¡@

Home 

OpenStack Study: filter_scheduler.py

OpenStack Index

**** CubicPower OpenStack Study ****

# Copyright (c) 2011 OpenStack Foundation

# 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.

"""

The FilterScheduler is for creating instances locally.

You can customize this scheduler by specifying your own Host Filters and

Weighing Functions.

"""

import random

from oslo.config import cfg

from nova.compute import rpcapi as compute_rpcapi

from nova import exception

from nova.objects import instance_group as instance_group_obj

from nova.openstack.common.gettextutils import _

from nova.openstack.common import log as logging

from nova.pci import pci_request

from nova import rpc

from nova.scheduler import driver

from nova.scheduler import scheduler_options

from nova.scheduler import utils as scheduler_utils

CONF = cfg.CONF

LOG = logging.getLogger(__name__)

filter_scheduler_opts = [

cfg.IntOpt('scheduler_host_subset_size',

default=1,

help='New instances will be scheduled on a host chosen '

'randomly from a subset of the N best hosts. This '

'property defines the subset size that a host is '

'chosen from. A value of 1 chooses the '

'first host returned by the weighing functions. '

'This value must be at least 1. Any value less than 1 '

'will be ignored, and 1 will be used instead')

]

CONF.register_opts(filter_scheduler_opts)

**** CubicPower OpenStack Study ****

class FilterScheduler(driver.Scheduler):

"""Scheduler that can be used for filtering and weighing."""

**** CubicPower OpenStack Study ****

    def __init__(self, *args, **kwargs):

        super(FilterScheduler, self).__init__(*args, **kwargs)

        self.options = scheduler_options.SchedulerOptions()

        self.compute_rpcapi = compute_rpcapi.ComputeAPI()

        self.notifier = rpc.get_notifier('scheduler')

**** CubicPower OpenStack Study ****

    def schedule_run_instance(self, context, request_spec,

                              admin_password, injected_files,

                              requested_networks, is_first_time,

                              filter_properties, legacy_bdm_in_spec):

        """This method is called from nova.compute.api to provision

        an instance.  We first create a build plan (a list of WeightedHosts)

        and then provision.

        Returns a list of the instances created.

        """

        payload = dict(request_spec=request_spec)

        self.notifier.info(context, 'scheduler.run_instance.start', payload)

        instance_uuids = request_spec.get('instance_uuids')

        LOG.info(_("Attempting to build %(num_instances)d instance(s) "

                    "uuids: %(instance_uuids)s"),

                  {'num_instances': len(instance_uuids),

                   'instance_uuids': instance_uuids})

        LOG.debug(_("Request Spec: %s") % request_spec)

        weighed_hosts = self._schedule(context, request_spec,

                                       filter_properties, instance_uuids)

        # NOTE: Pop instance_uuids as individual creates do not need the

        # set of uuids. Do not pop before here as the upper exception

        # handler fo NoValidHost needs the uuid to set error state

        instance_uuids = request_spec.pop('instance_uuids')

        # NOTE(comstud): Make sure we do not pass this through.  It

        # contains an instance of RpcContext that cannot be serialized.

        filter_properties.pop('context', None)

        for num, instance_uuid in enumerate(instance_uuids):

            request_spec['instance_properties']['launch_index'] = num

            try:

                try:

                    weighed_host = weighed_hosts.pop(0)

                    LOG.info(_("Choosing host %(weighed_host)s "

                                "for instance %(instance_uuid)s"),

                              {'weighed_host': weighed_host,

                               'instance_uuid': instance_uuid})

                except IndexError:

                    raise exception.NoValidHost(reason="")

                self._provision_resource(context, weighed_host,

                                         request_spec,

                                         filter_properties,

                                         requested_networks,

                                         injected_files, admin_password,

                                         is_first_time,

                                         instance_uuid=instance_uuid,

                                         legacy_bdm_in_spec=legacy_bdm_in_spec)

            except Exception as ex:

                # NOTE(vish): we don't reraise the exception here to make sure

                #             that all instances in the request get set to

                #             error properly

                driver.handle_schedule_error(context, ex, instance_uuid,

                                             request_spec)

            # scrub retry host list in case we're scheduling multiple

            # instances:

            retry = filter_properties.get('retry', {})

            retry['hosts'] = []

        self.notifier.info(context, 'scheduler.run_instance.end', payload)

**** CubicPower OpenStack Study ****

    def select_destinations(self, context, request_spec, filter_properties):

        """Selects a filtered set of hosts and nodes."""

        num_instances = request_spec['num_instances']

        instance_uuids = request_spec.get('instance_uuids')

        selected_hosts = self._schedule(context, request_spec,

                                        filter_properties, instance_uuids)

        # Couldn't fulfill the request_spec

        if len(selected_hosts) < num_instances:

            raise exception.NoValidHost(reason='')

        dests = [dict(host=host.obj.host, nodename=host.obj.nodename,

                      limits=host.obj.limits) for host in selected_hosts]

        return dests

**** CubicPower OpenStack Study ****

    def _provision_resource(self, context, weighed_host, request_spec,

            filter_properties, requested_networks, injected_files,

            admin_password, is_first_time, instance_uuid=None,

            legacy_bdm_in_spec=True):

        """Create the requested resource in this Zone."""

        # NOTE(vish): add our current instance back into the request spec

        request_spec['instance_uuids'] = [instance_uuid]

        payload = dict(request_spec=request_spec,

                       weighted_host=weighed_host.to_dict(),

                       instance_id=instance_uuid)

        self.notifier.info(context,

                           'scheduler.run_instance.scheduled', payload)

        # Update the metadata if necessary

        scheduler_hints = filter_properties.get('scheduler_hints') or {}

        try:

            updated_instance = driver.instance_update_db(context,

                                                         instance_uuid)

        except exception.InstanceNotFound:

            LOG.warning(_("Instance disappeared during scheduling"),

                        context=context, instance_uuid=instance_uuid)

        else:

            scheduler_utils.populate_filter_properties(filter_properties,

                    weighed_host.obj)

            self.compute_rpcapi.run_instance(context,

                    instance=updated_instance,

                    host=weighed_host.obj.host,

                    request_spec=request_spec,

                    filter_properties=filter_properties,

                    requested_networks=requested_networks,

                    injected_files=injected_files,

                    admin_password=admin_password, is_first_time=is_first_time,

                    node=weighed_host.obj.nodename,

                    legacy_bdm_in_spec=legacy_bdm_in_spec)

**** CubicPower OpenStack Study ****

    def _get_configuration_options(self):

        """Fetch options dictionary. Broken out for testing."""

        return self.options.get_configuration()

**** CubicPower OpenStack Study ****

    def populate_filter_properties(self, request_spec, filter_properties):

        """Stuff things into filter_properties.  Can be overridden in a

        subclass to add more data.

        """

        # Save useful information from the request spec for filter processing:

        project_id = request_spec['instance_properties']['project_id']

        os_type = request_spec['instance_properties']['os_type']

        filter_properties['project_id'] = project_id

        filter_properties['os_type'] = os_type

        pci_requests = pci_request.get_pci_requests_from_flavor(

            request_spec.get('instance_type') or {})

        if pci_requests:

            filter_properties['pci_requests'] = pci_requests

**** CubicPower OpenStack Study ****

    def _max_attempts(self):

        max_attempts = CONF.scheduler_max_attempts

        if max_attempts < 1:

            raise exception.NovaException(_("Invalid value for "

                "'scheduler_max_attempts', must be >= 1"))

        return max_attempts

**** CubicPower OpenStack Study ****

    def _log_compute_error(self, instance_uuid, retry):

        """If the request contained an exception from a previous compute

        build/resize operation, log it to aid debugging

        """

        exc = retry.pop('exc', None)  # string-ified exception from compute

        if not exc:

            return  # no exception info from a previous attempt, skip

        hosts = retry.get('hosts', None)

        if not hosts:

            return  # no previously attempted hosts, skip

        last_host, last_node = hosts[-1]

        LOG.error(_('Error from last host: %(last_host)s (node %(last_node)s):'

                    ' %(exc)s'),

                  {'last_host': last_host,

                   'last_node': last_node,

                   'exc': exc},

                  instance_uuid=instance_uuid)

**** CubicPower OpenStack Study ****

    def _populate_retry(self, filter_properties, instance_properties):

        """Populate filter properties with history of retries for this

        request. If maximum retries is exceeded, raise NoValidHost.

        """

        max_attempts = self._max_attempts()

        force_hosts = filter_properties.get('force_hosts', [])

        force_nodes = filter_properties.get('force_nodes', [])

        if max_attempts == 1 or force_hosts or force_nodes:

            # re-scheduling is disabled.

            return

        retry = filter_properties.pop('retry', {})

        # retry is enabled, update attempt count:

        if retry:

            retry['num_attempts'] += 1

        else:

            retry = {

                'num_attempts': 1,

                'hosts': []  # list of compute hosts tried

            }

        filter_properties['retry'] = retry

        instance_uuid = instance_properties.get('uuid')

        self._log_compute_error(instance_uuid, retry)

        if retry['num_attempts'] > max_attempts:

            msg = (_('Exceeded max scheduling attempts %(max_attempts)d for '

                     'instance %(instance_uuid)s')

                   % {'max_attempts': max_attempts,

                      'instance_uuid': instance_uuid})

            raise exception.NoValidHost(reason=msg)

    @staticmethod

**** CubicPower OpenStack Study ****

    def _setup_instance_group(context, filter_properties):

        update_group_hosts = False

        scheduler_hints = filter_properties.get('scheduler_hints') or {}

        group_uuid = scheduler_hints.get('group', None)

        if group_uuid:

            group = instance_group_obj.InstanceGroup.get_by_uuid(context,

                    group_uuid)

            policies = set(('anti-affinity', 'affinity'))

            if any((policy in policies) for policy in group.policies):

                update_group_hosts = True

                filter_properties.setdefault('group_hosts', set())

                user_hosts = set(filter_properties['group_hosts'])

                group_hosts = set(group.get_hosts(context))

                filter_properties['group_hosts'] = user_hosts | group_hosts

                filter_properties['group_policies'] = group.policies

        return update_group_hosts

**** CubicPower OpenStack Study ****

    def _schedule(self, context, request_spec, filter_properties,

                  instance_uuids=None):

        """Returns a list of hosts that meet the required specs,

        ordered by their fitness.

        """

        elevated = context.elevated()

        instance_properties = request_spec['instance_properties']

        instance_type = request_spec.get("instance_type", None)

        update_group_hosts = self._setup_instance_group(context,

                filter_properties)

        config_options = self._get_configuration_options()

        # check retry policy.  Rather ugly use of instance_uuids[0]...

        # but if we've exceeded max retries... then we really only

        # have a single instance.

        properties = instance_properties.copy()

        if instance_uuids:

            properties['uuid'] = instance_uuids[0]

        self._populate_retry(filter_properties, properties)

        filter_properties.update({'context': context,

                                  'request_spec': request_spec,

                                  'config_options': config_options,

                                  'instance_type': instance_type})

        self.populate_filter_properties(request_spec,

                                        filter_properties)

        # Find our local list of acceptable hosts by repeatedly

        # filtering and weighing our options. Each time we choose a

        # host, we virtually consume resources on it so subsequent

        # selections can adjust accordingly.

        # Note: remember, we are using an iterator here. So only

        # traverse this list once. This can bite you if the hosts

        # are being scanned in a filter or weighing function.

        hosts = self._get_all_host_states(elevated)

        selected_hosts = []

        if instance_uuids:

            num_instances = len(instance_uuids)

        else:

            num_instances = request_spec.get('num_instances', 1)

        for num in xrange(num_instances):

            # Filter local hosts based on requirements ...

            hosts = self.host_manager.get_filtered_hosts(hosts,

                    filter_properties, index=num)

            if not hosts:

                # Can't get any more locally.

                break

            LOG.debug(_("Filtered %(hosts)s"), {'hosts': hosts})

            weighed_hosts = self.host_manager.get_weighed_hosts(hosts,

                    filter_properties)

            LOG.debug(_("Weighed %(hosts)s"), {'hosts': weighed_hosts})

            scheduler_host_subset_size = CONF.scheduler_host_subset_size

            if scheduler_host_subset_size > len(weighed_hosts):

                scheduler_host_subset_size = len(weighed_hosts)

            if scheduler_host_subset_size < 1:

                scheduler_host_subset_size = 1

            chosen_host = random.choice(

                weighed_hosts[0:scheduler_host_subset_size])

            selected_hosts.append(chosen_host)

            # Now consume the resources so the filter/weights

            # will change for the next instance.

            chosen_host.obj.consume_from_instance(instance_properties)

            if update_group_hosts is True:

                filter_properties['group_hosts'].add(chosen_host.obj.host)

        return selected_hosts

**** CubicPower OpenStack Study ****

    def _get_all_host_states(self, context):

        """Template method, so a subclass can implement caching."""

        return self.host_manager.get_all_host_states(context)