-
Notifications
You must be signed in to change notification settings - Fork 457
/
worker.go
1857 lines (1606 loc) · 60.8 KB
/
worker.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 2019 Altinity Ltd and/or its affiliates. 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 chi
import (
"context"
"errors"
"fmt"
"time"
"github.com/juliangruber/go-intersect"
core "k8s.io/api/core/v1"
apiErrors "k8s.io/apimachinery/pkg/api/errors"
"k8s.io/apimachinery/pkg/api/resource"
meta "k8s.io/apimachinery/pkg/apis/meta/v1"
utilRuntime "k8s.io/apimachinery/pkg/util/runtime"
"github.com/altinity/queue"
log "github.com/altinity/clickhouse-operator/pkg/announcer"
api "github.com/altinity/clickhouse-operator/pkg/apis/clickhouse.altinity.com/v1"
"github.com/altinity/clickhouse-operator/pkg/apis/deployment"
"github.com/altinity/clickhouse-operator/pkg/chop"
"github.com/altinity/clickhouse-operator/pkg/controller"
model "github.com/altinity/clickhouse-operator/pkg/model/chi"
chiCreator "github.com/altinity/clickhouse-operator/pkg/model/chi/creator"
"github.com/altinity/clickhouse-operator/pkg/model/chi/normalizer"
"github.com/altinity/clickhouse-operator/pkg/model/chi/schemer"
"github.com/altinity/clickhouse-operator/pkg/model/clickhouse"
"github.com/altinity/clickhouse-operator/pkg/model/k8s"
"github.com/altinity/clickhouse-operator/pkg/util"
)
// FinalizerName specifies name of the finalizer to be used with CHI
const FinalizerName = "finalizer.clickhouseinstallation.altinity.com"
// worker represents worker thread which runs reconcile tasks
type worker struct {
c *Controller
a Announcer
//queue workqueue.RateLimitingInterface
queue queue.PriorityQueue
normalizer *normalizer.Normalizer
schemer *schemer.ClusterSchemer
start time.Time
task task
}
// task represents context of a worker. This also can be called "a reconcile task"
type task struct {
creator *chiCreator.Creator
registryReconciled *model.Registry
registryFailed *model.Registry
cmUpdate time.Time
start time.Time
}
// newTask creates new context
func newTask(creator *chiCreator.Creator) task {
return task{
creator: creator,
registryReconciled: model.NewRegistry(),
registryFailed: model.NewRegistry(),
cmUpdate: time.Time{},
start: time.Now(),
}
}
// newWorker
// func (c *Controller) newWorker(q workqueue.RateLimitingInterface) *worker {
func (c *Controller) newWorker(q queue.PriorityQueue, sys bool) *worker {
start := time.Now()
if !sys {
start = start.Add(api.DefaultReconcileThreadsWarmup)
}
return &worker{
c: c,
a: NewAnnouncer().WithController(c),
queue: q,
normalizer: normalizer.NewNormalizer(func(namespace, name string) (*core.Secret, error) {
return c.kubeClient.CoreV1().Secrets(namespace).Get(context.TODO(), name, controller.NewGetOptions())
}),
schemer: nil,
start: start,
}
}
// newContext creates new reconcile task
func (w *worker) newTask(chi *api.ClickHouseInstallation) {
w.task = newTask(chiCreator.NewCreator(chi))
}
// timeToStart specifies time that operator does not accept changes
const timeToStart = 1 * time.Minute
// isJustStarted checks whether worked just started
func (w *worker) isJustStarted() bool {
return time.Since(w.start) < timeToStart
}
func (w *worker) isConfigurationChangeRequiresReboot(host *api.ChiHost) bool {
return model.IsConfigurationChangeRequiresReboot(host)
}
// shouldForceRestartHost checks whether cluster requires hosts restart
func (w *worker) shouldForceRestartHost(host *api.ChiHost) bool {
// RollingUpdate purpose is to always shut the host down.
// It is such an interesting policy.
if host.GetCHI().IsRollingUpdate() {
w.a.V(1).M(host).F().Info("RollingUpdate requires force restart. Host: %s", host.GetName())
return true
}
if host.GetReconcileAttributes().GetStatus() == api.ObjectStatusNew {
w.a.V(1).M(host).F().Info("Host is new, no restart applicable. Host: %s", host.GetName())
return false
}
if (host.GetReconcileAttributes().GetStatus() == api.ObjectStatusSame) && !host.HasAncestor() {
w.a.V(1).M(host).F().Info("Host already exists, but has no ancestor, no restart applicable. Host: %s", host.GetName())
return false
}
// For some configuration changes we have to force restart host
if w.isConfigurationChangeRequiresReboot(host) {
w.a.V(1).M(host).F().Info("Config change(s) require host restart. Host: %s", host.GetName())
return true
}
podIsCrushed := false
// pod.Status.ContainerStatuses[0].State.Waiting.Reason
if pod, err := w.c.getPod(host); err == nil {
if len(pod.Status.ContainerStatuses) > 0 {
if pod.Status.ContainerStatuses[0].State.Waiting != nil {
if pod.Status.ContainerStatuses[0].State.Waiting.Reason == "CrashLoopBackOff" {
podIsCrushed = true
}
}
}
}
if host.Runtime.Version.IsUnknown() && podIsCrushed {
w.a.V(1).M(host).F().Info("Host with unknown version and in CrashLoopBackOff should be restarted. It most likely is unable to start due to bad config. Host: %s", host.GetName())
return true
}
w.a.V(1).M(host).F().Info("Host restart is not required. Host: %s", host.GetName())
return false
}
// run is an endless work loop, expected to be run in a thread
func (w *worker) run() {
w.a.V(2).S().P()
defer w.a.V(2).E().P()
// For system thread let's wait its 'official start time', thus giving it time to bootstrap
util.WaitContextDoneUntil(context.Background(), w.start)
// Events loop
for {
// Get() blocks until it can return an item
item, ctx, ok := w.queue.Get()
if !ok {
w.a.Info("shutdown request")
return
}
//item, shut := w.queue.Get()
//task := context.Background()
//if shut {
// w.a.Info("shutdown request")
// return
//}
if err := w.processItem(ctx, item); err != nil {
// Item not processed
// this code cannot return an error and needs to indicate error has been ignored
utilRuntime.HandleError(err)
}
// Forget indicates that an item is finished being retried. Doesn't matter whether its for perm failing
// or for success, we'll stop the rate limiter from tracking it. This only clears the `rateLimiter`, you
// still have to call `Done` on the queue.
//w.queue.Forget(item)
// Remove item from processing set when processing completed
w.queue.Done(item)
}
}
func (w *worker) processReconcileCHI(ctx context.Context, cmd *ReconcileCHI) error {
switch cmd.cmd {
case reconcileAdd:
return w.updateCHI(ctx, nil, cmd.new)
case reconcileUpdate:
return w.updateCHI(ctx, cmd.old, cmd.new)
case reconcileDelete:
return w.discoveryAndDeleteCHI(ctx, cmd.old)
}
// Unknown item type, don't know what to do with it
// Just skip it and behave like it never existed
utilRuntime.HandleError(fmt.Errorf("unexpected reconcile - %#v", cmd))
return nil
}
func (w *worker) processReconcileCHIT(cmd *ReconcileCHIT) error {
switch cmd.cmd {
case reconcileAdd:
return w.addChit(cmd.new)
case reconcileUpdate:
return w.updateChit(cmd.old, cmd.new)
case reconcileDelete:
return w.deleteChit(cmd.old)
}
// Unknown item type, don't know what to do with it
// Just skip it and behave like it never existed
utilRuntime.HandleError(fmt.Errorf("unexpected reconcile - %#v", cmd))
return nil
}
func (w *worker) processReconcileChopConfig(cmd *ReconcileChopConfig) error {
switch cmd.cmd {
case reconcileAdd:
return w.c.addChopConfig(cmd.new)
case reconcileUpdate:
return w.c.updateChopConfig(cmd.old, cmd.new)
case reconcileDelete:
return w.c.deleteChopConfig(cmd.old)
}
// Unknown item type, don't know what to do with it
// Just skip it and behave like it never existed
utilRuntime.HandleError(fmt.Errorf("unexpected reconcile - %#v", cmd))
return nil
}
func (w *worker) processReconcileEndpoints(ctx context.Context, cmd *ReconcileEndpoints) error {
switch cmd.cmd {
case reconcileUpdate:
return w.updateEndpoints(ctx, cmd.old, cmd.new)
}
// Unknown item type, don't know what to do with it
// Just skip it and behave like it never existed
utilRuntime.HandleError(fmt.Errorf("unexpected reconcile - %#v", cmd))
return nil
}
func (w *worker) processReconcilePod(ctx context.Context, cmd *ReconcilePod) error {
switch cmd.cmd {
case reconcileAdd:
w.a.V(1).M(cmd.new).F().Info("Add Pod. %s/%s", cmd.new.Namespace, cmd.new.Name)
metricsPodAdd(ctx)
return nil
case reconcileUpdate:
//ignore
//w.a.V(1).M(cmd.new).F().Info("Update Pod. %s/%s", cmd.new.Namespace, cmd.new.Name)
//metricsPodUpdate(ctx)
return nil
case reconcileDelete:
w.a.V(1).M(cmd.old).F().Info("Delete Pod. %s/%s", cmd.old.Namespace, cmd.old.Name)
metricsPodDelete(ctx)
return nil
}
// Unknown item type, don't know what to do with it
// Just skip it and behave like it never existed
utilRuntime.HandleError(fmt.Errorf("unexpected reconcile - %#v", cmd))
return nil
}
func (w *worker) processDropDns(ctx context.Context, cmd *DropDns) error {
if chi, err := w.createCHIFromObjectMeta(cmd.initiator, false, normalizer.NewOptions()); err == nil {
w.a.V(2).M(cmd.initiator).Info("flushing DNS for CHI %s", chi.Name)
_ = w.ensureClusterSchemer(chi.FirstHost()).CHIDropDnsCache(ctx, chi)
} else {
w.a.M(cmd.initiator).F().Error("unable to find CHI by %v err: %v", cmd.initiator.Labels, err)
}
return nil
}
// processItem processes one work item according to its type
func (w *worker) processItem(ctx context.Context, item interface{}) error {
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return nil
}
w.a.V(3).S().P()
defer w.a.V(3).E().P()
switch cmd := item.(type) {
case *ReconcileCHI:
return w.processReconcileCHI(ctx, cmd)
case *ReconcileCHIT:
return w.processReconcileCHIT(cmd)
case *ReconcileChopConfig:
return w.processReconcileChopConfig(cmd)
case *ReconcileEndpoints:
return w.processReconcileEndpoints(ctx, cmd)
case *ReconcilePod:
return w.processReconcilePod(ctx, cmd)
case *DropDns:
return w.processDropDns(ctx, cmd)
}
// Unknown item type, don't know what to do with it
// Just skip it and behave like it never existed
utilRuntime.HandleError(fmt.Errorf("unexpected item in the queue - %#v", item))
return nil
}
// normalize
func (w *worker) normalize(c *api.ClickHouseInstallation) *api.ClickHouseInstallation {
chi, err := w.normalizer.CreateTemplatedCHI(c, normalizer.NewOptions())
if err != nil {
w.a.WithEvent(chi, eventActionReconcile, eventReasonReconcileFailed).
WithStatusError(chi).
M(chi).F().
Error("FAILED to normalize CHI 1: %v", err)
}
ips := w.c.getPodsIPs(chi)
w.a.V(1).M(chi).Info("IPs of the CHI normalizer %s/%s: len: %d %v", chi.Namespace, chi.Name, len(ips), ips)
opts := normalizer.NewOptions()
opts.DefaultUserAdditionalIPs = ips
chi, err = w.normalizer.CreateTemplatedCHI(c, opts)
if err != nil {
w.a.WithEvent(chi, eventActionReconcile, eventReasonReconcileFailed).
WithStatusError(chi).
M(chi).F().
Error("FAILED to normalize CHI 2: %v", err)
}
return chi
}
// ensureFinalizer
func (w *worker) ensureFinalizer(ctx context.Context, chi *api.ClickHouseInstallation) bool {
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return false
}
// In case CHI is being deleted already, no need to meddle with finalizers
if !chi.ObjectMeta.DeletionTimestamp.IsZero() {
return false
}
// Finalizer can already be listed in CHI, do nothing in this case
if util.InArray(FinalizerName, chi.ObjectMeta.Finalizers) {
w.a.V(2).M(chi).F().Info("finalizer already installed")
return false
}
// No finalizer found - need to install it
if err := w.c.installFinalizer(ctx, chi); err != nil {
w.a.V(1).M(chi).F().Error("unable to install finalizer. err: %v", err)
return false
}
w.a.V(3).M(chi).F().Info("finalizer installed")
return true
}
// updateEndpoints updates endpoints
func (w *worker) updateEndpoints(ctx context.Context, old, new *core.Endpoints) error {
if chi, err := w.createCHIFromObjectMeta(&new.ObjectMeta, false, normalizer.NewOptions()); err == nil {
w.a.V(1).M(chi).Info("updating endpoints for CHI-1 %s", chi.Name)
ips := w.c.getPodsIPs(chi)
w.a.V(1).M(chi).Info("IPs of the CHI-1 update endpoints %s/%s: len: %d %v", chi.Namespace, chi.Name, len(ips), ips)
opts := normalizer.NewOptions()
opts.DefaultUserAdditionalIPs = ips
if chi, err := w.createCHIFromObjectMeta(&new.ObjectMeta, false, opts); err == nil {
w.a.V(1).M(chi).Info("Update users IPS-1")
// TODO unify with finalize reconcile
w.newTask(chi)
w.reconcileCHIConfigMapUsers(ctx, chi)
w.c.updateCHIObjectStatus(ctx, chi, UpdateCHIStatusOptions{
TolerateAbsence: true,
CopyCHIStatusOptions: api.CopyCHIStatusOptions{
Normalized: true,
},
})
} else {
w.a.M(&new.ObjectMeta).F().Error("internal unable to find CHI by %v err: %v", new.Labels, err)
}
} else {
w.a.M(&new.ObjectMeta).F().Error("external unable to find CHI by %v err %v", new.Labels, err)
}
return nil
}
// updateCHI sync CHI which was already created earlier
func (w *worker) updateCHI(ctx context.Context, old, new *api.ClickHouseInstallation) error {
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return nil
}
w.a.V(3).M(new).S().P()
defer w.a.V(3).M(new).E().P()
update := (old != nil) && (new != nil)
if update && (old.ObjectMeta.ResourceVersion == new.ObjectMeta.ResourceVersion) {
// No need to react
w.a.V(3).M(new).F().Info("ResourceVersion did not change: %s", new.ObjectMeta.ResourceVersion)
return nil
}
w.a.V(1).M(new).S().P()
defer w.a.V(1).M(new).E().P()
if w.ensureFinalizer(context.Background(), new) {
w.a.M(new).F().Info("finalizer installed, let's restart reconcile cycle. CHI: %s/%s", new.Namespace, new.Name)
w.a.M(new).F().Info("---------------------------------------------------------------------")
return nil
} else {
w.a.M(new).F().Info("finalizer in place, proceed to reconcile cycle. CHI: %s/%s", new.Namespace, new.Name)
}
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return nil
}
if w.deleteCHI(ctx, old, new) {
// CHI is being deleted
return nil
}
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return nil
}
if w.isCHIProcessedOnTheSameIP(new) {
// First minute after restart do not reconcile already reconciled generations
w.a.V(1).M(new).F().Info("Will not reconcile known generation after restart. Generation %d", new.Generation)
return nil
}
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return nil
}
// CHI is being reconciled
return w.reconcileCHI(ctx, old, new)
}
// isCHIProcessedOnTheSameIP checks whether it is just a restart of the operator on the same IP
func (w *worker) isCHIProcessedOnTheSameIP(chi *api.ClickHouseInstallation) bool {
ip, _ := chop.Get().ConfigManager.GetRuntimeParam(deployment.OPERATOR_POD_IP)
operatorIpIsTheSame := ip == chi.Status.GetCHOpIP()
log.V(1).Info("Operator IPs to process CHI: %s. Previous: %s Cur: %s", chi.Name, chi.Status.GetCHOpIP(), ip)
if !operatorIpIsTheSame {
// Operator has restarted on the different IP address.
// We may need to reconcile config files
log.V(1).Info("Operator IPs are different. Operator was restarted on another IP since previous reconcile of the CHI: %s", chi.Name)
return false
}
log.V(1).Info("Operator IPs are the same as on previous reconcile of the CHI: %s", chi.Name)
return w.isCleanRestart(chi)
}
// isCleanRestart checks whether it is just a restart of the operator and CHI has no changes since last processed
func (w *worker) isCleanRestart(chi *api.ClickHouseInstallation) bool {
// Clean restart may be only in case operator has just recently started
if !w.isJustStarted() {
log.V(1).Info("Operator is not just started. May not be clean restart")
return false
}
log.V(1).Info("Operator just started. May be clean restart")
// Migration support
// Do we have have previously completed CHI?
// In case no - this means that CHI has either not completed or we are migrating from
// such a version of the operator, where there is no completed CHI at all
noCompletedCHI := !chi.HasAncestor()
// Having status completed and not having completed CHI suggests we are migrating operator version
statusIsCompleted := chi.Status.GetStatus() == api.StatusCompleted
if noCompletedCHI && statusIsCompleted {
// In case of a restart - assume that normalized is already completed
chi.SetAncestor(chi.GetTarget())
}
// Check whether anything has changed in CHI spec
// In case the generation is the same as already completed - it is clean restart
generationIsOk := false
// However, completed CHI still can be missing, for example, in newly requested CHI
if chi.HasAncestor() {
generationIsOk = chi.Generation == chi.GetAncestor().Generation
log.V(1).Info(
"CHI %s has ancestor. Generations. Prev: %d Cur: %d Generation is the same: %t",
chi.Name,
chi.GetAncestor().Generation,
chi.Generation,
generationIsOk,
)
} else {
log.V(1).Info("CHI %s has NO ancestor, meaning reconcile cycle was never completed.", chi.Name)
}
log.V(1).Info("Is CHI %s clean on operator restart: %t", chi.Name, generationIsOk)
return generationIsOk
}
// areUsableOldAndNew checks whether there are old and new usable
func (w *worker) areUsableOldAndNew(old, new *api.ClickHouseInstallation) bool {
if old == nil {
return false
}
if new == nil {
return false
}
return true
}
// isAfterFinalizerInstalled checks whether we are just installed finalizer
func (w *worker) isAfterFinalizerInstalled(old, new *api.ClickHouseInstallation) bool {
if !w.areUsableOldAndNew(old, new) {
return false
}
finalizerIsInstalled := len(old.Finalizers) == 0 && len(new.Finalizers) > 0
return w.isGenerationTheSame(old, new) && finalizerIsInstalled
}
// isGenerationTheSame checks whether old ans new CHI have the same generation
func (w *worker) isGenerationTheSame(old, new *api.ClickHouseInstallation) bool {
if !w.areUsableOldAndNew(old, new) {
return false
}
return old.Generation == new.Generation
}
// logCHI writes a CHI into the log
func (w *worker) logCHI(name string, chi *api.ClickHouseInstallation) {
w.a.V(1).M(chi).Info(
"logCHI %s start--------------------------------------------:\n%s\nlogCHI %s end--------------------------------------------",
name,
name,
chi.YAML(api.CopyCHIOptions{SkipStatus: true, SkipManagedFields: true}),
)
}
// logActionPlan logs action plan
func (w *worker) logActionPlan(ap *model.ActionPlan) {
w.a.Info(
"ActionPlan start---------------------------------------------:\n%s\nActionPlan end---------------------------------------------",
ap,
)
}
// logOldAndNew writes old and new CHIs into the log
func (w *worker) logOldAndNew(name string, old, new *api.ClickHouseInstallation) {
w.logCHI(name+" old", old)
w.logCHI(name+" new", new)
}
func (w *worker) waitForIPAddresses(ctx context.Context, chi *api.ClickHouseInstallation) {
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return
}
if chi.IsStopped() {
// No need to wait for stopped CHI
return
}
w.a.V(1).M(chi).F().S().Info("wait for IP addresses to be assigned to all pods")
start := time.Now()
w.c.poll(ctx, chi, func(c *api.ClickHouseInstallation, e error) bool {
if len(c.Status.GetPodIPs()) >= len(c.Status.GetPods()) {
// Stop polling
w.a.V(1).M(c).Info("all IP addresses are in place")
return false
}
if time.Now().Sub(start) > 1*time.Minute {
// Stop polling
w.a.V(1).M(c).Warning("not all IP addresses are in place but time has elapsed")
return false
}
// Continue polling
w.a.V(1).M(c).Warning("still waiting - not all IP addresses are in place yet")
return true
})
}
// excludeStoppedCHIFromMonitoring excludes stopped CHI from monitoring
func (w *worker) excludeStoppedCHIFromMonitoring(chi *api.ClickHouseInstallation) {
if !chi.IsStopped() {
// No need to exclude non-stopped CHI
return
}
w.a.V(1).
WithEvent(chi, eventActionReconcile, eventReasonReconcileInProgress).
WithStatusAction(chi).
M(chi).F().
Info("exclude CHI from monitoring")
w.c.deleteWatch(chi)
}
// addCHIToMonitoring adds CHI to monitoring
func (w *worker) addCHIToMonitoring(chi *api.ClickHouseInstallation) {
if chi.IsStopped() {
// No need to add stopped CHI
return
}
w.a.V(1).
WithEvent(chi, eventActionReconcile, eventReasonReconcileInProgress).
WithStatusAction(chi).
M(chi).F().
Info("add CHI to monitoring")
w.c.updateWatch(chi)
}
func (w *worker) markReconcileStart(ctx context.Context, chi *api.ClickHouseInstallation, ap *model.ActionPlan) {
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return
}
// Write desired normalized CHI with initialized .Status, so it would be possible to monitor progress
chi.EnsureStatus().ReconcileStart(ap.GetRemovedHostsNum())
_ = w.c.updateCHIObjectStatus(ctx, chi, UpdateCHIStatusOptions{
CopyCHIStatusOptions: api.CopyCHIStatusOptions{
MainFields: true,
},
})
w.a.V(1).
WithEvent(chi, eventActionReconcile, eventReasonReconcileStarted).
WithStatusAction(chi).
WithStatusActions(chi).
M(chi).F().
Info("reconcile started, task id: %s", chi.Spec.GetTaskID())
w.a.V(2).M(chi).F().Info("action plan\n%s\n", ap.String())
}
func (w *worker) finalizeReconcileAndMarkCompleted(ctx context.Context, _chi *api.ClickHouseInstallation) {
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return
}
w.a.V(1).M(_chi).F().S().Info("finalize reconcile")
// Update CHI object
if chi, err := w.createCHIFromObjectMeta(&_chi.ObjectMeta, true, normalizer.NewOptions()); err == nil {
w.a.V(1).M(chi).Info("updating endpoints for CHI-2 %s", chi.Name)
ips := w.c.getPodsIPs(chi)
w.a.V(1).M(chi).Info("IPs of the CHI-2 finalize reconcile %s/%s: len: %d %v", chi.Namespace, chi.Name, len(ips), ips)
opts := normalizer.NewOptions()
opts.DefaultUserAdditionalIPs = ips
if chi, err := w.createCHIFromObjectMeta(&_chi.ObjectMeta, true, opts); err == nil {
w.a.V(1).M(chi).Info("Update users IPS-2")
chi.SetAncestor(chi.GetTarget())
chi.SetTarget(nil)
chi.EnsureStatus().ReconcileComplete()
// TODO unify with update endpoints
w.newTask(chi)
w.reconcileCHIConfigMapUsers(ctx, chi)
w.c.updateCHIObjectStatus(ctx, chi, UpdateCHIStatusOptions{
CopyCHIStatusOptions: api.CopyCHIStatusOptions{
WholeStatus: true,
},
})
} else {
w.a.M(&_chi.ObjectMeta).F().Error("internal unable to find CHI by %v err: %v", _chi.Labels, err)
}
} else {
w.a.M(&_chi.ObjectMeta).F().Error("external unable to find CHI by %v err %v", _chi.Labels, err)
}
w.a.V(1).
WithEvent(_chi, eventActionReconcile, eventReasonReconcileCompleted).
WithStatusAction(_chi).
WithStatusActions(_chi).
M(_chi).F().
Info("reconcile completed successfully, task id: %s", _chi.Spec.GetTaskID())
}
func (w *worker) markReconcileCompletedUnsuccessfully(ctx context.Context, chi *api.ClickHouseInstallation, err error) {
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return
}
switch {
case err == nil:
chi.EnsureStatus().ReconcileComplete()
case errors.Is(err, errCRUDAbort):
chi.EnsureStatus().ReconcileAbort()
}
w.c.updateCHIObjectStatus(ctx, chi, UpdateCHIStatusOptions{
CopyCHIStatusOptions: api.CopyCHIStatusOptions{
MainFields: true,
},
})
w.a.V(1).
WithEvent(chi, eventActionReconcile, eventReasonReconcileFailed).
WithStatusAction(chi).
WithStatusActions(chi).
M(chi).F().
Warning("reconcile completed UNSUCCESSFULLY, task id: %s", chi.Spec.GetTaskID())
}
func (w *worker) walkHosts(ctx context.Context, chi *api.ClickHouseInstallation, ap *model.ActionPlan) {
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return
}
objs := w.c.discovery(ctx, chi)
ap.WalkAdded(
// Walk over added clusters
func(cluster *api.Cluster) {
cluster.WalkHosts(func(host *api.ChiHost) error {
// Name of the StatefulSet for this host
name := model.CreateStatefulSetName(host)
// Have we found this StatefulSet
found := false
objs.WalkStatefulSet(func(meta meta.ObjectMeta) {
if name == meta.Name {
// StatefulSet of this host already exist
found = true
}
})
if found {
// StatefulSet of this host already exist, we can't ADD it for sure
// It looks like FOUND is the most correct approach
host.GetReconcileAttributes().SetFound()
w.a.V(1).M(chi).Info("Add host as FOUND. Host was found as sts %s", host.GetName())
} else {
// StatefulSet of this host does not exist, looks like we need to ADD it
host.GetReconcileAttributes().SetAdd()
w.a.V(1).M(chi).Info("Add host as ADD. Host was not found as sts %s", host.GetName())
}
return nil
})
},
// Walk over added shards
func(shard *api.ChiShard) {
// Mark all hosts of the shard as newly added
shard.WalkHosts(func(host *api.ChiHost) error {
host.GetReconcileAttributes().SetAdd()
return nil
})
},
// Walk over added hosts
func(host *api.ChiHost) {
host.GetReconcileAttributes().SetAdd()
},
)
ap.WalkModified(
func(cluster *api.Cluster) {
},
func(shard *api.ChiShard) {
},
func(host *api.ChiHost) {
host.GetReconcileAttributes().SetModify()
},
)
chi.WalkHosts(func(host *api.ChiHost) error {
switch {
case host.GetReconcileAttributes().IsAdd():
// Already added
return nil
case host.GetReconcileAttributes().IsModify():
// Already modified
return nil
default:
// Not clear yet
host.GetReconcileAttributes().SetFound()
}
return nil
})
chi.WalkHosts(func(host *api.ChiHost) error {
switch {
case host.GetReconcileAttributes().IsAdd():
w.a.M(host).Info("ADD host: %s", host.Runtime.Address.CompactString())
case host.GetReconcileAttributes().IsModify():
w.a.M(host).Info("MODIFY host: %s", host.Runtime.Address.CompactString())
case host.GetReconcileAttributes().IsFound():
w.a.M(host).Info("FOUND host: %s", host.Runtime.Address.CompactString())
default:
w.a.M(host).Info("UNKNOWN host: %s", host.Runtime.Address.CompactString())
}
return nil
})
}
// getRemoteServersGeneratorOptions build base set of RemoteServersGeneratorOptions
// which are applied on each of `remote_servers` reconfiguration during reconcile cycle
func (w *worker) getRemoteServersGeneratorOptions() *model.RemoteServersGeneratorOptions {
// Base model.RemoteServersGeneratorOptions specifies to exclude:
// 1. all newly added hosts
// 2. all explicitly excluded hosts
return model.NewRemoteServersGeneratorOptions().ExcludeReconcileAttributes(
api.NewChiHostReconcileAttributes().
SetAdd().
SetExclude(),
)
}
// options build ClickHouseConfigFilesGeneratorOptions
func (w *worker) options() *model.ClickHouseConfigFilesGeneratorOptions {
opts := w.getRemoteServersGeneratorOptions()
w.a.Info("RemoteServersGeneratorOptions: %s", opts)
return model.NewClickHouseConfigFilesGeneratorOptions().SetRemoteServersGeneratorOptions(opts)
}
// prepareHostStatefulSetWithStatus prepares host's StatefulSet status
func (w *worker) prepareHostStatefulSetWithStatus(ctx context.Context, host *api.ChiHost, shutdown bool) {
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return
}
w.prepareDesiredStatefulSet(host, shutdown)
host.GetReconcileAttributes().SetStatus(w.getStatefulSetStatus(host))
}
// prepareDesiredStatefulSet prepares desired StatefulSet
func (w *worker) prepareDesiredStatefulSet(host *api.ChiHost, shutdown bool) {
host.Runtime.DesiredStatefulSet = w.task.creator.CreateStatefulSet(host, shutdown)
}
type migrateTableOptions struct {
forceMigrate bool
dropReplica bool
}
func (o *migrateTableOptions) ForceMigrate() bool {
if o == nil {
return false
}
return o.forceMigrate
}
func (o *migrateTableOptions) DropReplica() bool {
if o == nil {
return false
}
return o.dropReplica
}
type migrateTableOptionsArr []*migrateTableOptions
// NewMigrateTableOptionsArr creates new migrateTableOptions array
func NewMigrateTableOptionsArr(opts ...*migrateTableOptions) (res migrateTableOptionsArr) {
return append(res, opts...)
}
// First gets first option
func (a migrateTableOptionsArr) First() *migrateTableOptions {
if len(a) > 0 {
return a[0]
}
return nil
}
// migrateTables
func (w *worker) migrateTables(ctx context.Context, host *api.ChiHost, opts ...*migrateTableOptions) error {
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return nil
}
if !w.shouldMigrateTables(host, opts...) {
w.a.V(1).
M(host).F().
Info(
"No need to add tables on host %d to shard %d in cluster %s",
host.Runtime.Address.ReplicaIndex, host.Runtime.Address.ShardIndex, host.Runtime.Address.ClusterName)
return nil
}
// Need to migrate tables
if w.shouldDropReplica(host, opts...) {
w.a.V(1).
M(host).F().
Info(
"Need to drop replica on host %d to shard %d in cluster %s",
host.Runtime.Address.ReplicaIndex, host.Runtime.Address.ShardIndex, host.Runtime.Address.ClusterName)
w.dropReplica(ctx, host, &dropReplicaOptions{forceDrop: true})
}
w.a.V(1).
WithEvent(host.GetCHI(), eventActionCreate, eventReasonCreateStarted).
WithStatusAction(host.GetCHI()).
M(host).F().
Info(
"Adding tables on shard/host:%d/%d cluster:%s",
host.Runtime.Address.ShardIndex, host.Runtime.Address.ReplicaIndex, host.Runtime.Address.ClusterName)
err := w.ensureClusterSchemer(host).HostCreateTables(ctx, host)
if err == nil {
w.a.V(1).
WithEvent(host.GetCHI(), eventActionCreate, eventReasonCreateCompleted).
WithStatusAction(host.GetCHI()).
M(host).F().
Info("Tables added successfully on shard/host:%d/%d cluster:%s",
host.Runtime.Address.ShardIndex, host.Runtime.Address.ReplicaIndex, host.Runtime.Address.ClusterName)
host.GetCHI().EnsureStatus().PushHostTablesCreated(model.CreateFQDN(host))
} else {
w.a.V(1).
WithEvent(host.GetCHI(), eventActionCreate, eventReasonCreateFailed).
WithStatusAction(host.GetCHI()).
M(host).F().
Error("ERROR add tables added successfully on shard/host:%d/%d cluster:%s err:%v",
host.Runtime.Address.ShardIndex, host.Runtime.Address.ReplicaIndex, host.Runtime.Address.ClusterName, err)
}
return err
}
// shouldMigrateTables
func (w *worker) shouldMigrateTables(host *api.ChiHost, opts ...*migrateTableOptions) bool {
o := NewMigrateTableOptionsArr(opts...).First()
// Deal with special cases in order of priority
switch {
case host.IsStopped():
// Stopped host is not able to receive any data, migration is inapplicable
return false
case o.ForceMigrate():
// Force migration requested
return true
case model.HostHasTablesCreated(host):
// This host is listed as having tables created already, no need to migrate again
return false
case model.HostIsNewOne(host):
// CHI is new, all hosts were added
return false
}
// In all the rest cases - perform migration
return true
}
// shouldDropTables
func (w *worker) shouldDropReplica(host *api.ChiHost, opts ...*migrateTableOptions) bool {
o := NewMigrateTableOptionsArr(opts...).First()
// Deal with special cases
switch {
case o.DropReplica():
return true
}
return false
}
// excludeHost excludes host from ClickHouse clusters if required
func (w *worker) excludeHost(ctx context.Context, host *api.ChiHost) error {
if util.IsContextDone(ctx) {
log.V(2).Info("task is done")
return nil
}