logging
from stevedore import driver as stevedore_driver
from taskflow import retry
from taskflow import task
from taskflow.types import failure
from octavia.amphorae.backends.agent import agent_jinja_cfg
from octavia.amphorae.driver_exceptions import exceptions as driver_except
from octavia.common import constants
from octavia.common import utils
from octavia.controller.worker import task_utils as task_utilities
from octavia.db import api as db_apis
from octavia.db import repositories as repo
from octavia.network import data_models
CONF = cfg.CONF
LOG = logging.getLogger(__name__)
class BaseAmphoraTask(task.Task):
"""Base task to load drivers common to the tasks."""
def __init__(self, **kwargs):
super().__init__(**kwargs)
self.amphora_driver = stevedore_driver.DriverManager(
namespace='octavia.amphora.drivers',
name=CONF.controller_worker.amphora_driver,
invoke_on_load=True
).driver
self.amphora_repo = repo.AmphoraRepository()
self.listener_repo = repo.ListenerRepository()
self.loadbalancer_repo = repo.LoadBalancerRepository()
self.task_utils = task_utilities.TaskUtils()
class AmpRetry(retry.Times):
def on_failure(self, history, *args, **kwargs):
last_errors = history[-1][1]
max_retry_attempt = CONF.haproxy_amphora.connection_max_retries
for task_name, ex_info in last_errors.items():
if len(history) <= max_retry_attempt:
# When taskflow persistance is enabled and flow/task state is
# saved in the backend. If flow(task) is restored(restart of
# worker,etc) we are getting ex_info as None - we need to RETRY
# task to check its real state.
if ex_info is None or ex_info._exc_info is None:
return retry.RETRY
excp = ex_info._exc_info[1]
if isinstance(excp, driver_except.AmpConnectionRetry):
return retry.RETRY
return retry.REVERT_ALL
class AmpListenersUpdate(BaseAmphoraTask):
"""Task to update the listeners on one amphora."""
def execute(self, loadbalancer, amphora, timeout_dict=None):
# Note, we don't want this to cause a revert as it may be used
# in a failover flow with both amps failing. Skip it and let
# health manager fix it.
# TODO(johnsom) Optimize this to use the dicts and not need the
# DB lookups
db_amp = self.amphora_repo.get(db_apis.get_session(),
id=amphora[constants.ID])
try:
db_lb = self.loadbalancer_repo.get(
db_apis.get_session(),
id=loadbalancer[constants.LOADBALANCER_ID])
self.amphora_driver.update_amphora_listeners(
db_lb, db_amp, timeout_dict)
except Exception as e:
LOG.error('Failed to update listeners on amphora %s. Skipping '
'this amphora as it is failing to update due to: %s',
db_amp.id, str(e))
self.amphora_repo.update(db_apis.get_session(), db_amp.id,
status=constants.ERROR)
class AmphoraIndexListenerUpdate(BaseAmphoraTask):
"""Task to update the listeners on one amphora."""
def execute(self, loadbalancer, amphora_index, amphorae, timeout_dict=()):
# Note, we don't want this to cause a revert as it may be used
# in a failover flow with both amps failing. Skip it and let
# health manager fix it.
try:
# TODO(johnsom) Optimize this to use the dicts and not need the
# DB lookups
db_amp = self.amphora_repo.get(
db_apis.get_session(),
id=amphorae[amphora_index][constants.ID])
db_lb = self.loadbalancer_repo.get(
db_apis.get_session(),
id=loadbalancer[constants.LOADBALANCER_ID])
self.amphora_driver.update_amphora_listeners(
db_lb, db_amp, timeout_dict)
except Exception as e:
amphora_id = amphorae[amphora_index].get(constants.ID)
LOG.error('Failed to update listeners on amphora %s. Skipping '
'this amphora as it is failing to update due to: %s',
amphora_id, str(e))
self.amphora_repo.update(db_apis.get_session(), amphora_id,
status=constants.ERROR)
class ListenersUpdate(BaseAmphoraTask):
"""Task to update amphora with all specified listeners' configurations."""
def execute(self, loadbalancer_id):
"""Execute updates per listener for an amphora."""
loadbalancer = self.loadbalancer_repo.get(db_apis.get_session(),
id=loadbalancer_id)
if loadbalancer:
self.amphora_driver.update(loadbalancer)
else:
LOG.error('Load balancer %s for listeners update not found. '
'Skipping update.', loadbalancer_id)
def revert(self, loadbalancer_id, *args, **kwargs):
"""Handle failed listeners updates."""
LOG.warning("Reverting listeners updates.")
loadbalancer = self.loadbalancer_repo.get(db_apis.get_session(),
id=loadbalancer_id)
for listener in loadbalancer.listeners:
self.task_utils.mark_listener_prov_status_error(
listener.id)
class ListenersStart(BaseAmphoraTask):
"""Task to start all listeners on the vip."""
def execute(self, loadbalancer, amphora=None):
"""Execute listener start routines for listeners on an amphora."""
db_lb = self.loadbalancer_repo.get(
db_apis.get_session(), id=loadbalancer[constants.LOADBALANCER_ID])
if db_lb.listeners:
if amphora is not None:
db_amp = self.amphora_repo.get(db_apis.get_session(),
id=amphora[constants.ID])
else:
db_amp = amphora
self.amphora_driver.start(db_lb, db_amp)
LOG.debug("Started the listeners on the vip")
def revert(self, loadbalancer, *args, **kwargs):
"""Handle failed listeners starts."""
LOG.warning("Reverting listeners starts.")
db_lb = self.loadbalancer_repo.get(
db_apis.get_session(), id=loadbalancer[constants.LOADBALANCER_ID])
for listener in db_lb.listeners:
self.task_utils.mark_listener_prov_status_error(listener.id)
class AmphoraIndexListenersReload(BaseAmphoraTask):
"""Task to reload all listeners on an amphora."""
def execute(self, loadbalancer, amphora_index, amphorae,
timeout_dict=None):
"""Execute listener reload routines for listeners on an amphora."""
if amphorae is None:
return
# TODO(johnsom) Optimize this to use the dicts and not need the
# DB lookups
db_amp = self.amphora_repo.get(
db_apis.get_session(), id=amphorae[amphora_index][constants.ID])
db_lb = self.loadbalancer_repo.get(
db_apis.get_session(),
id=loadbalancer[constants.LOADBALANCER_ID])
if db_lb.listeners:
try:
self.amphora_driver.reload(db_lb, db_amp, timeout_dict)
except Exception as e:
amphora_id = amphorae[amphora_index].id
LOG.warning('Failed to reload listeners on amphora %s. '
'Skipping this amphora as it is failing to '
'reload due to: %s', amphora_id, str(e))
self.amphora_repo.update(db_apis.get_session(), amphora_id,
status=constants.ERROR)
class ListenerDelete(BaseAmphoraTask):
"""Task to delete the listener on the vip."""
def execute(self, listener):
"""Execute listener delete routines for an amphora."""
db_listener = self.listener_repo.get(
db_apis.get_session(), id=listener[constants.LISTENER_ID])
self.amphora_driver.delete(db_listener)
LOG.debug("Deleted the listener on the vip")
def revert(self, listener, *args, **kwargs):
"""Handle a failed listener delete."""
LOG.warning("Reverting listener delete.")
self.task_utils.mark_listener_prov_status_error(
listener[constants.LISTENER_ID])
class AmphoraGetInfo(BaseAmphoraTask):
"""Task to get information on an amphora."""
def execute(self, amphora):
"""Execute get_info routine for an amphora."""
db_amp = self.amphora_repo.get(db_apis.get_session(),
id=amphora[constants.ID])
self.amphora_driver.get_info(db_amp)
class AmphoraGetDiagnostics(BaseAmphoraTask):
"""Task to get diagnostics on the amphora and the loadbalancers."""
def execute(self, amphora):
"""Execute get_diagnostic routine for an amphora."""
self.amphora_driver.get_diagnostics(amphora)
class AmphoraFinalize(BaseAmphoraTask):
"""Task to finalize the amphora before any listeners are configured."""
def execute(self, amphora):
"""Execute finalize_amphora routine."""
db_amp = self.amphora_repo.get(db_apis.get_session(),
id=amphora.get(constants.ID))
self.amphora_driver.finalize_amphora(db_amp)
LOG.debug("Finalized the amphora.")
def revert(self, result, amphora, *args, **kwargs):
"""Handle a failed amphora finalize."""
if isinstance(result, failure.Failure):
return
LOG.warning("Reverting amphora finalize.")
self.task_utils.mark_amphora_status_error(
amphora.get(constants.ID))
class AmphoraPostNetworkPlug(BaseAmphoraTask):
"""Task to notify the amphora post network plug."""
def execute(self, amphora, ports):
"""Execute post_network_plug routine."""
db_amp = self.amphora_repo.get(db_apis.get_session(),
id=amphora[constants.ID])
for port in ports:
net = data_models.Network(**port.pop(constants.NETWORK))
ips = port.pop(constants.FIXED_IPS)
fixed_ips = []
for ip in ips:
subnet_arg = ip.pop(constants.SUBNET)
host_routes = subnet_arg.get('host_routes')
if host_routes:
subnet_arg['host_routes'] = [
data_models.HostRoute(**hr)
for hr in host_routes
]
fixed_ips.append(data_models.FixedIP(
subnet=data_models.Subnet(**subnet_arg), **ip))
self.amphora_driver.post_network_plug(
db_amp, data_models.Port(network=net, fixed_ips=fixed_ips,
**port))
LOG.debug("post_network_plug called on compute instance "
"%(compute_id)s for port %(port_id)s",
{"compute_id": amphora[constants.COMPUTE_ID],
"port_id": port[constants.ID]})
def revert(self, result, amphora, *args, **kwargs):
"""Handle a failed post network plug."""
if isinstance(result, failure.Failure):
return
LOG.warning("Reverting post network plug.")
self.task_utils.mark_amphora_status_error(amphora.get(constants.ID))
class AmphoraePostNetworkPlug(BaseAmphoraTask):
"""Task to notify the amphorae post network plug."""
def execute(self, loadbalancer, added_ports):
"""Execute post_network_plug routine."""
amp_post_plug = AmphoraPostNetworkPlug()
db_lb = self.loadbalancer_repo.get(
db_apis.get_session(), id=loadbalancer[constants.LOADBALANCER_ID])
for amphora in db_lb.amphorae:
if amphora.id in added_ports:
amp_post_plug.execute(amphora.to_dict(),
added_ports[amphora.id])
def revert(self, result, loadbalancer, added_ports, *args, **kwargs):
"""Handle a failed post network plug."""
if isinstance(result, failure.Failure):
return
db_lb = self.loadbalancer_repo.get(
db_apis.get_session(), id=loadbalancer[constants.LOADBALANCER_ID])
LOG.warning("Reverting post network plug.")
for amphora in filter(
lambda amp: amp.status == constants.AMPHORA_ALLOCATED,
db_lb.amphorae):
self.task_utils.mark_amphora_status_error(amphora.id)
class AmphoraPostVIPPlug(BaseAmphoraTask):
"""Task to notify the amphora post VIP plug."""
def execute(self, amphora, loadbalancer, amphorae_network_config):
"""Execute post_vip_routine."""
db_amp = self.amphora_repo.get(db_apis.get_session(),
id=amphora.get(constants.ID))
db_lb = self.loadbalancer_repo.get(
db_apis.get_session(), id=loadbalancer[constants.LOADBALANCER_ID])
vrrp_port = data_models.Port(
**amphorae_network_config[
amphora.get(constants.ID)][constants.VRRP_PORT])
# Required for noop-case
vip_arg = amphorae_network_config[amphora.get(
constants.ID)][constants.VIP_SUBNET]
if vip_arg:
host_routes = vip_arg.get('host_routes')
if host_routes:
vip_arg['host_routes'] = [
data_models.HostRoute(**hr)
for hr in host_routes
]
vip_subnet = data_models.Subnet(**vip_arg)
else:
vip_subnet = data_models.Subnet()
self.amphora_driver.post_vip_plug(
db_amp, db_lb, amphorae_network_config, vrrp_port=vrrp_port,
vip_subnet=vip_subnet)
LOG.debug("Notified amphora of vip plug")
def revert(self, result, amphora, loadbalancer, *args, **kwargs):
"""Handle a failed amphora vip plug notification."""
if isinstance(result, failure.Failure):
return
LOG.warning("Reverting post vip plug.")
self.task_utils.mark_amphora_status_error(amphora.get(constants.ID))
self.task_utils.mark_loadbalancer_prov_status_error(
loadbalancer[constants.LOADBALANCER_ID])
class AmphoraePostVIPPlug(BaseAmphoraTask):
"""Task to notify the amphorae post VIP plug."""
def execute(self, loadbalancer, amphorae_network_config):
"""Execute post_vip_plug across the amphorae."""
amp_post_vip_plug = AmphoraPostVIPPlug()
db_lb = self.loadbalancer_repo.get(
db_apis.get_session(), id=loadbalancer[constants.LOADBALANCER_ID])
for amphora in db_lb.amphorae:
amp_post_vip_plug.execute(amphora.to_dict(),
loadbalancer,
amphorae_network_config)
def revert(self, result, loadbalancer, *args, **kwargs):
"""Handle a failed amphora vip plug notification."""
if isinstance(result, failure.Failure):
return
LOG.warning("Reverting amphorae post vip plug.")
self.task_utils.mark_loadbalancer_prov_status_error(
loadbalancer[constants.LOADBALANCER_ID])
class AmphoraCertUpload(BaseAmphoraTask):
"""Upload a certificate to the amphora."""
def execute(self, amphora, server_pem):
"""Execute cert_update_amphora routine."""
LOG.debug("Upload cert in amphora REST driver")
key = utils.get_compatible_server_certs_key_passphrase()
fer = fernet.Fernet(key)
db_amp = self.amphora_repo.get(db_apis.get_session(),
id=amphora.get(constants.ID))
self.amphora_driver.upload_cert_amp(
db_amp, fer.decrypt(server_pem.encode('utf-8')))
# TODO(johnsom) REMOVE ME!
class AmphoraUpdateVRRPInterface(BaseAmphoraTask):
"""Task to get and update the VRRP interface device name from amphora."""
def execute(self, amphora, timeout_dict=None):
try:
# TODO(johnsom) Optimize this to use the dicts and not need the
# DB lookups
db_amp = self.amphora_repo.get(db_apis.get_session(),
id=amphora[constants.ID])
interface = self.amphora_driver.get_interface_from_ip(
db_amp, db_amp.vrrp_ip, timeout_dict=timeout_dict)
except Exception as e:
# This can occur when an active/standby LB has no listener
LOG.error('Failed to get amphora VRRP interface on amphora '
'%s. Skipping this amphora as it is failing due to: '
'%s', amphora.get(constants.ID), str(e))
self.amphora_repo.update(db_apis.get_session(),
amphora.get(constants.ID),
status=constants.ERROR)
return None
self.amphora_repo.update(db_apis.get_session(), amphora[constants.ID],
vrrp_interface=interface)
return interface
class AmphoraIndexUpdateVRRPInterface(BaseAmphoraTask):
"""Task to get and update the VRRP interface device name from amphora."""
def execute(self, amphora_index, amphorae, timeout_dict=None):
amphora_id = amphorae[amphora_index][constants.ID]
try:
# TODO(johnsom) Optimize this to use the dicts and not need the
# DB lookups
db_amp = self.amphora_repo.get(db_apis.get_session(),
id=amphora_id)
interface = self.amphora_driver.get_interface_from_ip(
db_amp, db_amp.vrrp_ip, timeout_dict=timeout_dict)
except Exception as e:
# This can occur when an active/standby LB has no listener
LOG.error('Failed to get amphora VRRP interface on amphora '
'%s. Skipping this amphora as it is failing due to: '
'%s', amphora_id, str(e))
self.amphora_repo.update(db_apis.get_session(), amphora_id,
status=constants.ERROR)
return None
self.amphora_repo.update(db_apis.get_session(), amphora_id,
vrrp_interface=interface)
return interface
class AmphoraVRRPUpdate(BaseAmphoraTask):
"""Task to update the VRRP configuration of an amphora."""
def execute(self, loadbalancer_id, amphorae_network_config, amphora,
amp_vrrp_int, timeout_dict=None):
"""Execute update_vrrp_conf."""
# Note, we don't want this to cause a revert as it may be used
# in a failover flow with both amps failing. Skip it and let
# health manager fix it.
amphora_id = amphora[constants.ID]
try:
# TODO(johnsom) Optimize this to use the dicts and not need the
# DB lookups
db_amp = self.amphora_repo.get(db_apis.get_session(),
id=amphora_id)
loadbalancer = self.loadbalancer_repo.get(db_apis.get_session(),
id=loadbalancer_id)
db_amp.vrrp_interface = amp_vrrp_int
self.amphora_driver.update_vrrp_conf(
loadbalancer, amphorae_network_config, db_amp, timeout_dict)
except Exception as e:
LOG.error('Failed to update VRRP configuration amphora %s. '
'Skipping this amphora as it is failing to update due '
'to: %s', amphora_id, str(e))
self.amphora_repo.update(db_apis.get_session(), amphora_id,
status=constants.ERROR)
LOG.debug("Uploaded VRRP configuration of amphora %s.", amphora_id)
class AmphoraIndexVRRPUpdate(BaseAmphoraTask):
"""Task to update the VRRP configuration of an amphora."""
def execute(self, loadbalancer_id, amphorae_network_config, amphora_index,
amphorae, amp_vrrp_int, timeout_dict=None):
"""Execute update_vrrp_conf."""
# Note, we don't want this to cause a revert as it may be used
# in a failover flow with both amps failing. Skip it and let
# health manager fix it.
amphora_id = amphorae[amphora_index][constants.ID]
try:
# TODO(johnsom) Optimize this to use the dicts and not need the
# DB lookups
db_amp = self.amphora_repo.get(db_apis.get_session(),
id=amphora_id)
loadbalancer = self.loadbalancer_repo.get(db_apis.get_session(),
id=loadbalancer_id)
db_amp.vrrp_interface = amp_vrrp_int
self.amphora_driver.update_vrrp_conf(
loadbalancer, amphorae_network_config, db_amp, timeout_dict)
except Exception as e:
LOG.error('Failed to update VRRP configuration amphora %s. '
'Skipping this amphora as it is failing to update due '
'to: %s', amphora_id, str(e))
self.amphora_repo.update(db_apis.get_session(), amphora_id,
status=constants.ERROR)
return
LOG.debug("Uploaded VRRP configuration of amphora %s.", amphora_id)
class AmphoraVRRPStart(BaseAmphoraTask):
"""Task to start keepalived on an amphora.
This will reload keepalived if it is already running.
"""
def execute(self, amphora, timeout_dict=None):
# TODO(johnsom) Optimize this to use the dicts and not need the
# DB lookups
db_amp = self.amphora_repo.get(
db_apis.get_session(), id=amphora[constants.ID])
self.amphora_driver.start_vrrp_service(db_amp, timeout_dict)
LOG.debug("Started VRRP on amphora %s.", amphora[constants.ID])
class AmphoraIndexVRRPStart(BaseAmphoraTask):
"""Task to start keepalived on an amphora.
This will reload keepalived if it is already running.
"""
def execute(self, amphora_index, amphorae, timeout_dict=None):
# TODO(johnsom) Optimize this to use the dicts and not need the
# DB lookups
amphora_id = amphorae[amphora_index][constants.ID]
db_amp = self.amphora_repo.get(db_apis.get_session(), id=amphora_id)
try:
self.amphora_driver.start_vrrp_service(db_amp, timeout_dict)
except Exception as e:
LOG.error('Failed to start VRRP on amphora %s. '
'Skipping this amphora as it is failing to start due '
'to: %s', amphora_id, str(e))
self.amphora_repo.update(db_apis.get_session(), amphora_id,
status=constants.ERROR)
return
LOG.debug("Started VRRP on amphora %s.",
amphorae[amphora_index][constants.ID])
class AmphoraComputeConnectivityWait(BaseAmphoraTask):
"""Task to wait for the compute instance to be up."""
def execute(self, amphora, raise_retry_exception=False):
"""Execute get_info routine for an amphora until it responds."""
try:
db_amphora = self.amphora_repo.get(
db_apis.get_session(), id=amphora.get(constants.ID))
amp_info = self.amphora_driver.get_info(
db_amphora, raise_retry_exception=raise_retry_exception)
LOG.debug('Successfuly connected to amphora %s: %s',
amphora.get(constants.ID), amp_info)
except driver_except.TimeOutException:
LOG.error("Amphora compute instance failed to become reachable. "
"This either means the compute driver failed to fully "
"boot the instance inside the timeout interval or the "
"instance is not reachable via the lb-mgmt-net.")
self.amphora_repo.update(db_apis.get_session(),
amphora.get(constants.ID),
status=constants.ERROR)
raise
class AmphoraConfigUpdate(BaseAmphoraTask):
"""Task to push a new amphora agent configuration to the amphora."""
def execute(self, amphora, flavor):
# Extract any flavor based settings
if flavor:
topology = flavor.get(constants.LOADBALANCER_TOPOLOGY,
CONF.controller_worker.loadbalancer_topology)
else:
topology = CONF.controller_worker.loadbalancer_topology
# Build the amphora agent config
agent_cfg_tmpl = agent_jinja_cfg.AgentJinjaTemplater()
agent_config = agent_cfg_tmpl.build_agent_config(
amphora.get(constants.ID), topology)
db_amp = self.amphora_repo.get(db_apis.get_session(),
id=amphora[constants.ID])
# Push the new configuration to the amphora
try:
self.amphora_driver.update_amphora_agent_config(db_amp,
agent_config)
except driver_except.AmpDriverNotImplementedError:
LOG.error('Amphora {} does not support agent configuration '
'update. Please update the amphora image for this '
'amphora. Skipping.'.
format(amphora.get(constants.ID)))