# Copyright 2013 IBM Corp. # # 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 contextlib from oslo_config import cfg from oslo_db import exception as db_exc from oslo_log import log as logging from oslo_serialization import jsonutils from oslo_utils import timeutils from oslo_utils import versionutils from sqlalchemy import or_ from sqlalchemy.sql import func from sqlalchemy.sql import null from nova import availability_zones as avail_zone from nova.cells import opts as cells_opts from nova.cells import rpcapi as cells_rpcapi from nova.cells import utils as cells_utils from nova.compute import task_states from nova.compute import vm_states from nova.db import api as db from nova.db.sqlalchemy import api as db_api from nova.db.sqlalchemy import models from nova import exception from nova.i18n import _ from nova.network import model as network_model from nova import notifications from nova import objects from nova.objects import base from nova.objects import fields from nova import utils CONF = cfg.CONF LOG = logging.getLogger(__name__) # List of fields that can be joined in DB layer. _INSTANCE_OPTIONAL_JOINED_FIELDS = ['metadata', 'system_metadata', 'info_cache', 'security_groups', 'pci_devices', 'tags', 'services', 'fault'] # These are fields that are optional but don't translate to db columns _INSTANCE_OPTIONAL_NON_COLUMN_FIELDS = ['flavor', 'old_flavor', 'new_flavor', 'ec2_ids'] # These are fields that are optional and in instance_extra _INSTANCE_EXTRA_FIELDS = ['numa_topology', 'pci_requests', 'flavor', 'vcpu_model', 'migration_context', 'keypairs', 'device_metadata', 'trusted_certs'] # These are fields that applied/drooped by migration_context _MIGRATION_CONTEXT_ATTRS = ['numa_topology', 'pci_requests', 'pci_devices'] # These are fields that can be specified as expected_attrs INSTANCE_OPTIONAL_ATTRS = (_INSTANCE_OPTIONAL_JOINED_FIELDS + _INSTANCE_OPTIONAL_NON_COLUMN_FIELDS + _INSTANCE_EXTRA_FIELDS) # These are fields that most query calls load by default INSTANCE_DEFAULT_FIELDS = ['metadata', 'system_metadata', 'info_cache', 'security_groups'] # Maximum count of tags to one instance MAX_TAG_COUNT = 50 def _expected_cols(expected_attrs): """Return expected_attrs that are columns needing joining. NB: This function may modify expected_attrs if one requested attribute requires another. """ if not expected_attrs: return expected_attrs simple_cols = [attr for attr in expected_attrs if attr in _INSTANCE_OPTIONAL_JOINED_FIELDS] complex_cols = ['extra.%s' % field for field in _INSTANCE_EXTRA_FIELDS if field in expected_attrs] if complex_cols: simple_cols.append('extra') simple_cols = [x for x in simple_cols if x not in _INSTANCE_EXTRA_FIELDS] expected_cols = simple_cols + complex_cols # NOTE(pumaranikar): expected_cols list can contain duplicates since # caller appends column attributes to expected_attr without checking if # it is already present in the list or not. Hence, we remove duplicates # here, if any. The resultant list is sorted based on list index to # maintain the insertion order. return sorted(list(set(expected_cols)), key=expected_cols.index) _NO_DATA_SENTINEL = object() # TODO(berrange): Remove NovaObjectDictCompat @base.NovaObjectRegistry.register class Instance(base.NovaPersistentObject, base.NovaObject, base.NovaObjectDictCompat): # Version 2.0: Initial version # Version 2.1: Added services # Version 2.2: Added keypairs # Version 2.3: Added device_metadata # Version 2.4: Added trusted_certs VERSION = '2.4' fields = { 'id': fields.IntegerField(), 'user_id': fields.StringField(nullable=True), 'project_id': fields.StringField(nullable=True), 'image_ref': fields.StringField(nullable=True), 'kernel_id': fields.StringField(nullable=True), 'ramdisk_id': fields.StringField(nullable=True), 'hostname': fields.StringField(nullable=True), 'launch_index': fields.IntegerField(nullable=True), 'key_name': fields.StringField(nullable=True), 'key_data': fields.StringField(nullable=True), 'power_state': fields.IntegerField(nullable=True), 'vm_state': fields.StringField(nullable=True), 'task_state': fields.StringField(nullable=True), 'services': fields.ObjectField('ServiceList'), 'memory_mb': fields.IntegerField(nullable=True), 'vcpus': fields.IntegerField(nullable=True), 'root_gb': fields.IntegerField(nullable=True), 'ephemeral_gb': fields.IntegerField(nullable=True), 'ephemeral_key_uuid': fields.UUIDField(nullable=True), 'host': fields.StringField(nullable=True), 'node': fields.StringField(nullable=True), 'instance_type_id': fields.IntegerField(nullable=True), 'user_data': fields.StringField(nullable=True), 'reservation_id': fields.StringField(nullable=True), 'launched_at': fields.DateTimeField(nullable=True), 'terminated_at': fields.DateTimeField(nullable=True), 'availability_zone': fields.StringField(nullable=True), 'display_name': fields.StringField(nullable=True), 'display_description': fields.StringField(nullable=True), 'launched_on': fields.StringField(nullable=True), 'locked': fields.BooleanField(default=False), 'locked_by': fields.StringField(nullable=True), 'os_type': fields.StringField(nullable=True), 'architecture': fields.StringField(nullable=True), 'vm_mode': fields.StringField(nullable=True), 'uuid': fields.UUIDField(), 'root_device_name': fields.StringField(nullable=True), 'default_ephemeral_device': fields.StringField(nullable=True), 'default_swap_device': fields.StringField(nullable=True), 'config_drive': fields.StringField(nullable=True), 'access_ip_v4': fields.IPV4AddressField(nullable=True), 'access_ip_v6': fields.IPV6AddressField(nullable=True), 'auto_disk_config': fields.BooleanField(default=False), 'progress': fields.IntegerField(nullable=True), 'shutdown_terminate': fields.BooleanField(default=False), 'disable_terminate': fields.BooleanField(default=False), 'cell_name': fields.StringField(nullable=True), 'metadata': fields.DictOfStringsField(), 'system_metadata': fields.DictOfNullableStringsField(), 'info_cache': fields.ObjectField('InstanceInfoCache', nullable=True), 'security_groups': fields.ObjectField('SecurityGroupList'), 'fault': fields.ObjectField('InstanceFault', nullable=True), 'cleaned': fields.BooleanField(default=False), 'pci_devices': fields.ObjectField('PciDeviceList', nullable=True), 'numa_topology': fields.ObjectField('InstanceNUMATopology', nullable=True), 'pci_requests': fields.ObjectField('InstancePCIRequests', nullable=True), 'device_metadata': fields.ObjectField('InstanceDeviceMetadata', nullable=True), 'tags': fields.ObjectField('TagList'), 'flavor': fields.ObjectField('Flavor'), 'old_flavor': fields.ObjectField('Flavor', nullable=True), 'new_flavor': fields.ObjectField('Flavor', nullable=True), 'vcpu_model': fields.ObjectField('VirtCPUModel', nullable=True), 'ec2_ids': fields.ObjectField('EC2Ids'), 'migration_context': fields.ObjectField('MigrationContext', nullable=True), 'keypairs': fields.ObjectField('KeyPairList'), 'trusted_certs': fields.ObjectField('TrustedCerts', nullable=True), } obj_extra_fields = ['name'] def obj_make_compatible(self, primitive, target_version): super(Instance, self).obj_make_compatible(primitive, target_version) target_version = versionutils.convert_version_to_tuple(target_version) if target_version < (2, 4) and 'trusted_certs' in primitive: del primitive['trusted_certs'] if target_version < (2, 3) and 'device_metadata' in primitive: del primitive['device_metadata'] if target_version < (2, 2) and 'keypairs' in primitive: del primitive['keypairs'] if target_version < (2, 1) and 'services' in primitive: del primitive['services'] def __init__(self, *args, **kwargs): super(Instance, self).__init__(*args, **kwargs) self._reset_metadata_tracking() @property def image_meta(self): return objects.ImageMeta.from_instance(self) def _reset_metadata_tracking(self, fields=None): if fields is None or 'system_metadata' in fields: self._orig_system_metadata = (dict(self.system_metadata) if 'system_metadata' in self else {}) if fields is None or 'metadata' in fields: self._orig_metadata = (dict(self.metadata) if 'metadata' in self else {}) def obj_clone(self): """Create a copy of this instance object.""" nobj = super(Instance, self).obj_clone() # Since the base object only does a deep copy of the defined fields, # need to make sure to also copy the additional tracking metadata # attributes so they don't show as changed and cause the metadata # to always be updated even when stale information. if hasattr(self, '_orig_metadata'): nobj._orig_metadata = dict(self._orig_metadata) if hasattr(self, '_orig_system_metadata'): nobj._orig_system_metadata = dict(self._orig_system_metadata) return nobj def obj_reset_changes(self, fields=None, recursive=False): super(Instance, self).obj_reset_changes(fields, recursive=recursive) self._reset_metadata_tracking(fields=fields) def obj_what_changed(self): changes = super(Instance, self).obj_what_changed() if 'metadata' in self and self.metadata != self._orig_metadata: changes.add('metadata') if 'system_metadata' in self and (self.system_metadata != self._orig_system_metadata): changes.add('system_metadata') return changes @classmethod def _obj_from_primitive(cls, context, objver, primitive): self = super(Instance, cls)._obj_from_primitive(context, objver, primitive) self._reset_metadata_tracking() return self @property def name(self): try: base_name = CONF.instance_name_template % self.id except TypeError: # Support templates like "uuid-%(uuid)s", etc. info = {} # NOTE(russellb): Don't use self.iteritems() here, as it will # result in infinite recursion on the name property. for key in self.fields: if key == 'name': # NOTE(danms): prevent recursion continue elif not self.obj_attr_is_set(key): # NOTE(danms): Don't trigger lazy-loads continue info[key] = self[key] try: base_name = CONF.instance_name_template % info except KeyError: base_name = self.uuid except (exception.ObjectActionError, exception.OrphanedObjectError): # This indicates self.id was not set and/or could not be # lazy loaded. What this means is the instance has not # been persisted to a db yet, which should indicate it has # not been scheduled yet. In this situation it will have a # blank name. if (self.vm_state == vm_states.BUILDING and self.task_state == task_states.SCHEDULING): base_name = '' else: # If the vm/task states don't indicate that it's being booted # then we have a bug here. Log an error and attempt to return # the uuid which is what an error above would return. LOG.error('Could not lazy-load instance.id while ' 'attempting to generate the instance name.') base_name = self.uuid return base_name def _flavor_from_db(self, db_flavor): """Load instance flavor information from instance_extra.""" # Before we stored flavors in instance_extra, certain fields, defined # in nova.compute.flavors.system_metadata_flavor_props, were stored # in the instance.system_metadata for the embedded instance.flavor. # The "disabled" and "is_public" fields weren't one of those keys, # however, so really old instances that had their embedded flavor # converted to the serialized instance_extra form won't have the # disabled attribute set and we need to default those here so callers # don't explode trying to load instance.flavor.disabled. def _default_flavor_values(flavor): if 'disabled' not in flavor: flavor.disabled = False if 'is_public' not in flavor: flavor.is_public = True flavor_info = jsonutils.loads(db_flavor) self.flavor = objects.Flavor.obj_from_primitive(flavor_info['cur']) _default_flavor_values(self.flavor) if flavor_info['old']: self.old_flavor = objects.Flavor.obj_from_primitive( flavor_info['old']) _default_flavor_values(self.old_flavor) else: self.old_flavor = None if flavor_info['new']: self.new_flavor = objects.Flavor.obj_from_primitive( flavor_info['new']) _default_flavor_values(self.new_flavor) else: self.new_flavor = None self.obj_reset_changes(['flavor', 'old_flavor', 'new_flavor']) @staticmethod def _from_db_object(context, instance, db_inst, expected_attrs=None): """Method to help with migration to objects. Converts a database entity to a formal object. """ instance._context = context if expected_attrs is None: expected_attrs = [] # Most of the field names match right now, so be quick for field in instance.fields: if field in INSTANCE_OPTIONAL_ATTRS: continue elif field == 'deleted': instance.deleted = db_inst['deleted'] == db_inst['id'] elif field == 'cleaned': instance.cleaned = db_inst['cleaned'] == 1 else: instance[field] = db_inst[field] if 'metadata' in expected_attrs: instance['metadata'] = utils.instance_meta(db_inst) if 'system_metadata' in expected_attrs: instance['system_metadata'] = utils.instance_sys_meta(db_inst) if 'fault' in expected_attrs: instance['fault'] = ( objects.InstanceFault.get_latest_for_instance( context, instance.uuid)) if 'ec2_ids' in expected_attrs: instance._load_ec2_ids() if 'info_cache' in expected_attrs: if db_inst.get('info_cache') is None: instance.info_cache = None elif not instance.obj_attr_is_set('info_cache'): # TODO(danms): If this ever happens on a backlevel instance # passed to us by a backlevel service, things will break instance.info_cache = objects.InstanceInfoCache(context) if instance.info_cache is not None: instance.info_cache._from_db_object(context, instance.info_cache, db_inst['info_cache']) # TODO(danms): If we are updating these on a backlevel instance, # we'll end up sending back new versions of these objects (see # above note for new info_caches if 'pci_devices' in expected_attrs: pci_devices = base.obj_make_list( context, objects.PciDeviceList(context), objects.PciDevice, db_inst['pci_devices']) instance['pci_devices'] = pci_devices if 'security_groups' in expected_attrs: sec_groups = base.obj_make_list( context, objects.SecurityGroupList(context), objects.SecurityGroup, db_inst.get('security_groups', [])) instance['security_groups'] = sec_groups if 'tags' in expected_attrs: tags = base.obj_make_list( context, objects.TagList(context), objects.Tag, db_inst['tags']) instance['tags'] = tags if 'services' in expected_attrs: services = base.obj_make_list( context, objects.ServiceList(context), objects.Service, db_inst['services']) instance['services'] = services instance._extra_attributes_from_db_object(instance, db_inst, expected_attrs) instance.obj_reset_changes() return instance @staticmethod def _extra_attributes_from_db_object(instance, db_inst, expected_attrs=None): """Method to help with migration of extra attributes to objects. """ if expected_attrs is None: expected_attrs = [] # NOTE(danms): We can be called with a dict instead of a # SQLAlchemy object, so we have to be careful here if hasattr(db_inst, '__dict__'): have_extra = 'extra' in db_inst.__dict__ and db_inst['extra'] else: have_extra = 'extra' in db_inst and db_inst['extra'] if 'numa_topology' in expected_attrs: if have_extra: instance._load_numa_topology( db_inst['extra'].get('numa_topology')) else: instance.numa_topology = None if 'pci_requests' in expected_attrs: if have_extra: instance._load_pci_requests( db_inst['extra'].get('pci_requests')) else: instance.pci_requests = None if 'device_metadata' in expected_attrs: if have_extra: instance._load_device_metadata( db_inst['extra'].get('device_metadata')) else: instance.device_metadata = None if 'vcpu_model' in expected_attrs: if have_extra: instance._load_vcpu_model( db_inst['extra'].get('vcpu_model')) else: instance.vcpu_model = None if 'migration_context' in expected_attrs: if have_extra: instance._load_migration_context( db_inst['extra'].get('migration_context')) else: instance.migration_context = None if 'keypairs' in expected_attrs: if have_extra: instance._load_keypairs(db_inst['extra'].get('keypairs')) if 'trusted_certs' in expected_attrs: if have_extra: instance._load_trusted_certs( db_inst['extra'].get('trusted_certs')) else: instance.trusted_certs = None if any([x in expected_attrs for x in ('flavor', 'old_flavor', 'new_flavor')]): if have_extra and db_inst['extra'].get('flavor'): instance._flavor_from_db(db_inst['extra']['flavor']) @staticmethod @db.select_db_reader_mode def _db_instance_get_by_uuid(context, uuid, columns_to_join, use_slave=False): return db.instance_get_by_uuid(context, uuid, columns_to_join=columns_to_join) @base.remotable_classmethod def get_by_uuid(cls, context, uuid, expected_attrs=None, use_slave=False): if expected_attrs is None: expected_attrs = ['info_cache', 'security_groups'] columns_to_join = _expected_cols(expected_attrs) db_inst = cls._db_instance_get_by_uuid(context, uuid, columns_to_join, use_slave=use_slave) return cls._from_db_object(context, cls(), db_inst, expected_attrs) @base.remotable_classmethod def get_by_id(cls, context, inst_id, expected_attrs=None): if expected_attrs is None: expected_attrs = ['info_cache', 'security_groups'] columns_to_join = _expected_cols(expected_attrs) db_inst = db.instance_get(context, inst_id, columns_to_join=columns_to_join) return cls._from_db_object(context, cls(), db_inst, expected_attrs) @base.remotable def create(self): if self.obj_attr_is_set('id'): raise exception.ObjectActionError(action='create', reason='already created') if self.obj_attr_is_set('deleted') and self.deleted: raise exception.ObjectActionError(action='create', reason='already deleted') updates = self.obj_get_changes() # NOTE(danms): We know because of the check above that deleted # is either unset or false. Since we need to avoid passing False # down to the DB layer (which uses an integer), we can always # default it to zero here. updates['deleted'] = 0 expected_attrs = [attr for attr in INSTANCE_DEFAULT_FIELDS if attr in updates] if 'security_groups' in updates: updates['security_groups'] = [x.name for x in updates['security_groups']] if 'info_cache' in updates: updates['info_cache'] = { 'network_info': updates['info_cache'].network_info.json() } updates['extra'] = {} numa_topology = updates.pop('numa_topology', None) expected_attrs.append('numa_topology') if numa_topology: updates['extra']['numa_topology'] = numa_topology._to_json() else: updates['extra']['numa_topology'] = None pci_requests = updates.pop('pci_requests', None) expected_attrs.append('pci_requests') if pci_requests: updates['extra']['pci_requests'] = ( pci_requests.to_json()) else: updates['extra']['pci_requests'] = None device_metadata = updates.pop('device_metadata', None) expected_attrs.append('device_metadata') if device_metadata: updates['extra']['device_metadata'] = ( device_metadata._to_json()) else: updates['extra']['device_metadata'] = None flavor = updates.pop('flavor', None) if flavor: expected_attrs.append('flavor') old = ((self.obj_attr_is_set('old_flavor') and self.old_flavor) and self.old_flavor.obj_to_primitive() or None) new = ((self.obj_attr_is_set('new_flavor') and self.new_flavor) and self.new_flavor.obj_to_primitive() or None) flavor_info = { 'cur': self.flavor.obj_to_primitive(), 'old': old, 'new': new, } self._nullify_flavor_description(flavor_info) updates['extra']['flavor'] = jsonutils.dumps(flavor_info) keypairs = updates.pop('keypairs', None) if keypairs is not None: expected_attrs.append('keypairs') updates['extra']['keypairs'] = jsonutils.dumps( keypairs.obj_to_primitive()) vcpu_model = updates.pop('vcpu_model', None) expected_attrs.append('vcpu_model') if vcpu_model: updates['extra']['vcpu_model'] = ( jsonutils.dumps(vcpu_model.obj_to_primitive())) else: updates['extra']['vcpu_model'] = None trusted_certs = updates.pop('trusted_certs', None) expected_attrs.append('trusted_certs') if trusted_certs: updates['extra']['trusted_certs'] = jsonutils.dumps( trusted_certs.obj_to_primitive()) else: updates['extra']['trusted_certs'] = None db_inst = db.instance_create(self._context, updates) self._from_db_object(self._context, self, db_inst, expected_attrs) # NOTE(danms): The EC2 ids are created on their first load. In order # to avoid them being missing and having to be loaded later, we # load them once here on create now that the instance record is # created. self._load_ec2_ids() self.obj_reset_changes(['ec2_ids']) @base.remotable def destroy(self): if not self.obj_attr_is_set('id'): raise exception.ObjectActionError(action='destroy', reason='already destroyed') if not self.obj_attr_is_set('uuid'): raise exception.ObjectActionError(action='destroy', reason='no uuid') if not self.obj_attr_is_set('host') or not self.host: # NOTE(danms): If our host is not set, avoid a race constraint = db.constraint(host=db.equal_any(None)) else: constraint = None cell_type = cells_opts.get_cell_type() if cell_type is not None: stale_instance = self.obj_clone() try: db_inst = db.instance_destroy(self._context, self.uuid, constraint=constraint) self._from_db_object(self._context, self, db_inst) except exception.ConstraintNotMet: raise exception.ObjectActionError(action='destroy', reason='host changed') if cell_type == 'compute': cells_api = cells_rpcapi.CellsAPI() cells_api.instance_destroy_at_top(self._context, stale_instance) delattr(self, base.get_attrname('id')) def _save_info_cache(self, context): if self.info_cache: with self.info_cache.obj_alternate_context(context): self.info_cache.save() def _save_security_groups(self, context): security_groups = self.security_groups or [] for secgroup in security_groups: with secgroup.obj_alternate_context(context): secgroup.save() self.security_groups.obj_reset_changes() def _save_fault(self, context): # NOTE(danms): I don't think we need to worry about this, do we? pass def _save_pci_requests(self, context): # TODO(danms): Unfortunately, extra.pci_requests is not a serialized # PciRequests object (!), so we have to handle it specially here. # That should definitely be fixed! self._extra_values_to_save['pci_requests'] = ( self.pci_requests.to_json()) def _save_pci_devices(self, context): # NOTE(yjiang5): All devices held by PCI tracker, only PCI tracker # permitted to update the DB. all change to devices from here will # be dropped. pass def _save_tags(self, context): # NOTE(gibi): tags are not saved through the instance pass @staticmethod def _nullify_flavor_description(flavor_info): """Helper method to nullify descriptions from a set of primitive flavors. Note that we don't remove the flavor description since that would make the versioned notification FlavorPayload have to handle the field not being set on the embedded instance.flavor. :param dict: dict of primitive flavor objects where the values are the flavors which get persisted in the instance_extra.flavor table. """ for flavor in flavor_info.values(): if flavor and 'description' in flavor['nova_object.data']: flavor['nova_object.data']['description'] = None def _save_flavor(self, context): if not any([x in self.obj_what_changed() for x in ('flavor', 'old_flavor', 'new_flavor')]): return flavor_info = { 'cur': self.flavor.obj_to_primitive(), 'old': (self.old_flavor and self.old_flavor.obj_to_primitive() or None), 'new': (self.new_flavor and self.new_flavor.obj_to_primitive() or None), } self._nullify_flavor_description(flavor_info) self._extra_values_to_save['flavor'] = jsonutils.dumps(flavor_info) self.obj_reset_changes(['flavor', 'old_flavor', 'new_flavor']) def _save_old_flavor(self, context): if 'old_flavor' in self.obj_what_changed(): self._save_flavor(context) def _save_new_flavor(self, context): if 'new_flavor' in self.obj_what_changed(): self._save_flavor(context) def _save_ec2_ids(self, context): # NOTE(hanlind): Read-only so no need to save this. pass def _save_keypairs(self, context): # NOTE(danms): Read-only so no need to save this. pass def _save_extra_generic(self, field): if field in self.obj_what_changed(): obj = getattr(self, field) value = None if obj is not None: value = jsonutils.dumps(obj.obj_to_primitive()) self._extra_values_to_save[field] = value @base.remotable def save(self, expected_vm_state=None, expected_task_state=None, admin_state_reset=False): """Save updates to this instance Column-wise updates will be made based on the result of self.obj_what_changed(). If expected_task_state is provided, it will be checked against the in-database copy of the instance before updates are made. :param expected_vm_state: Optional tuple of valid vm states for the instance to be in :param expected_task_state: Optional tuple of valid task states for the instance to be in :param admin_state_reset: True if admin API is forcing setting of task_state/vm_state """ # Store this on the class because _cell_name_blocks_sync is useless # after the db update call below. self._sync_cells = not self._cell_name_blocks_sync() context = self._context cell_type = cells_opts.get_cell_type() if cell_type is not None: # NOTE(comstud): We need to stash a copy of ourselves # before any updates are applied. When we call the save # methods on nested objects, we will lose any changes to # them. But we need to make sure child cells can tell # what is changed. # # We also need to nuke any updates to vm_state and task_state # unless admin_state_reset is True. compute cells are # authoritative for their view of vm_state and task_state. stale_instance = self.obj_clone() cells_update_from_api = (cell_type == 'api' and self.cell_name and self._sync_cells) if cells_update_from_api: def _handle_cell_update_from_api(): cells_api = cells_rpcapi.CellsAPI() cells_api.instance_update_from_api(context, stale_instance, expected_vm_state, expected_task_state, admin_state_reset) self._extra_values_to_save = {} updates = {} changes = self.obj_what_changed() for field in self.fields: # NOTE(danms): For object fields, we construct and call a # helper method like self._save_$attrname() if (self.obj_attr_is_set(field) and isinstance(self.fields[field], fields.ObjectField)): try: getattr(self, '_save_%s' % field)(context) except AttributeError: if field in _INSTANCE_EXTRA_FIELDS: self._save_extra_generic(field) continue LOG.exception('No save handler for %s', field, instance=self) except db_exc.DBReferenceError as exp: if exp.key != 'instance_uuid': raise # NOTE(melwitt): This will happen if we instance.save() # before an instance.create() and FK constraint fails. # In practice, this occurs in cells during a delete of # an unscheduled instance. Otherwise, it could happen # as a result of bug. raise exception.InstanceNotFound(instance_id=self.uuid) elif field in changes: if (field == 'cell_name' and self[field] is not None and self[field].startswith(cells_utils.BLOCK_SYNC_FLAG)): updates[field] = self[field].replace( cells_utils.BLOCK_SYNC_FLAG, '', 1) else: updates[field] = self[field] if self._extra_values_to_save: db.instance_extra_update_by_uuid(context, self.uuid, self._extra_values_to_save) if not updates: if cells_update_from_api: _handle_cell_update_from_api() return # Cleaned needs to be turned back into an int here if 'cleaned' in updates: if updates['cleaned']: updates['cleaned'] = 1 else: updates['cleaned'] = 0 if expected_task_state is not None: updates['expected_task_state'] = expected_task_state if expected_vm_state is not None: updates['expected_vm_state'] = expected_vm_state expected_attrs = [attr for attr in _INSTANCE_OPTIONAL_JOINED_FIELDS if self.obj_attr_is_set(attr)] if 'pci_devices' in expected_attrs: # NOTE(danms): We don't refresh pci_devices on save right now expected_attrs.remove('pci_devices') # NOTE(alaski): We need to pull system_metadata for the # notification.send_update() below. If we don't there's a KeyError # when it tries to extract the flavor. if 'system_metadata' not in expected_attrs: expected_attrs.append('system_metadata') old_ref, inst_ref = db.instance_update_and_get_original( context, self.uuid, updates, columns_to_join=_expected_cols(expected_attrs)) self._from_db_object(context, self, inst_ref, expected_attrs=expected_attrs) if cells_update_from_api: _handle_cell_update_from_api() elif cell_type == 'compute': if self._sync_cells: cells_api = cells_rpcapi.CellsAPI() cells_api.instance_update_at_top(context, stale_instance) def _notify(): # NOTE(danms): We have to be super careful here not to trigger # any lazy-loads that will unmigrate or unbackport something. So, # make a copy of the instance for notifications first. new_ref = self.obj_clone() notifications.send_update(context, old_ref, new_ref) # NOTE(alaski): If cell synchronization is blocked it means we have # already run this block of code in either the parent or child of this # cell. Therefore this notification has already been sent. if not self._sync_cells: _notify = lambda: None # noqa: F811 _notify() self.obj_reset_changes() @base.remotable def refresh(self, use_slave=False): extra = [field for field in INSTANCE_OPTIONAL_ATTRS if self.obj_attr_is_set(field)] current = self.__class__.get_by_uuid(self._context, uuid=self.uuid, expected_attrs=extra, use_slave=use_slave) # NOTE(danms): We orphan the instance copy so we do not unexpectedly # trigger a lazy-load (which would mean we failed to calculate the # expected_attrs properly) current._context = None for field in self.fields: if field not in self: continue if field not in current: # If the field isn't in current we should not # touch it, triggering a likely-recursive lazy load. # Log it so we can see it happening though, as it # probably isn't expected in most cases. LOG.debug('Field %s is set but not in refreshed ' 'instance, skipping', field) continue if field == 'info_cache': self.info_cache.refresh() elif self[field] != current[field]: self[field] = current[field] self.obj_reset_changes() def _load_generic(self, attrname): with utils.temporary_mutation(self._context, read_deleted='yes'): instance = self.__class__.get_by_uuid(self._context, uuid=self.uuid, expected_attrs=[attrname]) if attrname not in instance: # NOTE(danms): Never allow us to recursively-load raise exception.ObjectActionError( action='obj_load_attr', reason=_('loading %s requires recursion') % attrname) # NOTE(danms): load anything we don't already have from the # instance we got from the database to make the most of the # performance hit. for field in self.fields: if field in instance and field not in self: setattr(self, field, getattr(instance, field)) def _load_fault(self): self.fault = objects.InstanceFault.get_latest_for_instance( self._context, self.uuid) def _load_numa_topology(self, db_topology=_NO_DATA_SENTINEL): if db_topology is None: self.numa_topology = None elif db_topology is not _NO_DATA_SENTINEL: self.numa_topology = \ objects.InstanceNUMATopology.obj_from_db_obj(self.uuid, db_topology) else: try: self.numa_topology = \ objects.InstanceNUMATopology.get_by_instance_uuid( self._context, self.uuid) except exception.NumaTopologyNotFound: self.numa_topology = None def _load_pci_requests(self, db_requests=_NO_DATA_SENTINEL): if db_requests is not _NO_DATA_SENTINEL: self.pci_requests = objects.InstancePCIRequests.obj_from_db( self._context, self.uuid, db_requests) else: self.pci_requests = \ objects.InstancePCIRequests.get_by_instance_uuid( self._context, self.uuid) def _load_device_metadata(self, db_dev_meta=_NO_DATA_SENTINEL): if db_dev_meta is None: self.device_metadata = None elif db_dev_meta is not _NO_DATA_SENTINEL: self.device_metadata = \ objects.InstanceDeviceMetadata.obj_from_db( self._context, db_dev_meta) else: self.device_metadata = \ objects.InstanceDeviceMetadata.get_by_instance_uuid( self._context, self.uuid) def _load_flavor(self): instance = self.__class__.get_by_uuid( self._context, uuid=self.uuid, expected_attrs=['flavor']) # NOTE(danms): Orphan the instance to make sure we don't lazy-load # anything below instance._context = None self.flavor = instance.flavor self.old_flavor = instance.old_flavor self.new_flavor = instance.new_flavor def _load_vcpu_model(self, db_vcpu_model=_NO_DATA_SENTINEL): if db_vcpu_model is None: self.vcpu_model = None elif db_vcpu_model is _NO_DATA_SENTINEL: self.vcpu_model = objects.VirtCPUModel.get_by_instance_uuid( self._context, self.uuid) else: db_vcpu_model = jsonutils.loads(db_vcpu_model) self.vcpu_model = objects.VirtCPUModel.obj_from_primitive( db_vcpu_model) def _load_ec2_ids(self): self.ec2_ids = objects.EC2Ids.get_by_instance(self._context, self) def _load_security_groups(self): self.security_groups = objects.SecurityGroupList.get_by_instance( self._context, self) def _load_pci_devices(self): self.pci_devices = objects.PciDeviceList.get_by_instance_uuid( self._context, self.uuid) def _load_migration_context(self, db_context=_NO_DATA_SENTINEL): if db_context is _NO_DATA_SENTINEL: try: self.migration_context = ( objects.MigrationContext.get_by_instance_uuid( self._context, self.uuid)) except exception.MigrationContextNotFound: self.migration_context = None elif db_context is None: self.migration_context = None else: self.migration_context = objects.MigrationContext.obj_from_db_obj( db_context) def _load_keypairs(self, db_keypairs=_NO_DATA_SENTINEL): if db_keypairs is _NO_DATA_SENTINEL: inst = objects.Instance.get_by_uuid(self._context, self.uuid, expected_attrs=['keypairs']) if 'keypairs' in inst: self.keypairs = inst.keypairs self.keypairs.obj_reset_changes(recursive=True) self.obj_reset_changes(['keypairs']) else: self.keypairs = objects.KeyPairList(objects=[]) # NOTE(danms): We leave the keypairs attribute dirty in hopes # someone else will save it for us elif db_keypairs: self.keypairs = objects.KeyPairList.obj_from_primitive( jsonutils.loads(db_keypairs)) self.obj_reset_changes(['keypairs']) def _load_tags(self): self.tags = objects.TagList.get_by_resource_id( self._context, self.uuid) def _load_trusted_certs(self, db_trusted_certs=_NO_DATA_SENTINEL): if db_trusted_certs is None: self.trusted_certs = None elif db_trusted_certs is _NO_DATA_SENTINEL: self.trusted_certs = objects.TrustedCerts.get_by_instance_uuid( self._context, self.uuid) else: self.trusted_certs = objects.TrustedCerts.obj_from_primitive( jsonutils.loads(db_trusted_certs)) def apply_migration_context(self): if self.migration_context: self._set_migration_context_to_instance(prefix='new_') else: LOG.debug("Trying to apply a migration context that does not " "seem to be set for this instance", instance=self) def revert_migration_context(self): if self.migration_context: self._set_migration_context_to_instance(prefix='old_') else: LOG.debug("Trying to revert a migration context that does not " "seem to be set for this instance", instance=self) def _set_migration_context_to_instance(self, prefix): for inst_attr_name in _MIGRATION_CONTEXT_ATTRS: setattr(self, inst_attr_name, None) attr_name = prefix + inst_attr_name if attr_name in self.migration_context: attr_value = getattr( self.migration_context, attr_name) setattr(self, inst_attr_name, attr_value) @contextlib.contextmanager def mutated_migration_context(self): """Context manager to temporarily apply the migration context. Calling .save() from within the context manager means that the mutated context will be saved which can cause incorrect resource tracking, and should be avoided. """ # First check to see if we even have a migration context set and if not # we can exit early without lazy-loading other attributes. if 'migration_context' in self and self.migration_context is None: yield return current_values = {} for attr_name in _MIGRATION_CONTEXT_ATTRS: current_values[attr_name] = getattr(self, attr_name) self.apply_migration_context() try: yield finally: for attr_name in _MIGRATION_CONTEXT_ATTRS: setattr(self, attr_name, current_values[attr_name]) @base.remotable def drop_migration_context(self): if self.migration_context: db.instance_extra_update_by_uuid(self._context, self.uuid, {'migration_context': None}) self.migration_context = None def clear_numa_topology(self): numa_topology = self.numa_topology if numa_topology is not None: self.numa_topology = numa_topology.clear_host_pinning() def obj_load_attr(self, attrname): # NOTE(danms): We can't lazy-load anything without a context and a uuid if not self._context: raise exception.OrphanedObjectError(method='obj_load_attr', objtype=self.obj_name()) if 'uuid' not in self: raise exception.ObjectActionError( action='obj_load_attr', reason=_('attribute %s not lazy-loadable') % attrname) LOG.debug("Lazy-loading '%(attr)s' on %(name)s uuid %(uuid)s", {'attr': attrname, 'name': self.obj_name(), 'uuid': self.uuid, }) # NOTE(danms): We handle some fields differently here so that we # can be more efficient if attrname == 'fault': self._load_fault() elif attrname == 'numa_topology': self._load_numa_topology() elif attrname == 'device_metadata': self._load_device_metadata() elif attrname == 'pci_requests': self._load_pci_requests() elif attrname == 'vcpu_model': self._load_vcpu_model() elif attrname == 'ec2_ids': self._load_ec2_ids() elif attrname == 'migration_context': self._load_migration_context() elif attrname == 'keypairs': # NOTE(danms): Let keypairs control its own destiny for # resetting changes. return self._load_keypairs() elif attrname == 'trusted_certs': return self._load_trusted_certs() elif attrname == 'security_groups': self._load_security_groups() elif attrname == 'pci_devices': self._load_pci_devices() elif 'flavor' in attrname: self._load_flavor() elif attrname == 'services' and self.deleted: # NOTE(mriedem): The join in the data model for instances.services # filters on instances.deleted == 0, so if the instance is deleted # don't attempt to even load services since we'll fail. self.services = objects.ServiceList(self._context) elif attrname == 'tags': if self.deleted: # NOTE(mriedem): Same story as services, the DB API query # in instance_tag_get_by_instance_uuid will fail if the # instance has been deleted so just return an empty tag list. self.tags = objects.TagList(self._context) else: self._load_tags() elif attrname in self.fields and attrname != 'id': # NOTE(danms): We've never let 'id' be lazy-loaded, and use its # absence as a sentinel that it hasn't been created in the database # yet, so refuse to do so here. self._load_generic(attrname) else: # NOTE(danms): This is historically what we did for # something not in a field that was force-loaded. So, just # do this for consistency. raise exception.ObjectActionError( action='obj_load_attr', reason=_('attribute %s not lazy-loadable') % attrname) self.obj_reset_changes([attrname]) def get_flavor(self, namespace=None): prefix = ('%s_' % namespace) if namespace is not None else '' attr = '%sflavor' % prefix try: return getattr(self, attr) except exception.FlavorNotFound: # NOTE(danms): This only happens in the case where we don't # have flavor information in instance_extra, and doing # this triggers a lookup based on our instance_type_id for # (very) legacy instances. That legacy code expects a None here, # so emulate it for this helper, even though the actual attribute # is not nullable. return None @base.remotable def delete_metadata_key(self, key): """Optimized metadata delete method. This provides a more efficient way to delete a single metadata key, instead of just calling instance.save(). This should be called with the key still present in self.metadata, which it will update after completion. """ db.instance_metadata_delete(self._context, self.uuid, key) md_was_changed = 'metadata' in self.obj_what_changed() del self.metadata[key] self._orig_metadata.pop(key, None) notifications.send_update(self._context, self, self) if not md_was_changed: self.obj_reset_changes(['metadata']) def _cell_name_blocks_sync(self): if (self.obj_attr_is_set('cell_name') and self.cell_name is not None and self.cell_name.startswith(cells_utils.BLOCK_SYNC_FLAG)): return True return False def _normalize_cell_name(self): """Undo skip_cell_sync()'s cell_name modification if applied""" if not self.obj_attr_is_set('cell_name') or self.cell_name is None: return cn_changed = 'cell_name' in self.obj_what_changed() if self.cell_name.startswith(cells_utils.BLOCK_SYNC_FLAG): self.cell_name = self.cell_name.replace( cells_utils.BLOCK_SYNC_FLAG, '', 1) # cell_name is not normally an empty string, this means it was None # or unset before cells_utils.BLOCK_SYNC_FLAG was applied. if len(self.cell_name) == 0: self.cell_name = None if not cn_changed: self.obj_reset_changes(['cell_name']) @contextlib.contextmanager def skip_cells_sync(self): """Context manager to save an instance without syncing cells. Temporarily disables the cells syncing logic, if enabled. This should only be used when saving an instance that has been passed down/up from another cell in order to avoid passing it back to the originator to be re-saved. """ cn_changed = 'cell_name' in self.obj_what_changed() if not self.obj_attr_is_set('cell_name') or self.cell_name is None: self.cell_name = '' self.cell_name = '%s%s' % (cells_utils.BLOCK_SYNC_FLAG, self.cell_name) if not cn_changed: self.obj_reset_changes(['cell_name']) try: yield finally: self._normalize_cell_name() def get_network_info(self): if self.info_cache is None: return network_model.NetworkInfo.hydrate([]) return self.info_cache.network_info def get_bdms(self): return objects.BlockDeviceMappingList.get_by_instance_uuid( self._context, self.uuid) def _make_instance_list(context, inst_list, db_inst_list, expected_attrs): get_fault = expected_attrs and 'fault' in expected_attrs inst_faults = {} if get_fault: # Build an instance_uuid:latest-fault mapping expected_attrs.remove('fault') instance_uuids = [inst['uuid'] for inst in db_inst_list] faults = objects.InstanceFaultList.get_by_instance_uuids( context, instance_uuids) for fault in faults: if fault.instance_uuid not in inst_faults: inst_faults[fault.instance_uuid] = fault inst_cls = objects.Instance inst_list.objects = [] for db_inst in db_inst_list: inst_obj = inst_cls._from_db_object( context, inst_cls(context), db_inst, expected_attrs=expected_attrs) if get_fault: inst_obj.fault = inst_faults.get(inst_obj.uuid, None) inst_list.objects.append(inst_obj) inst_list.obj_reset_changes() return inst_list @db_api.pick_context_manager_writer def populate_missing_availability_zones(context, count): # instances without host have no reasonable AZ to set not_empty_host = models.Instance.host != None # noqa E711 instances = (context.session.query(models.Instance). filter(not_empty_host). filter_by(availability_zone=None).limit(count).all()) count_all = len(instances) count_hit = 0 for instance in instances: az = avail_zone.get_instance_availability_zone(context, instance) instance.availability_zone = az instance.save(context.session) count_hit += 1 return count_all, count_hit @base.NovaObjectRegistry.register class InstanceList(base.ObjectListBase, base.NovaObject): # Version 2.0: Initial Version # Version 2.1: Add get_uuids_by_host() # Version 2.2: Pagination for get_active_by_window_joined() # Version 2.3: Add get_count_by_vm_state() # Version 2.4: Add get_counts() VERSION = '2.4' fields = { 'objects': fields.ListOfObjectsField('Instance'), } @classmethod @db.select_db_reader_mode def _get_by_filters_impl(cls, context, filters, sort_key='created_at', sort_dir='desc', limit=None, marker=None, expected_attrs=None, use_slave=False, sort_keys=None, sort_dirs=None): if sort_keys or sort_dirs: db_inst_list = db.instance_get_all_by_filters_sort( context, filters, limit=limit, marker=marker, columns_to_join=_expected_cols(expected_attrs), sort_keys=sort_keys, sort_dirs=sort_dirs) else: db_inst_list = db.instance_get_all_by_filters( context, filters, sort_key, sort_dir, limit=limit, marker=marker, columns_to_join=_expected_cols(expected_attrs)) return db_inst_list @base.remotable_classmethod def get_by_filters(cls, context, filters, sort_key='created_at', sort_dir='desc', limit=None, marker=None, expected_attrs=None, use_slave=False, sort_keys=None, sort_dirs=None): db_inst_list = cls._get_by_filters_impl( context, filters, sort_key=sort_key, sort_dir=sort_dir, limit=limit, marker=marker, expected_attrs=expected_attrs, use_slave=use_slave, sort_keys=sort_keys, sort_dirs=sort_dirs) # NOTE(melwitt): _make_instance_list could result in joined objects' # (from expected_attrs) _from_db_object methods being called during # Instance._from_db_object, each of which might choose to perform # database writes. So, we call this outside of _get_by_filters_impl to # avoid being nested inside a 'reader' database transaction context. return _make_instance_list(context, cls(), db_inst_list, expected_attrs) @staticmethod @db.select_db_reader_mode def _db_instance_get_all_by_host(context, host, columns_to_join, use_slave=False): return db.instance_get_all_by_host(context, host, columns_to_join=columns_to_join) @base.remotable_classmethod def get_by_host(cls, context, host, expected_attrs=None, use_slave=False): db_inst_list = cls._db_instance_get_all_by_host( context, host, columns_to_join=_expected_cols(expected_attrs), use_slave=use_slave) return _make_instance_list(context, cls(), db_inst_list, expected_attrs) @base.remotable_classmethod def get_by_host_and_node(cls, context, host, node, expected_attrs=None): db_inst_list = db.instance_get_all_by_host_and_node( context, host, node, columns_to_join=_expected_cols(expected_attrs)) return _make_instance_list(context, cls(), db_inst_list, expected_attrs) @base.remotable_classmethod def get_by_host_and_not_type(cls, context, host, type_id=None, expected_attrs=None): db_inst_list = db.instance_get_all_by_host_and_not_type( context, host, type_id=type_id) return _make_instance_list(context, cls(), db_inst_list, expected_attrs) @base.remotable_classmethod def get_all(cls, context, expected_attrs=None): """Returns all instances on all nodes.""" db_instances = db.instance_get_all( context, columns_to_join=_expected_cols(expected_attrs)) return _make_instance_list(context, cls(), db_instances, expected_attrs) @base.remotable_classmethod def get_hung_in_rebooting(cls, context, reboot_window, expected_attrs=None): db_inst_list = db.instance_get_all_hung_in_rebooting(context, reboot_window) return _make_instance_list(context, cls(), db_inst_list, expected_attrs) @staticmethod @db.select_db_reader_mode def _db_instance_get_active_by_window_joined( context, begin, end, project_id, host, columns_to_join, use_slave=False, limit=None, marker=None): return db.instance_get_active_by_window_joined( context, begin, end, project_id, host, columns_to_join=columns_to_join, limit=limit, marker=marker) @base.remotable_classmethod def _get_active_by_window_joined(cls, context, begin, end=None, project_id=None, host=None, expected_attrs=None, use_slave=False, limit=None, marker=None): # NOTE(mriedem): We need to convert the begin/end timestamp strings # to timezone-aware datetime objects for the DB API call. begin = timeutils.parse_isotime(begin) end = timeutils.parse_isotime(end) if end else None db_inst_list = cls._db_instance_get_active_by_window_joined( context, begin, end, project_id, host, columns_to_join=_expected_cols(expected_attrs), use_slave=use_slave, limit=limit, marker=marker) return _make_instance_list(context, cls(), db_inst_list, expected_attrs) @classmethod def get_active_by_window_joined(cls, context, begin, end=None, project_id=None, host=None, expected_attrs=None, use_slave=False, limit=None, marker=None): """Get instances and joins active during a certain time window. :param:context: nova request context :param:begin: datetime for the start of the time window :param:end: datetime for the end of the time window :param:project_id: used to filter instances by project :param:host: used to filter instances on a given compute host :param:expected_attrs: list of related fields that can be joined in the database layer when querying for instances :param use_slave if True, ship this query off to a DB slave :param limit: maximum number of instances to return per page :param marker: last instance uuid from the previous page :returns: InstanceList """ # NOTE(mriedem): We have to convert the datetime objects to string # primitives for the remote call. begin = utils.isotime(begin) end = utils.isotime(end) if end else None return cls._get_active_by_window_joined(context, begin, end, project_id, host, expected_attrs, use_slave=use_slave, limit=limit, marker=marker) @base.remotable_classmethod def get_by_security_group_id(cls, context, security_group_id): db_secgroup = db.security_group_get( context, security_group_id, columns_to_join=['instances.info_cache', 'instances.system_metadata']) return _make_instance_list(context, cls(), db_secgroup['instances'], ['info_cache', 'system_metadata']) @classmethod def get_by_security_group(cls, context, security_group): return cls.get_by_security_group_id(context, security_group.id) @base.remotable_classmethod def get_by_grantee_security_group_ids(cls, context, security_group_ids): db_instances = db.instance_get_all_by_grantee_security_groups( context, security_group_ids) return _make_instance_list(context, cls(), db_instances, []) def fill_faults(self): """Batch query the database for our instances' faults. :returns: A list of instance uuids for which faults were found. """ uuids = [inst.uuid for inst in self] faults = objects.InstanceFaultList.get_latest_by_instance_uuids( self._context, uuids) faults_by_uuid = {} for fault in faults: faults_by_uuid[fault.instance_uuid] = fault for instance in self: if instance.uuid in faults_by_uuid: instance.fault = faults_by_uuid[instance.uuid] else: # NOTE(danms): Otherwise the caller will cause a lazy-load # when checking it, and we know there are none instance.fault = None instance.obj_reset_changes(['fault']) return faults_by_uuid.keys() @base.remotable_classmethod def get_uuids_by_host(cls, context, host): return db.instance_get_all_uuids_by_host(context, host) @staticmethod @db_api.pick_context_manager_reader def _get_count_by_vm_state_in_db(context, project_id, user_id, vm_state): return context.session.query(models.Instance.id).\ filter_by(deleted=0).\ filter_by(project_id=project_id).\ filter_by(user_id=user_id).\ filter_by(vm_state=vm_state).\ count() @base.remotable_classmethod def get_count_by_vm_state(cls, context, project_id, user_id, vm_state): return cls._get_count_by_vm_state_in_db(context, project_id, user_id, vm_state) @staticmethod @db_api.pick_context_manager_reader def _get_counts_in_db(context, project_id, user_id=None): # NOTE(melwitt): Copied from nova/db/sqlalchemy/api.py: # It would be better to have vm_state not be nullable # but until then we test it explicitly as a workaround. not_soft_deleted = or_( models.Instance.vm_state != vm_states.SOFT_DELETED, models.Instance.vm_state == null() ) project_query = context.session.query( func.count(models.Instance.id), func.sum(models.Instance.vcpus), func.sum(models.Instance.memory_mb)).\ filter_by(deleted=0).\ filter(not_soft_deleted).\ filter_by(project_id=project_id) project_result = project_query.first() fields = ('instances', 'cores', 'ram') project_counts = {field: int(project_result[idx] or 0) for idx, field in enumerate(fields)} counts = {'project': project_counts} if user_id: user_result = project_query.filter_by(user_id=user_id).first() user_counts = {field: int(user_result[idx] or 0) for idx, field in enumerate(fields)} counts['user'] = user_counts return counts @base.remotable_classmethod def get_counts(cls, context, project_id, user_id=None): """Get the counts of Instance objects in the database. :param context: The request context for database access :param project_id: The project_id to count across :param user_id: The user_id to count across :returns: A dict containing the project-scoped counts and user-scoped counts if user_id is specified. For example: {'project': {'instances': , 'cores': , 'ram': , 'cores': , 'ram': }} """ return cls._get_counts_in_db(context, project_id, user_id=user_id) @staticmethod @db_api.pick_context_manager_reader def _get_count_by_hosts(context, hosts): return context.session.query(models.Instance).\ filter_by(deleted=0).\ filter(models.Instance.host.in_(hosts)).count() @classmethod def get_count_by_hosts(cls, context, hosts): return cls._get_count_by_hosts(context, hosts)