forked from netgroup/vim-tuning-and-eval-tools
-
Notifications
You must be signed in to change notification settings - Fork 0
/
driver_tuned_v1.py
7570 lines (6489 loc) · 329 KB
/
driver_tuned_v1.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
# 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.
# Copyright (c) 2011 Piston Cloud Computing, Inc
# Copyright (c) 2012 University Of Minho
# (c) Copyright 2013 Hewlett-Packard Development Company, L.P.
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
# a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
# License for the specific language governing permissions and limitations
# under the License.
"""
A connection to a hypervisor through libvirt.
Supports KVM, LXC, QEMU, UML, XEN and Parallels.
"""
import collections
import contextlib
import errno
import functools
import glob
import itertools
import mmap
import operator
import os
import shutil
import tempfile
import time
import uuid
import random
import eventlet
from eventlet import greenthread
from eventlet import tpool
from lxml import etree
from os_brick.initiator import connector
from oslo_concurrency import processutils
from oslo_config import cfg
from oslo_log import log as logging
from oslo_serialization import jsonutils
from oslo_service import loopingcall
from oslo_utils import excutils
from oslo_utils import fileutils
from oslo_utils import importutils
from oslo_utils import strutils
from oslo_utils import timeutils
from oslo_utils import units
import six
from six.moves import range
from nova.api.metadata import base as instance_metadata
from nova import block_device
from nova.compute import arch
from nova.compute import hv_type
from nova.compute import power_state
from nova.compute import task_states
from nova.compute import utils as compute_utils
from nova.compute import vm_mode
from nova.console import serial as serial_console
from nova.console import type as ctype
from nova import context as nova_context
from nova import exception
from nova.i18n import _
from nova.i18n import _LE
from nova.i18n import _LI
from nova.i18n import _LW
from nova import image
from nova.network import model as network_model
from nova import objects
from nova.objects import fields
from nova.pci import manager as pci_manager
from nova.pci import utils as pci_utils
from nova import utils
from nova import version
from nova.virt import block_device as driver_block_device
from nova.virt import configdrive
from nova.virt import diagnostics
from nova.virt.disk import api as disk
from nova.virt.disk.vfs import guestfs
from nova.virt import driver
from nova.virt import firewall
from nova.virt import hardware
from nova.virt.image import model as imgmodel
from nova.virt import images
from nova.virt.libvirt import blockinfo
from nova.virt.libvirt import config as vconfig
from nova.virt.libvirt import firewall as libvirt_firewall
from nova.virt.libvirt import guest as libvirt_guest
from nova.virt.libvirt import host
from nova.virt.libvirt import imagebackend
from nova.virt.libvirt import imagecache
from nova.virt.libvirt import instancejobtracker
from nova.virt.libvirt.storage import dmcrypt
from nova.virt.libvirt.storage import lvm
from nova.virt.libvirt.storage import rbd_utils
from nova.virt.libvirt import utils as libvirt_utils
from nova.virt.libvirt import vif as libvirt_vif
from nova.virt.libvirt.volume import remotefs
from nova.virt import netutils
from nova.virt import watchdog_actions
from nova import volume
from nova.volume import encryptors
from lxml import etree
libvirt = None
LOG = logging.getLogger(__name__)
image_uuid = '18eae1ae-76e4-4a6c-8bd8-382e599928c2'
# Downtime period in milliseconds
LIVE_MIGRATION_DOWNTIME_MIN = 100
# Step count
LIVE_MIGRATION_DOWNTIME_STEPS_MIN = 3
# Delay in seconds
LIVE_MIGRATION_DOWNTIME_DELAY_MIN = 10
libvirt_opts = [
cfg.StrOpt('rescue_image_id',
help='Rescue ami image. This will not be used if an image id '
'is provided by the user.'),
cfg.StrOpt('rescue_kernel_id',
help='Rescue aki image'),
cfg.StrOpt('rescue_ramdisk_id',
help='Rescue ari image'),
cfg.StrOpt('virt_type',
default='kvm',
choices=('kvm', 'lxc', 'qemu', 'uml', 'xen', 'parallels'),
help='Libvirt domain type'),
cfg.StrOpt('connection_uri',
default='',
help='Override the default libvirt URI '
'(which is dependent on virt_type)'),
cfg.BoolOpt('inject_password',
default=False,
help='Inject the admin password at boot time, '
'without an agent.'),
cfg.BoolOpt('inject_key',
default=False,
help='Inject the ssh public key at boot time'),
cfg.IntOpt('inject_partition',
default=-2,
help='The partition to inject to : '
'-2 => disable, -1 => inspect (libguestfs only), '
'0 => not partitioned, >0 => partition number'),
cfg.BoolOpt('use_usb_tablet',
default=True,
help='Sync virtual and real mouse cursors in Windows VMs'),
cfg.StrOpt('live_migration_uri',
default="qemu+tcp://%s/system",
help='Migration target URI '
'(any included "%s" is replaced with '
'the migration target hostname)'),
cfg.StrOpt('live_migration_flag',
default='VIR_MIGRATE_UNDEFINE_SOURCE, VIR_MIGRATE_PEER2PEER, '
'VIR_MIGRATE_LIVE, VIR_MIGRATE_TUNNELLED',
help='Migration flags to be set for live migration'),
cfg.StrOpt('block_migration_flag',
default='VIR_MIGRATE_UNDEFINE_SOURCE, VIR_MIGRATE_PEER2PEER, '
'VIR_MIGRATE_LIVE, VIR_MIGRATE_TUNNELLED, '
'VIR_MIGRATE_NON_SHARED_INC',
help='Migration flags to be set for block migration'),
cfg.IntOpt('live_migration_bandwidth',
default=0,
help='Maximum bandwidth(in MiB/s) to be used during migration. '
'If set to 0, will choose a suitable default. Some '
'hypervisors do not support this feature and will return '
'an error if bandwidth is not 0. Please refer to the '
'libvirt documentation for further details'),
cfg.IntOpt('live_migration_downtime',
default=500,
help='Maximum permitted downtime, in milliseconds, for live '
'migration switchover. Will be rounded up to a minimum '
'of %dms. Use a large value if guest liveness is '
'unimportant.' % LIVE_MIGRATION_DOWNTIME_MIN),
cfg.IntOpt('live_migration_downtime_steps',
default=10,
help='Number of incremental steps to reach max downtime value. '
'Will be rounded up to a minimum of %d steps' %
LIVE_MIGRATION_DOWNTIME_STEPS_MIN),
cfg.IntOpt('live_migration_downtime_delay',
default=75,
help='Time to wait, in seconds, between each step increase '
'of the migration downtime. Minimum delay is %d seconds. '
'Value is per GiB of guest RAM + disk to be transferred, '
'with lower bound of a minimum of 2 GiB per device' %
LIVE_MIGRATION_DOWNTIME_DELAY_MIN),
cfg.IntOpt('live_migration_completion_timeout',
default=800,
help='Time to wait, in seconds, for migration to successfully '
'complete transferring data before aborting the '
'operation. Value is per GiB of guest RAM + disk to be '
'transferred, with lower bound of a minimum of 2 GiB. '
'Should usually be larger than downtime delay * downtime '
'steps. Set to 0 to disable timeouts.'),
cfg.IntOpt('live_migration_progress_timeout',
default=150,
help='Time to wait, in seconds, for migration to make forward '
'progress in transferring data before aborting the '
'operation. Set to 0 to disable timeouts.'),
cfg.StrOpt('snapshot_image_format',
choices=('raw', 'qcow2', 'vmdk', 'vdi'),
help='Snapshot image format. Defaults to same as source image'),
cfg.StrOpt('disk_prefix',
help='Override the default disk prefix for the devices attached'
' to a server, which is dependent on virt_type. '
'(valid options are: sd, xvd, uvd, vd)'),
cfg.IntOpt('wait_soft_reboot_seconds',
default=120,
help='Number of seconds to wait for instance to shut down after'
' soft reboot request is made. We fall back to hard reboot'
' if instance does not shutdown within this window.'),
cfg.StrOpt('cpu_mode',
choices=('host-model', 'host-passthrough', 'custom', 'none'),
help='Set to "host-model" to clone the host CPU feature flags; '
'to "host-passthrough" to use the host CPU model exactly; '
'to "custom" to use a named CPU model; '
'to "none" to not set any CPU model. '
'If virt_type="kvm|qemu", it will default to '
'"host-model", otherwise it will default to "none"'),
cfg.StrOpt('cpu_model',
help='Set to a named libvirt CPU model (see names listed '
'in /usr/share/libvirt/cpu_map.xml). Only has effect if '
'cpu_mode="custom" and virt_type="kvm|qemu"'),
cfg.StrOpt('snapshots_directory',
default='$instances_path/snapshots',
help='Location where libvirt driver will store snapshots '
'before uploading them to image service'),
cfg.StrOpt('xen_hvmloader_path',
default='/usr/lib/xen/boot/hvmloader',
help='Location where the Xen hvmloader is kept'),
cfg.ListOpt('disk_cachemodes',
default=[],
help='Specific cachemodes to use for different disk types '
'e.g: file=directsync,block=none'),
cfg.StrOpt('rng_dev_path',
help='A path to a device that will be used as source of '
'entropy on the host. Permitted options are: '
'/dev/random or /dev/hwrng'),
cfg.ListOpt('hw_machine_type',
help='For qemu or KVM guests, set this option to specify '
'a default machine type per host architecture. '
'You can find a list of supported machine types '
'in your environment by checking the output of '
'the "virsh capabilities"command. The format of the '
'value for this config option is host-arch=machine-type. '
'For example: x86_64=machinetype1,armv7l=machinetype2'),
cfg.StrOpt('sysinfo_serial',
default='auto',
choices=('none', 'os', 'hardware', 'auto'),
help='The data source used to the populate the host "serial" '
'UUID exposed to guest in the virtual BIOS.'),
cfg.IntOpt('mem_stats_period_seconds',
default=10,
help='A number of seconds to memory usage statistics period. '
'Zero or negative value mean to disable memory usage '
'statistics.'),
cfg.ListOpt('uid_maps',
default=[],
help='List of uid targets and ranges.'
'Syntax is guest-uid:host-uid:count'
'Maximum of 5 allowed.'),
cfg.ListOpt('gid_maps',
default=[],
help='List of guid targets and ranges.'
'Syntax is guest-gid:host-gid:count'
'Maximum of 5 allowed.')
]
CONF = cfg.CONF
CONF.register_opts(libvirt_opts, 'libvirt')
CONF.import_opt('host', 'nova.netconf')
CONF.import_opt('my_ip', 'nova.netconf')
CONF.import_opt('use_cow_images', 'nova.virt.driver')
CONF.import_opt('enabled', 'nova.compute.api',
group='ephemeral_storage_encryption')
CONF.import_opt('cipher', 'nova.compute.api',
group='ephemeral_storage_encryption')
CONF.import_opt('key_size', 'nova.compute.api',
group='ephemeral_storage_encryption')
CONF.import_opt('live_migration_retry_count', 'nova.compute.manager')
CONF.import_opt('vncserver_proxyclient_address', 'nova.vnc', group='vnc')
CONF.import_opt('server_proxyclient_address', 'nova.spice', group='spice')
CONF.import_opt('vcpu_pin_set', 'nova.virt.hardware')
CONF.import_opt('vif_plugging_is_fatal', 'nova.virt.driver')
CONF.import_opt('vif_plugging_timeout', 'nova.virt.driver')
CONF.import_opt('enabled', 'nova.console.serial', group='serial_console')
CONF.import_opt('proxyclient_address', 'nova.console.serial',
group='serial_console')
CONF.import_opt('hw_disk_discard', 'nova.virt.libvirt.imagebackend',
group='libvirt')
CONF.import_group('workarounds', 'nova.utils')
CONF.import_opt('iscsi_use_multipath', 'nova.virt.libvirt.volume.iscsi',
group='libvirt')
DEFAULT_FIREWALL_DRIVER = "%s.%s" % (
libvirt_firewall.__name__,
libvirt_firewall.IptablesFirewallDriver.__name__)
MAX_CONSOLE_BYTES = 100 * units.Ki
# The libvirt driver will prefix any disable reason codes with this string.
DISABLE_PREFIX = 'AUTO: '
# Disable reason for the service which was enabled or disabled without reason
DISABLE_REASON_UNDEFINED = None
# Guest config console string
CONSOLE = "console=tty0 console=ttyS0"
GuestNumaConfig = collections.namedtuple(
'GuestNumaConfig', ['cpuset', 'cputune', 'numaconfig', 'numatune'])
libvirt_volume_drivers = [
'iscsi=nova.virt.libvirt.volume.iscsi.LibvirtISCSIVolumeDriver',
'iser=nova.virt.libvirt.volume.iser.LibvirtISERVolumeDriver',
'local=nova.virt.libvirt.volume.volume.LibvirtVolumeDriver',
'fake=nova.virt.libvirt.volume.volume.LibvirtFakeVolumeDriver',
'rbd=nova.virt.libvirt.volume.net.LibvirtNetVolumeDriver',
'sheepdog=nova.virt.libvirt.volume.net.LibvirtNetVolumeDriver',
'nfs=nova.virt.libvirt.volume.nfs.LibvirtNFSVolumeDriver',
'smbfs=nova.virt.libvirt.volume.smbfs.LibvirtSMBFSVolumeDriver',
'aoe=nova.virt.libvirt.volume.aoe.LibvirtAOEVolumeDriver',
'glusterfs='
'nova.virt.libvirt.volume.glusterfs.LibvirtGlusterfsVolumeDriver',
'fibre_channel='
'nova.virt.libvirt.volume.fibrechannel.'
'LibvirtFibreChannelVolumeDriver',
'scality=nova.virt.libvirt.volume.scality.LibvirtScalityVolumeDriver',
'gpfs=nova.virt.libvirt.volume.gpfs.LibvirtGPFSVolumeDriver',
'quobyte=nova.virt.libvirt.volume.quobyte.LibvirtQuobyteVolumeDriver',
'hgst=nova.virt.libvirt.volume.hgst.LibvirtHGSTVolumeDriver',
'scaleio=nova.virt.libvirt.volume.scaleio.LibvirtScaleIOVolumeDriver',
]
def patch_tpool_proxy():
"""eventlet.tpool.Proxy doesn't work with old-style class in __str__()
or __repr__() calls. See bug #962840 for details.
We perform a monkey patch to replace those two instance methods.
"""
def str_method(self):
return str(self._obj)
def repr_method(self):
return repr(self._obj)
tpool.Proxy.__str__ = str_method
tpool.Proxy.__repr__ = repr_method
patch_tpool_proxy()
# This is effectively the min version for i686/x86_64 + KVM/QEMU
# TODO(berrange) find out what min version ppc64 needs as it
# almost certainly wants something newer than this....
MIN_LIBVIRT_VERSION = (0, 9, 11)
# When the above version matches/exceeds this version
# delete it & corresponding code using it
MIN_LIBVIRT_DEVICE_CALLBACK_VERSION = (1, 1, 1)
# TODO(mriedem): Change MIN_LIB_VERSION to this in the 13.0.0 'M' release.
NEXT_MIN_LIBVIRT_VERSION = (0, 10, 2)
# Live snapshot requirements
MIN_LIBVIRT_LIVESNAPSHOT_VERSION = (1, 0, 0)
MIN_QEMU_LIVESNAPSHOT_VERSION = (1, 3, 0)
# block size tuning requirements
MIN_LIBVIRT_BLOCKIO_VERSION = (0, 10, 2)
# BlockJobInfo management requirement
MIN_LIBVIRT_BLOCKJOBINFO_VERSION = (1, 1, 1)
# Relative block commit & rebase (feature is detected,
# this version is only used for messaging)
MIN_LIBVIRT_BLOCKJOB_RELATIVE_VERSION = (1, 2, 7)
# libvirt discard feature
MIN_LIBVIRT_DISCARD_VERSION = (1, 0, 6)
MIN_QEMU_DISCARD_VERSION = (1, 6, 0)
# While earlier versions could support NUMA reporting and
# NUMA placement, not until 1.2.7 was there the ability
# to pin guest nodes to host nodes, so mandate that. Without
# this the scheduler cannot make guaranteed decisions, as the
# guest placement may not match what was requested
MIN_LIBVIRT_NUMA_VERSION = (1, 2, 7)
# Versions of libvirt with known NUMA topology issues
# See bug #1449028
BAD_LIBVIRT_NUMA_VERSIONS = [(1, 2, 9, 2)]
# While earlier versions could support hugepage backed
# guests, not until 1.2.8 was there the ability to request
# a particular huge page size. Without this the scheduler
# cannot make guaranteed decisions, as the huge page size
# used by the guest may not match what was requested
MIN_LIBVIRT_HUGEPAGE_VERSION = (1, 2, 8)
# Versions of libvirt with broken cpu pinning support. This excludes
# versions of libvirt with broken NUMA support since pinning needs
# NUMA
# See bug #1438226
BAD_LIBVIRT_CPU_POLICY_VERSIONS = [(1, 2, 10)]
# qemu 2.1 introduces support for pinning memory on host
# NUMA nodes, along with the ability to specify hugepage
# sizes per guest NUMA node
MIN_QEMU_NUMA_HUGEPAGE_VERSION = (2, 1, 0)
# fsFreeze/fsThaw requirement
MIN_LIBVIRT_FSFREEZE_VERSION = (1, 2, 5)
# Hyper-V paravirtualized time source
MIN_LIBVIRT_HYPERV_TIMER_VERSION = (1, 2, 2)
MIN_QEMU_HYPERV_TIMER_VERSION = (2, 0, 0)
MIN_LIBVIRT_HYPERV_FEATURE_VERSION = (1, 0, 0)
MIN_LIBVIRT_HYPERV_FEATURE_EXTRA_VERSION = (1, 1, 0)
MIN_QEMU_HYPERV_FEATURE_VERSION = (1, 1, 0)
# parallels driver support
MIN_LIBVIRT_PARALLELS_VERSION = (1, 2, 12)
# Ability to set the user guest password with Qemu
MIN_LIBVIRT_SET_ADMIN_PASSWD = (1, 2, 16)
# s/390 & s/390x architectures with KVM
MIN_LIBVIRT_KVM_S390_VERSION = (1, 2, 13)
MIN_QEMU_S390_VERSION = (2, 3, 0)
# Names of the types that do not get compressed during migration
NO_COMPRESSION_TYPES = ('qcow2',)
class LibvirtDriver(driver.ComputeDriver):
capabilities = {
"has_imagecache": True,
"supports_recreate": True,
"supports_migrate_to_same_host": False
}
def __init__(self, virtapi, read_only=False):
super(LibvirtDriver, self).__init__(virtapi)
global libvirt
if libvirt is None:
libvirt = importutils.import_module('libvirt')
self._host = host.Host(self._uri(), read_only,
lifecycle_event_handler=self.emit_event,
conn_event_handler=self._handle_conn_event)
self._initiator = None
self._fc_wwnns = None
self._fc_wwpns = None
self._caps = None
self.firewall_driver = firewall.load_driver(
DEFAULT_FIREWALL_DRIVER,
self.virtapi,
host=self._host)
self.vif_driver = libvirt_vif.LibvirtGenericVIFDriver()
self.volume_drivers = driver.driver_dict_from_config(
self._get_volume_drivers(), self)
self._disk_cachemode = None
self.image_cache_manager = imagecache.ImageCacheManager()
self.image_backend = imagebackend.Backend(CONF.use_cow_images)
self.disk_cachemodes = {}
self.valid_cachemodes = ["default",
"none",
"writethrough",
"writeback",
"directsync",
"unsafe",
]
self._conn_supports_start_paused = CONF.libvirt.virt_type in ('kvm',
'qemu')
for mode_str in CONF.libvirt.disk_cachemodes:
disk_type, sep, cache_mode = mode_str.partition('=')
if cache_mode not in self.valid_cachemodes:
LOG.warn(_LW('Invalid cachemode %(cache_mode)s specified '
'for disk type %(disk_type)s.'),
{'cache_mode': cache_mode, 'disk_type': disk_type})
continue
self.disk_cachemodes[disk_type] = cache_mode
self._volume_api = volume.API()
self._image_api = image.API()
sysinfo_serial_funcs = {
'none': lambda: None,
'hardware': self._get_host_sysinfo_serial_hardware,
'os': self._get_host_sysinfo_serial_os,
'auto': self._get_host_sysinfo_serial_auto,
}
self._sysinfo_serial_func = sysinfo_serial_funcs.get(
CONF.libvirt.sysinfo_serial)
self.job_tracker = instancejobtracker.InstanceJobTracker()
self._remotefs = remotefs.RemoteFilesystem()
def _get_volume_drivers(self):
return libvirt_volume_drivers
@property
def disk_cachemode(self):
if self._disk_cachemode is None:
# We prefer 'none' for consistent performance, host crash
# safety & migration correctness by avoiding host page cache.
# Some filesystems (eg GlusterFS via FUSE) don't support
# O_DIRECT though. For those we fallback to 'writethrough'
# which gives host crash safety, and is safe for migration
# provided the filesystem is cache coherent (cluster filesystems
# typically are, but things like NFS are not).
self._disk_cachemode = "none"
if not self._supports_direct_io(CONF.instances_path):
self._disk_cachemode = "writethrough"
return self._disk_cachemode
def _set_cache_mode(self, conf):
"""Set cache mode on LibvirtConfigGuestDisk object."""
try:
source_type = conf.source_type
driver_cache = conf.driver_cache
except AttributeError:
return
cache_mode = self.disk_cachemodes.get(source_type,
driver_cache)
conf.driver_cache = cache_mode
def _do_quality_warnings(self):
"""Warn about untested driver configurations.
This will log a warning message about untested driver or host arch
configurations to indicate to administrators that the quality is
unknown. Currently, only qemu or kvm on intel 32- or 64-bit systems
is tested upstream.
"""
caps = self._host.get_capabilities()
hostarch = caps.host.cpu.arch
if (CONF.libvirt.virt_type not in ('qemu', 'kvm') or
hostarch not in (arch.I686, arch.X86_64)):
LOG.warn(_LW('The libvirt driver is not tested on '
'%(type)s/%(arch)s by the OpenStack project and '
'thus its quality can not be ensured. For more '
'information, see: https://wiki.openstack.org/wiki/'
'HypervisorSupportMatrix'),
{'type': CONF.libvirt.virt_type, 'arch': hostarch})
def _handle_conn_event(self, enabled, reason):
LOG.info(_LI("Connection event '%(enabled)d' reason '%(reason)s'"),
{'enabled': enabled, 'reason': reason})
self._set_host_enabled(enabled, reason)
def _version_to_string(self, version):
return '.'.join([str(x) for x in version])
def init_host(self, host):
self._host.initialize()
self._do_quality_warnings()
if (CONF.libvirt.virt_type == 'lxc' and
not (CONF.libvirt.uid_maps and CONF.libvirt.gid_maps)):
LOG.warn(_LW("Running libvirt-lxc without user namespaces is "
"dangerous. Containers spawned by Nova will be run "
"as the host's root user. It is highly suggested "
"that user namespaces be used in a public or "
"multi-tenant environment."))
# Stop libguestfs using KVM unless we're also configured
# to use this. This solves problem where people need to
# stop Nova use of KVM because nested-virt is broken
if CONF.libvirt.virt_type != "kvm":
guestfs.force_tcg()
if not self._host.has_min_version(MIN_LIBVIRT_VERSION):
raise exception.NovaException(
_('Nova requires libvirt version %s or greater.') %
self._version_to_string(MIN_LIBVIRT_VERSION))
if (CONF.libvirt.virt_type == 'parallels' and
not self._host.has_min_version(MIN_LIBVIRT_PARALLELS_VERSION)):
raise exception.NovaException(
_('Running Nova with parallels virt_type requires '
'libvirt version %s') %
self._version_to_string(MIN_LIBVIRT_PARALLELS_VERSION))
# TODO(mriedem): We plan to move to a minimum required version of
# libvirt 0.10.2 in the 13.0.0 'M' release so if we're running with
# less than that now, log a warning.
if not self._host.has_min_version(NEXT_MIN_LIBVIRT_VERSION):
LOG.warning(_LW('Running Nova with a libvirt version less than '
'%(version)s is deprecated. The required minimum '
'version of libvirt will be raised to %(version)s '
'in the 13.0.0 release.'),
{'version': self._version_to_string(
NEXT_MIN_LIBVIRT_VERSION)})
if (CONF.libvirt.virt_type in ('kvm', 'qemu') and
arch.from_host() in (arch.S390, arch.S390X) and
not self._host.has_min_version(MIN_LIBVIRT_KVM_S390_VERSION,
MIN_QEMU_S390_VERSION)):
raise exception.NovaException(
_('Running Nova with qemu/kvm virt_type on s390/s390x '
'requires libvirt version %(libvirt_ver)s and '
'qemu version %(qemu_ver)s, or greater') %
{'libvirt_ver': self._version_to_string(
MIN_LIBVIRT_KVM_S390_VERSION),
'qemu_ver': self._version_to_string(
MIN_QEMU_S390_VERSION)})
# TODO(sahid): This method is targeted for removal when the tests
# have been updated to avoid its use
#
# All libvirt API calls on the libvirt.Connect object should be
# encapsulated by methods on the nova.virt.libvirt.host.Host
# object, rather than directly invoking the libvirt APIs. The goal
# is to avoid a direct dependency on the libvirt API from the
# driver.py file.
def _get_connection(self):
return self._host.get_connection()
_conn = property(_get_connection)
@staticmethod
def _uri():
if CONF.libvirt.virt_type == 'uml':
uri = CONF.libvirt.connection_uri or 'uml:///system'
elif CONF.libvirt.virt_type == 'xen':
uri = CONF.libvirt.connection_uri or 'xen:///'
elif CONF.libvirt.virt_type == 'lxc':
uri = CONF.libvirt.connection_uri or 'lxc:///'
elif CONF.libvirt.virt_type == 'parallels':
uri = CONF.libvirt.connection_uri or 'parallels:///system'
else:
uri = CONF.libvirt.connection_uri or 'qemu:///system'
return uri
def instance_exists(self, instance):
"""Efficient override of base instance_exists method."""
try:
self._host.get_guest(instance)
return True
except exception.NovaException:
return False
def list_instances(self):
names = []
for dom in self._host.list_instance_domains(only_running=False):
names.append(dom.name())
return names
def list_instance_uuids(self):
uuids = []
for dom in self._host.list_instance_domains(only_running=False):
uuids.append(dom.UUIDString())
return uuids
def plug_vifs(self, instance, network_info):
"""Plug VIFs into networks."""
for vif in network_info:
self.vif_driver.plug(instance, vif)
def _unplug_vifs(self, instance, network_info, ignore_errors):
"""Unplug VIFs from networks."""
for vif in network_info:
try:
self.vif_driver.unplug(instance, vif)
except exception.NovaException:
if not ignore_errors:
raise
def unplug_vifs(self, instance, network_info):
self._unplug_vifs(instance, network_info, False)
def _teardown_container(self, instance):
inst_path = libvirt_utils.get_instance_path(instance)
container_dir = os.path.join(inst_path, 'rootfs')
rootfs_dev = instance.system_metadata.get('rootfs_device_name')
disk.teardown_container(container_dir, rootfs_dev)
def _destroy(self, instance, attempt=1):
try:
guest = self._host.get_guest(instance)
except exception.InstanceNotFound:
guest = None
# If the instance is already terminated, we're still happy
# Otherwise, destroy it
old_domid = -1
if guest is not None:
try:
old_domid = guest.id
guest.poweroff()
except libvirt.libvirtError as e:
is_okay = False
errcode = e.get_error_code()
if errcode == libvirt.VIR_ERR_NO_DOMAIN:
# Domain already gone. This can safely be ignored.
is_okay = True
elif errcode == libvirt.VIR_ERR_OPERATION_INVALID:
# If the instance is already shut off, we get this:
# Code=55 Error=Requested operation is not valid:
# domain is not running
# TODO(sahid): At this point we should be a Guest object
state = guest.get_power_state(self._host)
if state == power_state.SHUTDOWN:
is_okay = True
elif errcode == libvirt.VIR_ERR_INTERNAL_ERROR:
errmsg = e.get_error_message()
if (CONF.libvirt.virt_type == 'lxc' and
errmsg == 'internal error: '
'Some processes refused to die'):
# Some processes in the container didn't die
# fast enough for libvirt. The container will
# eventually die. For now, move on and let
# the wait_for_destroy logic take over.
is_okay = True
elif errcode == libvirt.VIR_ERR_OPERATION_TIMEOUT:
LOG.warn(_LW("Cannot destroy instance, operation time "
"out"),
instance=instance)
reason = _("operation time out")
raise exception.InstancePowerOffFailure(reason=reason)
elif errcode == libvirt.VIR_ERR_SYSTEM_ERROR:
if e.get_int1() == errno.EBUSY:
# NOTE(danpb): When libvirt kills a process it sends it
# SIGTERM first and waits 10 seconds. If it hasn't gone
# it sends SIGKILL and waits another 5 seconds. If it
# still hasn't gone then you get this EBUSY error.
# Usually when a QEMU process fails to go away upon
# SIGKILL it is because it is stuck in an
# uninterruptable kernel sleep waiting on I/O from
# some non-responsive server.
# Given the CPU load of the gate tests though, it is
# conceivable that the 15 second timeout is too short,
# particularly if the VM running tempest has a high
# steal time from the cloud host. ie 15 wallclock
# seconds may have passed, but the VM might have only
# have a few seconds of scheduled run time.
LOG.warn(_LW('Error from libvirt during destroy. '
'Code=%(errcode)s Error=%(e)s; '
'attempt %(attempt)d of 3'),
{'errcode': errcode, 'e': e,
'attempt': attempt},
instance=instance)
with excutils.save_and_reraise_exception() as ctxt:
# Try up to 3 times before giving up.
if attempt < 3:
ctxt.reraise = False
self._destroy(instance, attempt + 1)
return
if not is_okay:
with excutils.save_and_reraise_exception():
LOG.error(_LE('Error from libvirt during destroy. '
'Code=%(errcode)s Error=%(e)s'),
{'errcode': errcode, 'e': e},
instance=instance)
def _wait_for_destroy(expected_domid):
"""Called at an interval until the VM is gone."""
# NOTE(vish): If the instance disappears during the destroy
# we ignore it so the cleanup can still be
# attempted because we would prefer destroy to
# never fail.
try:
dom_info = self.get_info(instance)
state = dom_info.state
new_domid = dom_info.id
except exception.InstanceNotFound:
LOG.info(_LI("During wait destroy, instance disappeared."),
instance=instance)
raise loopingcall.LoopingCallDone()
if state == power_state.SHUTDOWN:
LOG.info(_LI("Instance destroyed successfully."),
instance=instance)
raise loopingcall.LoopingCallDone()
# NOTE(wangpan): If the instance was booted again after destroy,
# this may be a endless loop, so check the id of
# domain here, if it changed and the instance is
# still running, we should destroy it again.
# see https://bugs.launchpad.net/nova/+bug/1111213 for more details
if new_domid != expected_domid:
LOG.info(_LI("Instance may be started again."),
instance=instance)
kwargs['is_running'] = True
raise loopingcall.LoopingCallDone()
kwargs = {'is_running': False}
timer = loopingcall.FixedIntervalLoopingCall(_wait_for_destroy,
old_domid)
timer.start(interval=0.5).wait()
if kwargs['is_running']:
LOG.info(_LI("Going to destroy instance again."),
instance=instance)
self._destroy(instance)
else:
# NOTE(GuanQiang): teardown container to avoid resource leak
if CONF.libvirt.virt_type == 'lxc':
self._teardown_container(instance)
def destroy(self, context, instance, network_info, block_device_info=None,
destroy_disks=True, migrate_data=None):
self._destroy(instance)
self.cleanup(context, instance, network_info, block_device_info,
destroy_disks, migrate_data)
def _undefine_domain(self, instance):
try:
guest = self._host.get_guest(instance)
try:
guest.delete_configuration()
except libvirt.libvirtError as e:
with excutils.save_and_reraise_exception():
errcode = e.get_error_code()
LOG.error(_LE('Error from libvirt during undefine. '
'Code=%(errcode)s Error=%(e)s'),
{'errcode': errcode, 'e': e}, instance=instance)
except exception.InstanceNotFound:
pass
def cleanup(self, context, instance, network_info, block_device_info=None,
destroy_disks=True, migrate_data=None, destroy_vifs=True):
if destroy_vifs:
self._unplug_vifs(instance, network_info, True)
retry = True
while retry:
try:
self.unfilter_instance(instance, network_info)
except libvirt.libvirtError as e:
try:
state = self.get_info(instance).state
except exception.InstanceNotFound:
state = power_state.SHUTDOWN
if state != power_state.SHUTDOWN:
LOG.warn(_LW("Instance may be still running, destroy "
"it again."), instance=instance)
self._destroy(instance)
else:
retry = False
errcode = e.get_error_code()
LOG.exception(_LE('Error from libvirt during unfilter. '
'Code=%(errcode)s Error=%(e)s'),
{'errcode': errcode, 'e': e},
instance=instance)
reason = "Error unfiltering instance."
raise exception.InstanceTerminationFailure(reason=reason)
except Exception:
retry = False
raise
else:
retry = False
# FIXME(wangpan): if the instance is booted again here, such as the
# the soft reboot operation boot it here, it will
# become "running deleted", should we check and destroy
# it at the end of this method?
# NOTE(vish): we disconnect from volumes regardless
block_device_mapping = driver.block_device_info_get_mapping(
block_device_info)
for vol in block_device_mapping:
connection_info = vol['connection_info']
disk_dev = vol['mount_device']
if disk_dev is not None:
disk_dev = disk_dev.rpartition("/")[2]
if ('data' in connection_info and
'volume_id' in connection_info['data']):
volume_id = connection_info['data']['volume_id']
encryption = encryptors.get_encryption_metadata(
context, self._volume_api, volume_id, connection_info)
if encryption:
# The volume must be detached from the VM before
# disconnecting it from its encryptor. Otherwise, the
# encryptor may report that the volume is still in use.
encryptor = self._get_volume_encryptor(connection_info,
encryption)
encryptor.detach_volume(**encryption)
try:
self._disconnect_volume(connection_info, disk_dev)
except Exception as exc:
with excutils.save_and_reraise_exception() as ctxt:
if destroy_disks:
# Don't block on Volume errors if we're trying to
# delete the instance as we may be partially created
# or deleted
ctxt.reraise = False
LOG.warn(_LW("Ignoring Volume Error on vol %(vol_id)s "
"during delete %(exc)s"),
{'vol_id': vol.get('volume_id'), 'exc': exc},
instance=instance)
if destroy_disks:
# NOTE(haomai): destroy volumes if needed
if CONF.libvirt.images_type == 'lvm':
self._cleanup_lvm(instance, block_device_info)
if CONF.libvirt.images_type == 'rbd':
self._cleanup_rbd(instance)
if destroy_disks or (
migrate_data and migrate_data.get('is_shared_block_storage',
False)):
attempts = int(instance.system_metadata.get('clean_attempts',
'0'))
success = self.delete_instance_files(instance)
# NOTE(mriedem): This is used in the _run_pending_deletes periodic
# task in the compute manager. The tight coupling is not great...
instance.system_metadata['clean_attempts'] = str(attempts + 1)
if success:
instance.cleaned = True
instance.save()
if CONF.serial_console.enabled:
try:
guest = self._host.get_guest(instance)
serials = self._get_serial_ports_from_guest(guest)
for hostname, port in serials:
serial_console.release_port(host=hostname, port=port)
except exception.InstanceNotFound:
pass
self._undefine_domain(instance)
def _detach_encrypted_volumes(self, instance, block_device_info):
"""Detaches encrypted volumes attached to instance."""
disks = jsonutils.loads(self.get_instance_disk_info(instance,
block_device_info))
encrypted_volumes = filter(dmcrypt.is_encrypted,
[disk['path'] for disk in disks])
for path in encrypted_volumes:
dmcrypt.delete_volume(path)
def _get_serial_ports_from_guest(self, guest, mode=None):
"""Returns an iterator over serial port(s) configured on guest.
:param mode: Should be a value in (None, bind, connect)
"""
xml = guest.get_xml_desc()
tree = etree.fromstring(xml)
# The 'serial' device is the base for x86 platforms. Other platforms
# (e.g. kvm on system z = arch.S390X) can only use 'console' devices.
xpath_mode = "[@mode='%s']" % mode if mode else ""
serial_tcp = "./devices/serial[@type='tcp']/source" + xpath_mode
console_tcp = "./devices/console[@type='tcp']/source" + xpath_mode
tcp_devices = tree.findall(serial_tcp)
if len(tcp_devices) == 0:
tcp_devices = tree.findall(console_tcp)
for source in tcp_devices:
yield (source.get("host"), int(source.get("service")))
@staticmethod
def _get_rbd_driver():
return rbd_utils.RBDDriver(
pool=CONF.libvirt.images_rbd_pool,
ceph_conf=CONF.libvirt.images_rbd_ceph_conf,
rbd_user=CONF.libvirt.rbd_user)
def _cleanup_rbd(self, instance):
LibvirtDriver._get_rbd_driver().cleanup_volumes(instance)
def _cleanup_lvm(self, instance, block_device_info):
"""Delete all LVM disks for given instance object."""
if instance.get('ephemeral_key_uuid') is not None:
self._detach_encrypted_volumes(instance, block_device_info)
disks = self._lvm_disks(instance)
if disks:
lvm.remove_volumes(disks)
def _lvm_disks(self, instance):
"""Returns all LVM disks for given instance object."""
if CONF.libvirt.images_volume_group:
vg = os.path.join('/dev', CONF.libvirt.images_volume_group)
if not os.path.exists(vg):
return []
pattern = '%s_' % instance.uuid
def belongs_to_instance(disk):
return disk.startswith(pattern)