-
Notifications
You must be signed in to change notification settings - Fork 161
/
migrate.go
1182 lines (1062 loc) · 40.3 KB
/
migrate.go
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 (c) 2017-2018 Tigera, Inc. All rights reserved.
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package migrator
import (
"context"
"errors"
"fmt"
"math/rand"
"strings"
"time"
"github.com/coreos/go-semver/semver"
log "github.com/sirupsen/logrus"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/util/uuid"
apiv3 "github.com/projectcalico/libcalico-go/lib/apis/v3"
bapi "github.com/projectcalico/libcalico-go/lib/backend/api"
"github.com/projectcalico/libcalico-go/lib/backend/model"
"github.com/projectcalico/libcalico-go/lib/clientv3"
cerrors "github.com/projectcalico/libcalico-go/lib/errors"
"github.com/projectcalico/libcalico-go/lib/numorstring"
"github.com/projectcalico/libcalico-go/lib/options"
"github.com/projectcalico/libcalico-go/lib/upgrade/converters"
"github.com/projectcalico/libcalico-go/lib/upgrade/migrator/clients"
validatorv3 "github.com/projectcalico/libcalico-go/lib/validator/v3"
)
const (
forceEnableReadyRetries = 30
maxApplyRetries = 5
numAppliesPerUpdate = 100
retryInterval = 5 * time.Second
// The minimum version to upgrade from. This should not include the leading 'v'.
minUpgradeVersion = "2.6.5"
)
// Interface is the migration interface used for migrating data from version
// v2.x to v3.x.
type Interface interface {
ValidateConversion() (*MigrationData, error)
IsDestinationEmpty() (bool, error)
ShouldMigrate() (bool, error)
CanMigrate() error
Migrate() (*MigrationData, error)
IsMigrationInProgress() (bool, error)
Abort() error
Complete() error
}
// StatusWriterInterface is an optional interface supplied by the consumer of
// the migration helper used to record status of the migration.
type StatusWriterInterface interface {
Msg(string)
Bullet(string)
Error(string)
}
// New creates a new migration helper implementing Interface.
func New(clientv3 clientv3.Interface, clientv1 clients.V1ClientInterface, statusWriter StatusWriterInterface) Interface {
return &migrationHelper{
clientv3: clientv3,
clientv1: clientv1,
statusWriter: statusWriter,
}
}
// migrationHelper implements the migrate.Interface.
type migrationHelper struct {
clientv3 clientv3.Interface
clientv1 clients.V1ClientInterface
statusWriter StatusWriterInterface
}
// Error types encountered during validation and migration.
type ErrorType int
const (
ErrorGeneric ErrorType = iota
ErrorConvertingData
ErrorMigratingData
)
type MigrationError struct {
Err error
Type ErrorType
NeedsAbort bool
}
func (m MigrationError) Error() string {
return m.Err.Error()
}
// MigrationData includes details about data migrated using the migration helper.
type MigrationData struct {
// The converted resources
Resources []converters.Resource
// The converted resource names
NameConversions []NameConversion
// Errors hit attempting to convert the v1 data to v3 format. The
// KeyV3 and ValueV3 will be nil for these conversion errors.
ConversionErrors []ConversionError
// Errors hit validating the converted v3 data. This suggests an error in the
// conversion script which should be fixed before reattempting the conversion.
ConvertedResourceValidationErrors []ConversionError
// Name clashes in the converted resources. These need to be resolved through
// reconfiguration before attempting the upgrade.
NameClashes []NameClash
// Entries that were skipped because they will be handled by the Kubernetes
// Policy controller.
HandledByPolicyCtrl []model.Key
}
// HasErrors returns whether there are any errors contained in the MigrationData.
func (c *MigrationData) HasErrors() bool {
return len(c.ConversionErrors) != 0 ||
len(c.ConvertedResourceValidationErrors) != 0 ||
len(c.NameClashes) != 0
}
// ConversionError contains details about a specific error converting a
// v1 resource to a v3 resource.
type ConversionError struct {
Cause error
KeyV1 model.Key
ValueV1 interface{}
KeyV3 model.Key
ValueV3 converters.Resource
}
// NameConversion contains details about name/id conversions.
type NameConversion struct {
KeyV1 model.Key
KeyV3 model.Key
}
// NameClash contains details about name/id clashes (i.e. when two converted resource
// names (for the same resource type) clash.
type NameClash struct {
KeyV1 model.Key
KeyV3 model.Key
OtherKeyV1 model.Key
}
// Validate validates that the v1 data can be correctly converted to v3.
// If an error is returned it will be of type MigrationError.
func (m *migrationHelper) ValidateConversion() (*MigrationData, error) {
m.status("Validating conversion of v1 data to v3")
data, err := m.queryAndConvertResources()
if err != nil {
m.statusError("Unable to perform validation, please resolve errors and retry")
m.statusBullet("Cause: %v", err)
return nil, MigrationError{
Type: ErrorGeneric,
Err: err,
}
}
if data.HasErrors() {
m.statusError("Error converting data, check output for details and resolve issues before starting upgrade")
return data, MigrationError{
Type: ErrorConvertingData,
Err: fmt.Errorf("error converting data: %v", err),
}
}
m.statusBullet("data conversion successful")
// Finally, check that we found some data. For KDD if there were no resources
// then that's fine, otherwise we should fail.
if !m.clientv1.IsKDD() && len(data.Resources) == 0 {
m.statusError("No v1 resources detected: is the API configuration correctly configured?")
return nil, MigrationError{
Type: ErrorGeneric,
Err: errors.New("no v1 resources detected: is the API configuration correctly configured?"),
}
}
// Everything validated correctly.
m.status("Data conversion validated successfully")
return data, nil
}
func (m *migrationHelper) IsDestinationEmpty() (bool, error) {
m.status("Validating the v3 datastore")
clean, err := m.v3DatastoreIsClean()
if err != nil {
m.statusError("Unable to validate the v3 datastore")
m.statusBullet("Cause: %v", err)
return false, MigrationError{
Type: ErrorGeneric,
Err: fmt.Errorf("unable to validate the v3 datastore: %v", err),
}
}
if clean {
m.statusBullet("the v3 datastore is empty")
} else {
m.statusBullet("the v3 datastore is not empty")
}
return clean, nil
}
// Migrate migrates the data from v1 format to v3. Both a v1 and v3 client are required.
// It returns the converted set of data, a bool indicating whether the migration succeeded.
// If an error is returned it will be of type MigrationError.
func (m *migrationHelper) Migrate() (*MigrationData, error) {
// Now set the Ready flag to False. This will stop Felix from making any data plane updates
// and will prevent the orchestrator plugins from adding any new workloads or IP allocations
if !m.clientv1.IsKDD() {
m.status("Pausing Calico networking")
if err := m.clearReadyV1(); err != nil {
m.statusError("Unable to pause calico networking")
return nil, MigrationError{
Type: ErrorGeneric,
Err: fmt.Errorf("unable to pause calico networking: %v", err),
}
}
// Wait for a short period to allow orchestrators to finish any current allocations.
m.status("Calico networking is now paused - waiting for 15s")
time.Sleep(15 * time.Second)
}
// Now query all the resources again and convert - this is the final snapshot that we will use.
m.status("Querying current v1 snapshot and converting to v3")
data, err := m.queryAndConvertResources()
if err != nil {
m.statusError("Unable to convert the v1 snapshot to v3")
m.statusBullet("cause: %v", err)
return nil, m.abortAfterError(
fmt.Errorf("error converting data: %v", err), ErrorGeneric,
)
}
if data.HasErrors() {
m.statusError("Error converting data - will attempt to abort upgrade")
return nil, m.abortAfterError(
fmt.Errorf("error converting data: %v", err), ErrorConvertingData,
)
}
m.statusBullet("data converted successfully")
m.status("Storing v3 data")
if err = m.storeV3Resources(data); err != nil {
m.statusError("Unable to store the v3 resources")
m.statusBullet("cause: %v", err)
return nil, m.abortAfterError(
fmt.Errorf("error storing converted data: %v", err), ErrorMigratingData,
)
}
// And we also need to migrate the IPAM data.
m.status("Migrating IPAM data")
if m.clientv1.IsKDD() {
m.statusBullet("no data to migrate - not supported")
} else if err = m.migrateIPAMData(); err != nil {
m.statusError("Unable to migrate the v3 IPAM data")
m.statusBullet("cause: %v", err)
return nil, m.abortAfterError(
fmt.Errorf("error migrating IPAM data: %v", err), ErrorMigratingData,
)
}
m.status("Data migration from v1 to v3 successful")
m.statusBullet("check the output for details of the migrated resources")
m.statusBullet("continue by upgrading your calico/node versions to Calico v3.x")
return data, nil
}
func (m *migrationHelper) abortAfterError(err error, errType ErrorType) error {
if m.clientv1.IsKDD() {
return MigrationError{Type: errType, Err: err}
}
if ae := m.Abort(); ae == nil {
return MigrationError{Type: errType, Err: err}
}
return MigrationError{Type: errType, Err: err, NeedsAbort: true}
}
// IsMigrationInProgress infers from ShouldMigrate and the Ready flag if the datastore
// is being migrated and returns true if it is. If migration is needed and the Ready
// flag is false then it is assumed migration is in progress. This could provide a
// false positive if there was an error during migration and migration was aborted.
// The other non-error cases will return false.
func (m *migrationHelper) IsMigrationInProgress() (bool, error) {
migrateNeeded, err := m.ShouldMigrate()
if err != nil {
return false, fmt.Errorf("error checking migration progress status: %v", err)
} else if migrateNeeded {
// Migration is needed. If Ready is true then no upgrade has been started
// (not in progress). If Ready is false then upgrade has been started.
ready, err := m.isReady()
if err != nil {
return false, fmt.Errorf("error checking migration progress status: %v", err)
}
return !ready, nil
} else {
return false, nil
}
}
// Abort aborts the upgrade by re-enabling Calico networking in v1.
// If an error is returned it will be of type MigrationError.
func (m *migrationHelper) Abort() error {
m.status("Aborting upgrade")
var err error
if !m.clientv1.IsKDD() {
m.status("Re-enabling Calico networking for v1")
for i := 0; i < forceEnableReadyRetries; i++ {
err = m.setReadyV1()
if err == nil {
break
}
time.Sleep(1 * time.Second)
}
}
if err != nil {
m.statusError("Failed to abort upgrade. Retry command.")
m.statusBullet("cause: %v", err)
return MigrationError{Type: ErrorGeneric, Err: err, NeedsAbort: true}
}
m.status("Upgrade aborted successfully")
return nil
}
// Complete completes the upgrade by re-enabling Calico networking in v1.
// If an error is returned it will be of type MigrationError.
func (m *migrationHelper) Complete() error {
m.status("Completing upgrade")
var err error
if !m.clientv1.IsKDD() {
m.status("Enabling Calico networking for v3")
for i := 0; i < forceEnableReadyRetries; i++ {
err = m.setReadyV3(true)
if err == nil {
break
}
time.Sleep(1 * time.Second)
}
}
if err != nil {
m.statusError("Failed to complete upgrade. Retry command.")
m.statusBullet("cause: %v", err)
return MigrationError{Type: ErrorGeneric, Err: err}
}
m.status("Upgrade completed successfully")
return nil
}
type policyCtrlFilterOut func(model.Key) bool
var noFilter = func(_ model.Key) bool { return false }
// Filter to filter out K8s backed network policies
var filterGNP = func(k model.Key) bool {
gk := k.(model.PolicyKey)
return strings.HasPrefix(gk.Name, "knp.default.")
}
// Filter to filter out K8s (namespace) and OpenStack backed profiles
var filterProfile = func(k model.Key) bool {
gk := k.(model.ProfileKey)
return strings.HasPrefix(gk.Name, "k8s_ns.") || strings.HasPrefix(gk.Name, "openstack-sg-")
}
// Filter to filter out OpenStack backed workload endpoints
var filterWEP = func(k model.Key) bool {
gk := k.(model.WorkloadEndpointKey)
return gk.OrchestratorID == apiv3.OrchestratorOpenStack
}
type ic interface {
IsClean() (bool, error)
}
func (m *migrationHelper) v3DatastoreIsClean() (bool, error) {
bc := m.clientv3.(backendClientAccessor).Backend()
if i, ok := bc.(ic); ok {
return i.IsClean()
}
return true, nil
}
// queryAndConvertResources queries the v1 resources and converts them to the equivalent
// v3 resources.
// This method returns an error if it is unable to query the current v1 settings.
// Errors from the conversion are returned within the MigrationData - this function will
// attempt to convert everything before returning with the set of converted data and
// conversion errors - this allows a full pre-migration report to be generated in a single
// shot.
func (m *migrationHelper) queryAndConvertResources() (*MigrationData, error) {
data := &MigrationData{}
// Query and convert global felix configuration and cluster info.
if err := m.queryAndConvertFelixConfigV1ToV3(data); err != nil {
return nil, err
}
m.statusBullet("handling BGPConfiguration (global) resource")
// Query the global BGP configuration: default AS number; node-to-node mesh.
if err := m.queryAndConvertGlobalBGPConfigV1ToV3(data); err != nil {
return nil, err
}
if m.clientv1.IsKDD() {
m.statusBullet("skipping Node resources - these do not need migrating")
} else {
m.statusBullet("handling Node resources")
// Query and convert the Nodes
if err := m.queryAndConvertV1ToV3Nodes(data); err != nil {
return nil, err
}
}
if m.clientv1.IsKDD() {
m.statusBullet("skipping BGPPeer (global) resources - these do not need migrating")
} else {
m.statusBullet("handling BGPPeer (global) resources")
// Query and convert the BGPPeers
if err := m.queryAndConvertV1ToV3Resources(
data, model.GlobalBGPPeerListOptions{}, converters.BGPPeer{}, noFilter,
); err != nil {
return nil, err
}
}
m.statusBullet("handling BGPPeer (node) resources")
if err := m.queryAndConvertV1ToV3Resources(
data, model.NodeBGPPeerListOptions{}, converters.BGPPeer{}, noFilter,
); err != nil {
return nil, err
}
if m.clientv1.IsKDD() {
m.statusBullet("skipping HostEndpoint resources - not supported")
} else {
m.statusBullet("handling HostEndpoint resources")
// Query and convert the HostEndpoints
if err := m.queryAndConvertV1ToV3Resources(
data, model.HostEndpointListOptions{}, converters.HostEndpoint{}, noFilter,
); err != nil {
return nil, err
}
}
if m.clientv1.IsKDD() {
m.statusBullet("skipping IPPool resources - these do not need migrating")
} else {
m.statusBullet("handling IPPool resources")
// Query and convert the IPPools
if err := m.queryAndConvertV1ToV3Resources(
data, model.IPPoolListOptions{}, converters.IPPool{}, noFilter,
); err != nil {
return nil, err
}
}
if m.clientv1.IsKDD() {
m.statusBullet("skipping GlobalNetworkPolicy resources - these do not need migrating")
} else {
m.statusBullet("handling GlobalNetworkPolicy resources")
// Query and convert the Policies
if err := m.queryAndConvertV1ToV3Resources(
data, model.PolicyListOptions{}, converters.Policy{}, filterGNP,
); err != nil {
return nil, err
}
}
if m.clientv1.IsKDD() {
m.statusBullet("skipping Profile resources - these do not need migrating")
} else {
m.statusBullet("handling Profile resources")
// Query and convert the Profiles
if err := m.queryAndConvertV1ToV3Resources(
data, model.ProfileListOptions{}, converters.Profile{}, filterProfile,
); err != nil {
return nil, err
}
}
if m.clientv1.IsKDD() {
m.statusBullet("skipping WorkloadEndpoint resources - these do not need migrating")
} else {
m.statusBullet("handling WorkloadEndpoint resources")
// Query and convert the WorkloadEndpoints
if err := m.queryAndConvertV1ToV3Resources(
data, model.WorkloadEndpointListOptions{}, converters.WorkloadEndpoint{}, filterWEP,
); err != nil {
return nil, err
}
}
return data, nil
}
// Query the v1 format resources and convert to the v3 format. Successfully
// migrated resources are appended to res, and conversion errors to convErr.
func (m *migrationHelper) queryAndConvertV1ToV3Resources(
data *MigrationData,
listInterface model.ListInterface,
converter converters.Converter,
filterOut policyCtrlFilterOut,
) error {
// Start by listing the results from the v1 client.
kvps, err := m.clientv1.List(listInterface)
if err != nil {
switch err.(type) {
case cerrors.ErrorResourceDoesNotExist, cerrors.ErrorOperationNotSupported:
return nil
default:
return err
}
}
// Keep track of the converted names so that we can determine if we have any
// name clashes. We don't generally expect this, but we do need to police against
// it just in case.
convertedNames := make(map[string]model.Key, len(kvps))
// Pass the results through the supplied converter and check that each result
// validates.
for _, kvp := range kvps {
if filterOut(kvp.Key) {
log.Infof("Filter out Policy Controller created resource: %s", kvp.Key)
data.HandledByPolicyCtrl = append(data.HandledByPolicyCtrl, kvp.Key)
continue
}
r, err := converter.BackendV1ToAPIV3(kvp)
if err != nil {
data.ConversionErrors = append(data.ConversionErrors, ConversionError{
KeyV1: kvp.Key,
ValueV1: kvp.Value,
Cause: err,
})
continue
}
// Check the converted name for clashes. Store an error if there is a clash and
// continue with additional checks so that we output as much information as possible.
valid := true
convertedName := r.GetObjectMeta().GetNamespace() + "/" + r.GetObjectMeta().GetName()
if k, ok := convertedNames[convertedName]; ok {
data.NameClashes = append(data.NameClashes, NameClash{
KeyV1: kvp.Key,
KeyV3: resourceToKey(r),
OtherKeyV1: k,
})
valid = false
}
convertedNames[convertedName] = kvp.Key
// Check the converted resource validates correctly.
if err := validatorv3.Validate(r); err != nil {
data.ConvertedResourceValidationErrors = append(data.ConvertedResourceValidationErrors, ConversionError{
KeyV1: kvp.Key,
ValueV1: kvp.Value,
KeyV3: resourceToKey(r),
ValueV3: r,
Cause: err,
})
valid = false
}
// Only store the resource and the converted name if it's valid.
if valid {
data.Resources = append(data.Resources, r)
data.NameConversions = append(data.NameConversions, NameConversion{
KeyV1: kvp.Key,
KeyV3: resourceToKey(r),
})
}
}
return nil
}
func (m *migrationHelper) queryAndConvertGlobalBGPConfigV1ToV3(data *MigrationData) error {
globalBGPConfig := apiv3.NewBGPConfiguration()
globalBGPConfig.Name = "default"
log.Info("Converting BGP config -> BGPConfiguration(default)")
var setValue bool
if kvp, err := m.clientv1.Get(model.GlobalBGPConfigKey{Name: "AsNumber"}); err != nil {
if _, ok := err.(cerrors.ErrorResourceDoesNotExist); !ok {
return err
}
log.Info("No global default ASNumber configured")
} else if kvp.Value.(string) != "" {
asNum, err := numorstring.ASNumberFromString(kvp.Value.(string))
if err != nil {
log.WithError(err).WithField("ASNumber", kvp.Value).Info("Invalid global default ASNumber")
data.ConversionErrors = append(data.ConversionErrors, ConversionError{
ValueV1: kvp.Value,
Cause: fmt.Errorf("default ASNumber is not valid: %s", kvp.Value.(string)),
})
return err
}
globalBGPConfig.Spec.ASNumber = &asNum
setValue = true
}
if kvp, err := m.clientv1.Get(model.GlobalBGPConfigKey{Name: "LogLevel"}); err != nil {
if _, ok := err.(cerrors.ErrorResourceDoesNotExist); !ok {
return err
}
log.Info("No global BGP log level configured")
} else if kvp.Value.(string) != "" {
globalBGPConfig.Spec.LogSeverityScreen = convertLogLevel(kvp.Value.(string))
setValue = true
}
if kvp, err := m.clientv1.Get(model.GlobalBGPConfigKey{Name: "NodeMeshEnabled"}); err != nil {
if _, ok := err.(cerrors.ErrorResourceDoesNotExist); !ok {
return err
}
log.Info("No global node to node mesh enabled setting configured")
} else if kvp.Value.(string) != "" {
nodeMeshEnabled := strings.ToLower(kvp.Value.(string)) == "true"
globalBGPConfig.Spec.NodeToNodeMeshEnabled = &nodeMeshEnabled
setValue = true
}
if setValue {
data.Resources = append(data.Resources, globalBGPConfig)
}
return nil
}
// Query the v1 format resources and convert to the v3 format. Successfully
// migrated resources are appended to res, and conversion errors to convErr.
func (m *migrationHelper) queryAndConvertV1ToV3Nodes(data *MigrationData) error {
// Start by querying the nodes and converting them, we don't add the nodes to the list
// of results just yet.
err := m.queryAndConvertV1ToV3Resources(
data, model.NodeListOptions{}, converters.Node{}, noFilter,
)
if err != nil {
return err
}
// Query all of the per-node config and extract all of the IPIP tunnel addresses that are
// configured.
kvps, err := m.clientv1.List(model.HostConfigListOptions{Name: "IpInIpTunnelAddr"})
if err != nil {
switch err.(type) {
case cerrors.ErrorResourceDoesNotExist, cerrors.ErrorOperationNotSupported:
return nil
default:
return err
}
}
addrs := map[string]string{}
for _, kvp := range kvps {
k := kvp.Key.(model.HostConfigKey)
addrs[converters.ConvertNodeName(k.Hostname)] = kvp.Value.(string)
}
// Update the node resources to include the tunnel addresses. Loop through the converted
// resources and modify any node that has a corresponding tunnel address (it's a pointer
// so we can adjust the in-situ resource).
for _, r := range data.Resources {
if nr, ok := r.(*apiv3.Node); ok {
addr := addrs[nr.Name]
if addr == "" || nr.Spec.BGP == nil {
continue
}
nr.Spec.BGP.IPv4IPIPTunnelAddr = addr
}
}
return nil
}
// ShouldMigrate checks version information and reports if migration is needed
// and is possible. This checks both the v3 and v1 versions and indicates no migration
// if the v3 data is present and indicates a version of v3.0+. This method is used for
// automated KDD migrations where aborted upgrades use rollback to revert the previous
// system state. The rollback will remove the v3 data thus allowing aborted upgrades
// to be retried. See also CanMigrate below.
//
// If neither the v1 nor v3 versions are present, then no migration is required.
func (m *migrationHelper) ShouldMigrate() (bool, error) {
ci, err := m.clientv3.ClusterInformation().Get(context.Background(), "default", options.GetOptions{})
if err == nil && ci.Spec.CalicoVersion == "" {
// The ClusterInformation exists but the CalicoVersion field is empty. This may happen if a
// non-calico/node or typha component initializes the datastore prior to the node writing in
// the current version, or the node or typha completing the data migration.
log.Debug("ClusterInformation contained empty CalicoVersion - treating as if ClusterInformation is not present")
} else if err == nil {
// The ClusterInformation exists and the CalicoVersion field is not empty, check if migration is
// required.
if yes, err := versionRequiresMigration(ci.Spec.CalicoVersion); err != nil {
log.Errorf("Unexpected CalicoVersion '%s' in ClusterInformation: %v", ci.Spec.CalicoVersion, err)
return false, fmt.Errorf("unexpected CalicoVersion '%s' in ClusterInformation: %v", ci.Spec.CalicoVersion, err)
} else if yes {
log.Debugf("ClusterInformation contained CalicoVersion '%s' and indicates migration is needed", ci.Spec.CalicoVersion)
return true, nil
}
log.Debugf("ClusterInformation contained CalicoVersion '%s' and indicates migration is not needed", ci.Spec.CalicoVersion)
return false, nil
} else if _, ok := err.(cerrors.ErrorResourceDoesNotExist); !ok {
// The error indicates a problem with accessing the resource.
return false, fmt.Errorf("unable to query ClusterInformation to determine Calico version: %v", err)
}
// The resource does not exist from the clientv3 so we need to check the
// clientv1 version. Grab the version from the clientv1.
v, err := m.getV1ClusterVersion()
if err != nil {
if _, ok := err.(cerrors.ErrorResourceDoesNotExist); ok {
log.Debugf("CalicoVersion does exist in the v1 or v3 data, no migration needed")
// The resource does not exist in the clientv1 (or in the clientv3)
// so no migration is needed because it seems that this is an
// uninitialized datastore.
return false, nil
}
// The error indicates a problem accessing the resource.
return false, fmt.Errorf("unable to query global Felix configuration to determine Calico version: %v", err)
}
// Migrate only if it is possible to migrate from the current version.
if yes, err := versionRequiresMigration(v); err != nil {
// Hit an error parsing the version, or we determined that the version is not
// valid to migrate from.
log.Errorf("Unable to migrate data from version '%s': %v", v, err)
return false, fmt.Errorf("unable to migrate data from version '%s': %v", v, err)
} else if !yes {
// Version indicates that migration is not required - however, we should never have a
// v3.0+ version in the v1 API data which suggests a modified/hacked set up which we
// cannot migrate from.
log.Errorf("Unexpected version in the global Felix configuration, currently at %s", v)
return false, fmt.Errorf("unexpected Calico version '%s': migration to v3 should be from a tagged "+
"release of Calico v%s+", v, minUpgradeVersion)
}
log.Debugf("GlobalConfig contained CalicoVersion '%s' and indicates migration is needed", v)
return true, nil
}
// CanMigrate checks version information and reports if migration is possible (nil error).
// This differs from ShouldMigrate in that it only checks the v1 API - thus allowing aborted
// upgrades to be retried when the v3 data has not been removed. This method is used by
// the calico-upgrade script which performs additional checks to verify that the v3
// datastore is clean. If the v1 version is not present, this is considered an error case.
func (m *migrationHelper) CanMigrate() error {
m.status("Checking Calico version is suitable for migration")
v, err := m.getV1ClusterVersion()
if err != nil {
if _, ok := err.(cerrors.ErrorResourceDoesNotExist); ok {
m.statusBullet("no version found in the v1 API datastore")
return errors.New("unable to determine the Calico API version: no version information found - this may be " +
"due to a misconfiguration of the v1 API datastore access details")
}
// The error indicates a problem accessing the resource.
m.statusBullet("unable to query Calico version in the v1 API datastore: %v", err)
return err
}
// Migrate only if it is possible to migrate from the current version.
m.statusBullet("determined Calico version of: %s", v)
if yes, err := versionRequiresMigration(v); err != nil {
// Hit an error parsing the version, or we determined that the version is not
// valid to migrate from.
m.statusBullet("unable to parse version")
return fmt.Errorf("unable to migrate data from version '%s': %v", v, err)
} else if !yes {
// Version indicates that migration is not required - however, we should never have a
// v3.0+ version in the v1 API data which suggests a modified/hacked set up which we
// cannot migrate from.
m.statusBullet("unexpected version in the v1 API datastore: should not have a version v3.0+")
return fmt.Errorf("unexpected Calico version '%s': migration to v3 should be from a tagged "+
"release of Calico v%s+", v, minUpgradeVersion)
}
m.statusBullet("the v1 API data can be migrated to the v3 API")
return nil
}
// getV1ClusterVersion reads the CalicoVersion from the v1 client interface.
func (m *migrationHelper) getV1ClusterVersion() (string, error) {
kv, err := m.clientv1.Get(model.GlobalConfigKey{Name: "CalicoVersion"})
if err == nil {
return kv.Value.(string), nil
}
return "", err
}
// versionRequiresMigration returns true if the given version requires
// upgrating the data model, false otherwise. Returns an error if it was
// not possible to parse the version, or if the version was less than the
// minimum requires for upgrade.
func versionRequiresMigration(v string) (bool, error) {
sv, err := semver.NewVersion(strings.TrimPrefix(v, "v"))
if err != nil {
log.Warnf("unable to parse Calico version %s: %v", v, err)
return false, errors.New("unable to parse the version")
}
if sv.Major >= 3 {
log.Debugf("major version is already >= 3: %s", v)
// No need to migrate 3.0 or greater
return false, nil
}
// Omit the pre-release from the comparison - we allow pre-releases to enable testing
// of the upgrade.
sv.PreRelease = semver.PreRelease("")
// Compare the parsed version against the minimum required version.
svMin := semver.New(minUpgradeVersion)
if sv.Compare(*svMin) >= 0 {
// Version is greater than or equal to the minimum upgrade version, and less than 3.0
return true, nil
}
// Version is less than the minimum required version (including pre-releases). This is an
// error case and we cannot allow upgrade to continue.
log.Infof("cannot migrate from version %s: migration to v3 requires a tagged release of "+
"Calico v%s+", v, minUpgradeVersion)
return false, fmt.Errorf("migration to v3 requires a tagged release of Calico v%s+", minUpgradeVersion)
}
// convertLogLevel converts the v1 log level to the equivalent v3 log level. We
// ignore errors, defaulting to info in the event of a conversion error.
func convertLogLevel(logLevel string) string {
switch strings.ToLower(logLevel) {
case "debug":
return "Debug"
case "info":
return "Info"
case "warning":
return "Warning"
case "error":
return "Error"
case "fatal":
return "Fatal"
case "panic":
return "Fatal"
case "":
return ""
default:
return "Info"
}
}
// Display a 79-char word wrapped status message and log.
func (m *migrationHelper) status(format string, a ...interface{}) {
msg := fmt.Sprintf(format, a...)
log.Info(strings.TrimSpace(msg))
if m.statusWriter != nil {
m.statusWriter.Msg(msg)
}
}
// Display a 79-char word wrapped sub status (a bulleted message) and log.
func (m *migrationHelper) statusBullet(format string, a ...interface{}) {
msg := fmt.Sprintf(format, a...)
log.Info(strings.TrimSpace(msg))
if m.statusWriter != nil {
m.statusWriter.Bullet(msg)
}
}
// Display a 79-char word wrapped error message and log.
func (m *migrationHelper) statusError(format string, a ...interface{}) {
msg := fmt.Sprintf(format, a...)
log.Error(strings.TrimSpace(msg))
if m.statusWriter != nil {
m.statusWriter.Error(msg)
}
}
func (m *migrationHelper) clearReadyV1() error {
log.WithField("Ready", false).Info("Updating Ready flag in v1")
readyKV, err := m.clientv1.Get(model.ReadyFlagKey{})
if err != nil {
m.statusBullet("failed to get status of Calico networking in the v1 configuration")
return err
}
if !readyKV.Value.(bool) {
m.statusBullet("Calico networking already paused in the v1 configuration")
return fmt.Errorf("Calico networking already paused do not continue.")
}
readyKV.Value = false
_, err = m.clientv1.Update(readyKV)
if err != nil {
m.statusBullet("failed to pause Calico networking in the v1 configuration")
return err
}
m.statusBullet("successfully paused Calico networking in the v1 configuration")
return nil
}
// setReadyV1 sets the ready flag in the v1 datastore.
func (m *migrationHelper) setReadyV1() error {
log.WithField("Ready", true).Info("Updating Ready flag in v1")
_, err := m.clientv1.Apply(&model.KVPair{
Key: model.ReadyFlagKey{},
Value: true,
})
if err != nil {
m.statusBullet("failed to resume Calico networking in the v1 configuration")
return err
}
m.statusBullet("successfully resumed Calico networking in the v1 configuration")
return nil
}
// setReadyV3 sets the ready flag in the v3 datastore.
func (m *migrationHelper) setReadyV3(ready bool) error {
log.WithField("Ready", ready).Info("Updating Ready flag in v3")
c, err := m.clientv3.ClusterInformation().Get(context.Background(), "default", options.GetOptions{})
if err == nil {
// ClusterInformation already exists - update the settings.
log.WithField("Ready", ready).Info("Updating Ready flag in v3 ClusterInformation")
c.Spec.DatastoreReady = &ready
_, err = m.clientv3.ClusterInformation().Update(context.Background(), c, options.SetOptions{})
if err != nil {
log.WithError(err).Info("Hit error setting ready flag in v3 ClusterInformation")
if ready {
m.statusBullet("failed to resume Calico networking in the v3 configuration: %v", err)
} else {
m.statusBullet("failed to pause Calico networking in the v3 configuration: %v", err)
}
return err
}
if ready {
m.statusBullet("successfully resumed Calico networking in the v3 configuration (updated ClusterInformation)")
} else {
m.statusBullet("successfully paused Calico networking in the v3 configuration (updated ClusterInformation)")
}
return nil
}
if _, ok := err.(cerrors.ErrorResourceDoesNotExist); !ok {
// ClusterInformation could not be queried.
if ready {
m.statusBullet("failed to resume Calico networking in the v3 configuration (unable to query ClusterInformation)")
} else {
m.statusBullet("failed to pause Calico networking in the v3 configuration (unable to query ClusterInformation)")
}
return err
}
// ClusterInformation does not exist - create a new one.
c = apiv3.NewClusterInformation()
c.Name = "default"
c.Spec.DatastoreReady = &ready
_, err = m.clientv3.ClusterInformation().Create(context.Background(), c, options.SetOptions{})
if err != nil {
if ready {
m.statusBullet("failed to resume Calico networking in the v3 configuration (unable to create ClusterInformation)")
} else {
m.statusBullet("failed to pause Calico networking in the v3 configuration (unable to create ClusterInformation)")
}
return err
}
if ready {
m.statusBullet("successfully resumed Calico networking in the v1 configuration (created ClusterInformation)")
} else {
m.statusBullet("successfully paused Calico networking in the v1 configuration (created ClusterInformation)")
}
return nil
}
// isReady reads the Ready flag from the datastore and returns its value
func (m *migrationHelper) isReady() (bool, error) {
kv, err := m.clientv1.Get(model.ReadyFlagKey{})
if err != nil {
m.statusError("Unable to query the v1 datastore for ready status")
m.statusBullet("Cause: %v", err)
return false, err
}
if kv.Value.(bool) {
m.statusBullet("Ready flag is true.")
} else {
m.statusBullet("Ready flag is false.")
}
return kv.Value.(bool), nil
}