node resource consolidation
This strategy is used to centralize VMs to as few nodes as possible by VM migration. User can set a input parameter to decide how to select the destination node. Implements: blueprint node-resource-consolidation Closes-Bug: #1843016 Change-Id: I104c864d532c2092f5dc6f0c8f756ebeae12f09e
This commit is contained in:
licanwei
@@ -20,6 +20,8 @@ from watcher.decision_engine.strategy.strategies import basic_consolidation
|
||||
from watcher.decision_engine.strategy.strategies import dummy_strategy
|
||||
from watcher.decision_engine.strategy.strategies import dummy_with_scorer
|
||||
from watcher.decision_engine.strategy.strategies import host_maintenance
|
||||
from watcher.decision_engine.strategy.strategies import \
|
||||
node_resource_consolidation
|
||||
from watcher.decision_engine.strategy.strategies import noisy_neighbor
|
||||
from watcher.decision_engine.strategy.strategies import outlet_temp_control
|
||||
from watcher.decision_engine.strategy.strategies import saving_energy
|
||||
@@ -45,6 +47,8 @@ VMWorkloadConsolidation = vm_workload_consolidation.VMWorkloadConsolidation
|
||||
WorkloadBalance = workload_balance.WorkloadBalance
|
||||
WorkloadStabilization = workload_stabilization.WorkloadStabilization
|
||||
UniformAirflow = uniform_airflow.UniformAirflow
|
||||
NodeResourceConsolidation = (
|
||||
node_resource_consolidation.NodeResourceConsolidation)
|
||||
NoisyNeighbor = noisy_neighbor.NoisyNeighbor
|
||||
ZoneMigration = zone_migration.ZoneMigration
|
||||
HostMaintenance = host_maintenance.HostMaintenance
|
||||
@@ -54,4 +58,4 @@ __all__ = ("Actuator", "BaseStrategy", "BasicConsolidation",
|
||||
"VMWorkloadConsolidation", "WorkloadBalance",
|
||||
"WorkloadStabilization", "UniformAirflow", "NoisyNeighbor",
|
||||
"SavingEnergy", "StorageCapacityBalance", "ZoneMigration",
|
||||
"HostMaintenance")
|
||||
"HostMaintenance", "NodeResourceConsolidation")
|
||||
|
||||
@@ -0,0 +1,290 @@
|
||||
# -*- encoding: utf-8 -*-
|
||||
# Copyright (c) 2019 ZTE Corporation
|
||||
#
|
||||
# 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 oslo_log import log
|
||||
|
||||
from watcher._i18n import _
|
||||
from watcher.common import exception
|
||||
from watcher.decision_engine.model import element
|
||||
from watcher.decision_engine.strategy.strategies import base
|
||||
from watcher import objects
|
||||
|
||||
LOG = log.getLogger(__name__)
|
||||
|
||||
|
||||
class NodeResourceConsolidation(base.ServerConsolidationBaseStrategy):
|
||||
"""consolidating resources on nodes using server migration
|
||||
|
||||
*Description*
|
||||
|
||||
This strategy checks the resource usages of compute nodes, if the used
|
||||
resources are less than total, it will try to migrate server to
|
||||
consolidate the use of resource.
|
||||
|
||||
*Requirements*
|
||||
|
||||
* You must have at least 2 compute nodes to run
|
||||
this strategy.
|
||||
* Hardware: compute nodes should use the same physical CPUs/RAMs
|
||||
|
||||
*Limitations*
|
||||
|
||||
* This is a proof of concept that is not meant to be used in production
|
||||
* It assume that live migrations are possible
|
||||
|
||||
*Spec URL*
|
||||
|
||||
http://specs.openstack.org/openstack/watcher-specs/specs/train/implemented/node-resource-consolidation.html
|
||||
"""
|
||||
|
||||
CHANGE_NOVA_SERVICE_STATE = "change_nova_service_state"
|
||||
REASON_FOR_DISABLE = 'Watcher node resource consolidation strategy'
|
||||
|
||||
def __init__(self, config, osc=None):
|
||||
"""node resource consolidation
|
||||
|
||||
:param config: A mapping containing the configuration of this strategy
|
||||
:type config: :py:class:`~.Struct` instance
|
||||
:param osc: :py:class:`~.OpenStackClients` instance
|
||||
"""
|
||||
super(NodeResourceConsolidation, self).__init__(config, osc)
|
||||
self.host_choice = 'auto'
|
||||
self.audit = None
|
||||
self.compute_nodes_count = 0
|
||||
self.number_of_released_nodes = 0
|
||||
self.number_of_migrations = 0
|
||||
|
||||
@classmethod
|
||||
def get_name(cls):
|
||||
return "node_resource_consolidation"
|
||||
|
||||
@classmethod
|
||||
def get_display_name(cls):
|
||||
return _("Node Resource Consolidation strategy")
|
||||
|
||||
@classmethod
|
||||
def get_translatable_display_name(cls):
|
||||
return "Node Resource Consolidation strategy"
|
||||
|
||||
@classmethod
|
||||
def get_schema(cls):
|
||||
# Mandatory default setting for each element
|
||||
return {
|
||||
"properties": {
|
||||
"host_choice": {
|
||||
"description": "the way to select the server migration "
|
||||
"destination node, The value auto "
|
||||
"means that Nova schedular selects "
|
||||
"the destination node, and specify "
|
||||
"means the strategy specifies the "
|
||||
"destination.",
|
||||
"type": "string",
|
||||
"default": 'auto'
|
||||
},
|
||||
},
|
||||
}
|
||||
|
||||
def check_resources(self, servers, destination):
|
||||
# check whether a node able to accommodate a VM
|
||||
dest_flag = False
|
||||
if not destination:
|
||||
return dest_flag
|
||||
free_res = self.compute_model.get_node_free_resources(destination)
|
||||
for server in servers:
|
||||
# just vcpu and memory, do not consider disk
|
||||
if free_res['vcpu'] >= server.vcpus and (
|
||||
free_res['memory'] >= server.memory):
|
||||
free_res['vcpu'] -= server.vcpus
|
||||
free_res['memory'] -= server.memory
|
||||
dest_flag = True
|
||||
servers.remove(server)
|
||||
|
||||
return dest_flag
|
||||
|
||||
def select_destination(self, server, source, destinations):
|
||||
dest_node = None
|
||||
if not destinations:
|
||||
return dest_node
|
||||
sorted_nodes = sorted(
|
||||
destinations,
|
||||
key=lambda x: self.compute_model.get_node_free_resources(
|
||||
x)['vcpu'])
|
||||
for dest in sorted_nodes:
|
||||
if self.check_resources([server], dest):
|
||||
if self.compute_model.migrate_instance(server, source, dest):
|
||||
dest_node = dest
|
||||
break
|
||||
|
||||
return dest_node
|
||||
|
||||
def add_migrate_actions(self, sources, destinations):
|
||||
if not sources or not destinations:
|
||||
return
|
||||
for node in sources:
|
||||
servers = self.compute_model.get_node_instances(node)
|
||||
sorted_servers = sorted(
|
||||
servers,
|
||||
key=lambda x: x.vcpus,
|
||||
reverse=True)
|
||||
for server in sorted_servers:
|
||||
parameters = {'migration_type': 'live',
|
||||
'source_node': node.hostname,
|
||||
'resource_name': server.name}
|
||||
action_flag = False
|
||||
if self.host_choice != 'auto':
|
||||
# specify destination host
|
||||
dest = self.select_destination(server, node, destinations)
|
||||
if dest:
|
||||
parameters['destination_node'] = dest.hostname
|
||||
action_flag = True
|
||||
else:
|
||||
action_flag = True
|
||||
if action_flag:
|
||||
self.number_of_migrations += 1
|
||||
self.solution.add_action(
|
||||
action_type=self.MIGRATION,
|
||||
resource_id=server.uuid,
|
||||
input_parameters=parameters)
|
||||
|
||||
def add_change_node_state_actions(self, nodes, status):
|
||||
if status not in (element.ServiceState.DISABLED.value,
|
||||
element.ServiceState.ENABLED.value):
|
||||
raise exception.IllegalArgumentException(
|
||||
message=_("The node status is not defined"))
|
||||
changed_nodes = []
|
||||
for node in nodes:
|
||||
if node.status != status:
|
||||
parameters = {'state': status,
|
||||
'resource_name': node.hostname}
|
||||
if status == element.ServiceState.DISABLED.value:
|
||||
parameters['disabled_reason'] = self.REASON_FOR_DISABLE
|
||||
self.solution.add_action(
|
||||
action_type=self.CHANGE_NOVA_SERVICE_STATE,
|
||||
resource_id=node.uuid,
|
||||
input_parameters=parameters)
|
||||
node.status = status
|
||||
changed_nodes.append(node)
|
||||
|
||||
return changed_nodes
|
||||
|
||||
def get_nodes_migrate_failed(self):
|
||||
# check if migration action ever failed
|
||||
# just for continuous audit
|
||||
nodes_failed = []
|
||||
if self.audit is None or (
|
||||
self.audit.audit_type ==
|
||||
objects.audit.AuditType.ONESHOT.value):
|
||||
return nodes_failed
|
||||
filters = {'audit_uuid': self.audit.uuid}
|
||||
actions = objects.action.Action.list(
|
||||
self.ctx,
|
||||
filters=filters)
|
||||
for action in actions:
|
||||
if action.state == objects.action.State.FAILED and (
|
||||
action.action_type == self.MIGRATION):
|
||||
server_uuid = action.input_parameters.get('resource_id')
|
||||
node = self.compute_model.get_node_by_instance_uuid(
|
||||
server_uuid)
|
||||
if node not in nodes_failed:
|
||||
nodes_failed.append(node)
|
||||
|
||||
return nodes_failed
|
||||
|
||||
def group_nodes(self, nodes):
|
||||
free_nodes = []
|
||||
source_nodes = []
|
||||
dest_nodes = []
|
||||
nodes_failed = self.get_nodes_migrate_failed()
|
||||
LOG.info("nodes: %s migration failed", nodes_failed)
|
||||
sorted_nodes = sorted(
|
||||
nodes,
|
||||
key=lambda x: self.compute_model.get_node_used_resources(
|
||||
x)['vcpu'])
|
||||
for node in sorted_nodes:
|
||||
if node in dest_nodes:
|
||||
break
|
||||
# If ever migration failed, do not migrate again
|
||||
if node in nodes_failed:
|
||||
# maybe can as the destination node
|
||||
if node.status == element.ServiceState.ENABLED.value:
|
||||
dest_nodes.append(node)
|
||||
continue
|
||||
used_resource = self.compute_model.get_node_used_resources(node)
|
||||
if used_resource['vcpu'] > 0:
|
||||
servers = self.compute_model.get_node_instances(node)
|
||||
for dest in reversed(sorted_nodes):
|
||||
# skip if compute node is disabled
|
||||
if dest.status == element.ServiceState.DISABLED.value:
|
||||
LOG.info("node %s is down", dest.hostname)
|
||||
continue
|
||||
if dest in dest_nodes:
|
||||
continue
|
||||
if node == dest:
|
||||
# The last on as destination node
|
||||
dest_nodes.append(dest)
|
||||
break
|
||||
if self.check_resources(servers, dest):
|
||||
dest_nodes.append(dest)
|
||||
if node not in source_nodes:
|
||||
source_nodes.append(node)
|
||||
if not servers:
|
||||
break
|
||||
else:
|
||||
free_nodes.append(node)
|
||||
|
||||
return free_nodes, source_nodes, dest_nodes
|
||||
|
||||
def pre_execute(self):
|
||||
self._pre_execute()
|
||||
self.host_choice = self.input_parameters.host_choice
|
||||
|
||||
def do_execute(self, audit=None):
|
||||
"""Strategy execution phase
|
||||
|
||||
Executing strategy and creating solution.
|
||||
"""
|
||||
self.audit = audit
|
||||
nodes = list(self.compute_model.get_all_compute_nodes().values())
|
||||
free_nodes, source_nodes, dest_nodes = self.group_nodes(nodes)
|
||||
self.compute_nodes_count = len(nodes)
|
||||
self.number_of_released_nodes = len(source_nodes)
|
||||
LOG.info("Free nodes: %s", free_nodes)
|
||||
LOG.info("Source nodes: %s", source_nodes)
|
||||
LOG.info("Destination nodes: %s", dest_nodes)
|
||||
if not source_nodes:
|
||||
LOG.info("No compute node needs to be consolidated")
|
||||
return
|
||||
nodes_disabled = []
|
||||
if self.host_choice == 'auto':
|
||||
# disable compute node to avoid to be select by Nova scheduler
|
||||
nodes_disabled = self.add_change_node_state_actions(
|
||||
free_nodes+source_nodes, element.ServiceState.DISABLED.value)
|
||||
self.add_migrate_actions(source_nodes, dest_nodes)
|
||||
if nodes_disabled:
|
||||
# restore disabled compute node after migration
|
||||
self.add_change_node_state_actions(
|
||||
nodes_disabled, element.ServiceState.ENABLED.value)
|
||||
|
||||
def post_execute(self):
|
||||
"""Post-execution phase
|
||||
|
||||
"""
|
||||
self.solution.set_efficacy_indicators(
|
||||
compute_nodes_count=self.compute_nodes_count,
|
||||
released_compute_nodes_count=self.number_of_released_nodes,
|
||||
instance_migrations_count=self.number_of_migrations,
|
||||
)
|
||||
Reference in New Issue
Block a user