nova/nova/virt/fake.py

501 lines
16 KiB
Python

# Copyright 2010 United States Government as represented by the
# Administrator of the National Aeronautics and Space Administration.
# All Rights Reserved.
# Copyright (c) 2010 Citrix Systems, Inc.
#
# 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.
"""
A fake (in-memory) hypervisor+api.
Allows nova testing w/o a hypervisor. This module also documents the
semantics of real hypervisor connections.
"""
import contextlib
from oslo.config import cfg
from oslo.serialization import jsonutils
from nova.compute import arch
from nova.compute import hvtype
from nova.compute import power_state
from nova.compute import task_states
from nova.compute import vm_mode
from nova.console import type as ctype
from nova import db
from nova import exception
from nova.i18n import _
from nova.openstack.common import log as logging
from nova import utils
from nova.virt import diagnostics
from nova.virt import driver
from nova.virt import virtapi
CONF = cfg.CONF
CONF.import_opt('host', 'nova.netconf')
LOG = logging.getLogger(__name__)
_FAKE_NODES = None
def set_nodes(nodes):
"""Sets FakeDriver's node.list.
It has effect on the following methods:
get_available_nodes()
get_available_resource
To restore the change, call restore_nodes()
"""
global _FAKE_NODES
_FAKE_NODES = nodes
def restore_nodes():
"""Resets FakeDriver's node list modified by set_nodes().
Usually called from tearDown().
"""
global _FAKE_NODES
_FAKE_NODES = [CONF.host]
class FakeInstance(object):
def __init__(self, name, state, uuid):
self.name = name
self.state = state
self.uuid = uuid
def __getitem__(self, key):
return getattr(self, key)
class FakeDriver(driver.ComputeDriver):
capabilities = {
"has_imagecache": True,
"supports_recreate": True,
}
# Since we don't have a real hypervisor, pretend we have lots of
# disk and ram so this driver can be used to test large instances.
vcpus = 1000
memory_mb = 800000
local_gb = 600000
"""Fake hypervisor driver."""
def __init__(self, virtapi, read_only=False):
super(FakeDriver, self).__init__(virtapi)
self.instances = {}
self.host_status_base = {
'vcpus': self.vcpus,
'memory_mb': self.memory_mb,
'local_gb': self.local_gb,
'vcpus_used': 0,
'memory_mb_used': 0,
'local_gb_used': 100000000000,
'hypervisor_type': 'fake',
'hypervisor_version': utils.convert_version_to_int('1.0'),
'hypervisor_hostname': CONF.host,
'cpu_info': {},
'disk_available_least': 0,
'supported_instances': jsonutils.dumps([(arch.X86_64,
hvtype.FAKE,
vm_mode.HVM)]),
'numa_topology': None,
}
self._mounts = {}
self._interfaces = {}
if not _FAKE_NODES:
set_nodes([CONF.host])
def init_host(self, host):
return
def list_instances(self):
return self.instances.keys()
def list_instance_uuids(self):
return [self.instances[name].uuid for name in self.instances.keys()]
def plug_vifs(self, instance, network_info):
"""Plug VIFs into networks."""
pass
def unplug_vifs(self, instance, network_info):
"""Unplug VIFs from networks."""
pass
def spawn(self, context, instance, image_meta, injected_files,
admin_password, network_info=None, block_device_info=None):
name = instance['name']
state = power_state.RUNNING
fake_instance = FakeInstance(name, state, instance['uuid'])
self.instances[name] = fake_instance
def snapshot(self, context, instance, name, update_task_state):
if instance['name'] not in self.instances:
raise exception.InstanceNotRunning(instance_id=instance['uuid'])
update_task_state(task_state=task_states.IMAGE_UPLOADING)
def reboot(self, context, instance, network_info, reboot_type,
block_device_info=None, bad_volumes_callback=None):
pass
@staticmethod
def get_host_ip_addr():
return '192.168.0.1'
def set_admin_password(self, instance, new_pass):
pass
def inject_file(self, instance, b64_path, b64_contents):
pass
def resume_state_on_host_boot(self, context, instance, network_info,
block_device_info=None):
pass
def rescue(self, context, instance, network_info, image_meta,
rescue_password):
pass
def unrescue(self, instance, network_info):
pass
def poll_rebooting_instances(self, timeout, instances):
pass
def migrate_disk_and_power_off(self, context, instance, dest,
flavor, network_info,
block_device_info=None,
timeout=0, retry_interval=0):
pass
def finish_revert_migration(self, context, instance, network_info,
block_device_info=None, power_on=True):
pass
def post_live_migration_at_destination(self, context, instance,
network_info,
block_migration=False,
block_device_info=None):
pass
def power_off(self, instance, shutdown_timeout=0, shutdown_attempts=0):
pass
def power_on(self, context, instance, network_info, block_device_info):
pass
def soft_delete(self, instance):
pass
def restore(self, instance):
pass
def pause(self, instance):
pass
def unpause(self, instance):
pass
def suspend(self, instance):
pass
def resume(self, context, instance, network_info, block_device_info=None):
pass
def destroy(self, context, instance, network_info, block_device_info=None,
destroy_disks=True, migrate_data=None):
key = instance['name']
if key in self.instances:
del self.instances[key]
else:
LOG.warning(_("Key '%(key)s' not in instances '%(inst)s'") %
{'key': key,
'inst': self.instances}, instance=instance)
def cleanup(self, context, instance, network_info, block_device_info=None,
destroy_disks=True, migrate_data=None, destroy_vifs=True):
pass
def attach_volume(self, context, connection_info, instance, mountpoint,
disk_bus=None, device_type=None, encryption=None):
"""Attach the disk to the instance at mountpoint using info."""
instance_name = instance['name']
if instance_name not in self._mounts:
self._mounts[instance_name] = {}
self._mounts[instance_name][mountpoint] = connection_info
def detach_volume(self, connection_info, instance, mountpoint,
encryption=None):
"""Detach the disk attached to the instance."""
try:
del self._mounts[instance['name']][mountpoint]
except KeyError:
pass
def swap_volume(self, old_connection_info, new_connection_info,
instance, mountpoint, resize_to):
"""Replace the disk attached to the instance."""
instance_name = instance['name']
if instance_name not in self._mounts:
self._mounts[instance_name] = {}
self._mounts[instance_name][mountpoint] = new_connection_info
def attach_interface(self, instance, image_meta, vif):
if vif['id'] in self._interfaces:
raise exception.InterfaceAttachFailed(
instance_uuid=instance['uuid'])
self._interfaces[vif['id']] = vif
def detach_interface(self, instance, vif):
try:
del self._interfaces[vif['id']]
except KeyError:
raise exception.InterfaceDetachFailed(
instance_uuid=instance['uuid'])
def get_info(self, instance):
if instance['name'] not in self.instances:
raise exception.InstanceNotFound(instance_id=instance['name'])
i = self.instances[instance['name']]
return {'state': i.state,
'max_mem': 0,
'mem': 0,
'num_cpu': 2,
'cpu_time': 0}
def get_diagnostics(self, instance_name):
return {'cpu0_time': 17300000000,
'memory': 524288,
'vda_errors': -1,
'vda_read': 262144,
'vda_read_req': 112,
'vda_write': 5778432,
'vda_write_req': 488,
'vnet1_rx': 2070139,
'vnet1_rx_drop': 0,
'vnet1_rx_errors': 0,
'vnet1_rx_packets': 26701,
'vnet1_tx': 140208,
'vnet1_tx_drop': 0,
'vnet1_tx_errors': 0,
'vnet1_tx_packets': 662,
}
def get_instance_diagnostics(self, instance_name):
diags = diagnostics.Diagnostics(state='running', driver='fake',
hypervisor_os='fake-os', uptime=46664, config_drive=True)
diags.add_cpu(time=17300000000)
diags.add_nic(mac_address='01:23:45:67:89:ab',
rx_packets=26701,
rx_octets=2070139,
tx_octets=140208,
tx_packets = 662)
diags.add_disk(id='fake-disk-id',
read_bytes=262144,
read_requests=112,
write_bytes=5778432,
write_requests=488)
diags.memory_details.maximum = 524288
return diags
def get_all_bw_counters(self, instances):
"""Return bandwidth usage counters for each interface on each
running VM.
"""
bw = []
return bw
def get_all_volume_usage(self, context, compute_host_bdms):
"""Return usage info for volumes attached to vms on
a given host.
"""
volusage = []
return volusage
def get_host_cpu_stats(self):
stats = {'kernel': 5664160000000L,
'idle': 1592705190000000L,
'user': 26728850000000L,
'iowait': 6121490000000L}
stats['frequency'] = 800
return stats
def block_stats(self, instance_name, disk_id):
return [0L, 0L, 0L, 0L, None]
def interface_stats(self, instance_name, iface_id):
return [0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L]
def get_console_output(self, context, instance):
return 'FAKE CONSOLE OUTPUT\nANOTHER\nLAST LINE'
def get_vnc_console(self, context, instance):
return ctype.ConsoleVNC(internal_access_path='FAKE',
host='fakevncconsole.com',
port=6969)
def get_spice_console(self, context, instance):
return ctype.ConsoleSpice(internal_access_path='FAKE',
host='fakespiceconsole.com',
port=6969,
tlsPort=6970)
def get_rdp_console(self, context, instance):
return ctype.ConsoleRDP(internal_access_path='FAKE',
host='fakerdpconsole.com',
port=6969)
def get_serial_console(self, context, instance):
return ctype.ConsoleSerial(internal_access_path='FAKE',
host='fakerdpconsole.com',
port=6969)
def get_console_pool_info(self, console_type):
return {'address': '127.0.0.1',
'username': 'fakeuser',
'password': 'fakepassword'}
def refresh_security_group_rules(self, security_group_id):
return True
def refresh_security_group_members(self, security_group_id):
return True
def refresh_instance_security_rules(self, instance):
return True
def refresh_provider_fw_rules(self):
pass
def get_available_resource(self, nodename):
"""Updates compute manager resource info on ComputeNode table.
Since we don't have a real hypervisor, pretend we have lots of
disk and ram.
"""
if nodename not in _FAKE_NODES:
return {}
host_status = self.host_status_base.copy()
host_status['hypervisor_hostname'] = nodename
host_status['host_hostname'] = nodename
host_status['host_name_label'] = nodename
host_status['cpu_info'] = '?'
return host_status
def ensure_filtering_rules_for_instance(self, instance_ref, network_info):
return
def get_instance_disk_info(self, instance_name, block_device_info=None):
return
def live_migration(self, context, instance_ref, dest,
post_method, recover_method, block_migration=False,
migrate_data=None):
post_method(context, instance_ref, dest, block_migration,
migrate_data)
return
def check_can_live_migrate_destination_cleanup(self, ctxt,
dest_check_data):
return
def check_can_live_migrate_destination(self, ctxt, instance_ref,
src_compute_info, dst_compute_info,
block_migration=False,
disk_over_commit=False):
return {}
def check_can_live_migrate_source(self, ctxt, instance_ref,
dest_check_data, block_device_info=None):
return
def finish_migration(self, context, migration, instance, disk_info,
network_info, image_meta, resize_instance,
block_device_info=None, power_on=True):
return
def confirm_migration(self, migration, instance, network_info):
return
def pre_live_migration(self, context, instance_ref, block_device_info,
network_info, disk, migrate_data=None):
return
def unfilter_instance(self, instance_ref, network_info):
return
def test_remove_vm(self, instance_name):
"""Removes the named VM, as if it crashed. For testing."""
self.instances.pop(instance_name)
def host_power_action(self, host, action):
"""Reboots, shuts down or powers up the host."""
return action
def host_maintenance_mode(self, host, mode):
"""Start/Stop host maintenance window. On start, it triggers
guest VMs evacuation.
"""
if not mode:
return 'off_maintenance'
return 'on_maintenance'
def set_host_enabled(self, host, enabled):
"""Sets the specified host's ability to accept new instances."""
if enabled:
return 'enabled'
return 'disabled'
def get_volume_connector(self, instance):
return {'ip': '127.0.0.1', 'initiator': 'fake', 'host': 'fakehost'}
def get_available_nodes(self, refresh=False):
return _FAKE_NODES
def instance_on_disk(self, instance):
return False
class FakeVirtAPI(virtapi.VirtAPI):
def provider_fw_rule_get_all(self, context):
return db.provider_fw_rule_get_all(context)
@contextlib.contextmanager
def wait_for_instance_event(self, instance, event_names, deadline=300,
error_callback=None):
# NOTE(danms): Don't actually wait for any events, just
# fall through
yield
class SmallFakeDriver(FakeDriver):
# The api samples expect specific cpu memory and disk sizes. In order to
# allow the FakeVirt driver to be used outside of the unit tests, provide
# a separate class that has the values expected by the api samples. So
# instead of requiring new samples every time those
# values are adjusted allow them to be overwritten here.
vcpus = 1
memory_mb = 8192
local_gb = 1028