/
instance_controller.go
777 lines (688 loc) · 32.6 KB
/
instance_controller.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
// Copyright 2021 Google LLC
//
// 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 instancecontroller
import (
"context"
"fmt"
"strings"
"sync"
"time"
"github.com/go-logr/logr"
"github.com/google/go-cmp/cmp"
appsv1 "k8s.io/api/apps/v1"
corev1 "k8s.io/api/core/v1"
apierrors "k8s.io/apimachinery/pkg/api/errors"
v1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/runtime"
"k8s.io/apimachinery/pkg/types"
"k8s.io/client-go/tools/record"
ctrl "sigs.k8s.io/controller-runtime"
"sigs.k8s.io/controller-runtime/pkg/builder"
"sigs.k8s.io/controller-runtime/pkg/client"
"sigs.k8s.io/controller-runtime/pkg/controller"
"sigs.k8s.io/controller-runtime/pkg/controller/controllerutil"
"sigs.k8s.io/controller-runtime/pkg/event"
"sigs.k8s.io/controller-runtime/pkg/handler"
"sigs.k8s.io/controller-runtime/pkg/predicate"
"sigs.k8s.io/controller-runtime/pkg/reconcile"
"sigs.k8s.io/controller-runtime/pkg/source"
commonv1alpha1 "github.com/GoogleCloudPlatform/elcarro-oracle-operator/common/api/v1alpha1"
commonutils "github.com/GoogleCloudPlatform/elcarro-oracle-operator/common/pkg/utils"
"github.com/GoogleCloudPlatform/elcarro-oracle-operator/oracle/api/v1alpha1"
"github.com/GoogleCloudPlatform/elcarro-oracle-operator/oracle/controllers"
"github.com/GoogleCloudPlatform/elcarro-oracle-operator/oracle/pkg/agents/consts"
dbdpb "github.com/GoogleCloudPlatform/elcarro-oracle-operator/oracle/pkg/agents/oracle"
"github.com/GoogleCloudPlatform/elcarro-oracle-operator/oracle/pkg/k8s"
)
var CheckStatusInstanceFunc = controllers.CheckStatusInstanceFunc
// InstanceReconciler reconciles an Instance object.
type InstanceReconciler struct {
client.Client
Log logr.Logger
SchemeVal *runtime.Scheme
Images map[string]string
Recorder record.EventRecorder
InstanceLocks *sync.Map
DatabaseClientFactory controllers.DatabaseClientFactory
}
func (r *InstanceReconciler) Scheme() *runtime.Scheme {
return r.SchemeVal
}
// +kubebuilder:rbac:groups=oracle.db.anthosapis.com,resources=instances,verbs=get;list;watch;create;update;patch;delete
// +kubebuilder:rbac:groups=oracle.db.anthosapis.com,resources=instances/status,verbs=get;update;patch
// +kubebuilder:rbac:groups=apps,resources=deployments,verbs=get;list;watch;create;update;patch;delete
// +kubebuilder:rbac:groups=apps,resources=statefulsets,verbs=get;list;watch;create;update;patch;delete
// +kubebuilder:rbac:groups=apps,resources=statefulsets/status,verbs=get;update;patch
// +kubebuilder:rbac:groups="",resources=persistentvolumeclaims,verbs=get;list;watch;create;update;patch;delete
// +kubebuilder:rbac:groups="",resources=persistentvolumes,verbs=get;list;watch;create;update;patch;delete
// +kubebuilder:rbac:groups=core,resources=services,verbs=list;watch;get;patch;create;delete
// +kubebuilder:rbac:groups="",resources=configmaps,verbs=get;list;watch;create;update;patch;delete
// +kubebuilder:rbac:groups="",resources=events,verbs=create;patch
// +kubebuilder:rbac:groups=storage.k8s.io,resources=storageclasses,verbs=get;list;watch
// +kubebuilder:rbac:groups=coordination.k8s.io,resources=leases,verbs=get;list;create;update
// +kubebuilder:rbac:groups=core,resources=secrets,verbs=get;list;watch;create;update;patch;delete
// +kubebuilder:rbac:groups=oracle.db.anthosapis.com,resources=databases,verbs=get;list;watch;create;update;patch;delete
// +kubebuilder:rbac:groups=oracle.db.anthosapis.com,resources=databases/status,verbs=get;update;patch
// +kubebuilder:rbac:groups=oracle.db.anthosapis.com,resources=configs,verbs=get;list;watch;create;update;patch;delete
const (
physicalRestore = "PhysicalRestore"
InstanceReadyTimeout = 120 * time.Minute
DatabaseInstanceReadyTimeoutSeeded = 30 * time.Minute
DatabaseInstanceReadyTimeoutUnseeded = 60 * time.Minute // 60 minutes because it can take 50+ minutes to create an unseeded CDB
dateFormat = "20060102"
DefaultStsPatchingTimeout = 25 * time.Minute
reconcileTimeout = 3 * time.Minute
)
func (r *InstanceReconciler) Reconcile(ctx context.Context, req ctrl.Request) (_ ctrl.Result, respErr error) {
ctx, cancel := context.WithTimeout(ctx, reconcileTimeout)
defer cancel()
log := r.Log.WithValues("Instance", req.NamespacedName)
log.Info("reconciling instance")
var inst v1alpha1.Instance
if err := r.Get(ctx, req.NamespacedName, &inst); err != nil {
return ctrl.Result{}, client.IgnoreNotFound(err)
}
if err := validateSpec(&inst); err != nil {
log.Error(err, "instance spec validation failed")
// TODO better error handling, no need retry
return ctrl.Result{}, nil
}
defer func() {
r.updateIsChangeApplied(&inst, log)
if err := r.Status().Update(ctx, &inst); err != nil {
log.Error(err, "failed to update the instance status")
if respErr == nil {
respErr = err
}
}
}()
instanceReadyCond := k8s.FindCondition(inst.Status.Conditions, k8s.Ready)
if IsDeleting(&inst) {
return r.reconcileInstanceDeletion(ctx, req, log)
} else if IsStopped(&inst) && !k8s.ConditionReasonEquals(instanceReadyCond, k8s.InstanceStopped) {
return r.reconcileInstanceStop(ctx, req, log)
} else if !IsStopped(&inst) && k8s.ConditionReasonEquals(instanceReadyCond, k8s.InstanceStopped) {
k8s.InstanceUpsertCondition(&inst.Status, k8s.Ready, v1.ConditionFalse, k8s.CreateInProgress, "Restarting Instance")
}
// Add finalizer to clean up underlying objects in case of deletion.
if !controllerutil.ContainsFinalizer(&inst, controllers.FinalizerName) {
log.Info("adding a finalizer to the Instance object.")
controllerutil.AddFinalizer(&inst, controllers.FinalizerName)
if err := r.Update(ctx, &inst); err != nil {
return ctrl.Result{}, err
}
}
diskSpace, err := commonutils.DiskSpaceTotal(&inst)
if err != nil {
log.Error(err, "failed to calculate the total disk space")
}
log.Info("common instance", "total allocated disk space across all instance disks [Gi]", diskSpace/1024/1024/1024)
instanceReadyCond = k8s.FindCondition(inst.Status.Conditions, k8s.Ready)
dbInstanceCond := k8s.FindCondition(inst.Status.Conditions, k8s.DatabaseInstanceReady)
if inst.Spec.Mode == commonv1alpha1.Pause {
if instanceReadyCond == nil || dbInstanceCond == nil || instanceReadyCond.Reason != k8s.CreateComplete {
log.Info("Ignoring pause mode since only instances in a stable state can be paused.")
} else {
r.InstanceLocks.Store(fmt.Sprintf("%s-%s", inst.Namespace, inst.Name), true)
k8s.InstanceUpsertCondition(&inst.Status, k8s.Ready, v1.ConditionFalse, k8s.PauseMode, "Instance switched to pause mode")
log.Info("Instance has been set to pause for further reconciliation reset the pause mode")
return ctrl.Result{}, nil
}
}
var enabledServices []commonv1alpha1.Service
for service, enabled := range inst.Spec.Services {
if enabled {
enabledServices = append(enabledServices, service)
}
}
// If the instance is ready and DR enabled, we can set up standby DR.
if k8s.ConditionReasonEquals(instanceReadyCond, k8s.StandbyDRInProgress) && isStandbyDR(&inst) {
return r.standbyStateMachine(ctx, &inst, log)
}
if result, err := r.parameterUpdateStateMachine(ctx, req, inst, log); err != nil {
return result, err
}
// If the instance and database is ready, we can set the instance parameters
if k8s.ConditionStatusEquals(instanceReadyCond, v1.ConditionTrue) &&
k8s.ConditionStatusEquals(dbInstanceCond, v1.ConditionTrue) && (inst.Spec.EnableDnfs != inst.Status.DnfsEnabled) {
log.Info("instance and db is ready, modifying dNFS")
if err := r.setDnfs(ctx, inst, inst.Spec.EnableDnfs); err != nil {
return ctrl.Result{}, err
}
inst.Status.DnfsEnabled = inst.Spec.EnableDnfs
if inst.Status.DnfsEnabled {
log.Info("dNFS successfully enabled")
} else {
log.Info("dNFS successfully disabled")
}
}
instanceReadyCond = k8s.FindCondition(inst.Status.Conditions, k8s.Ready)
dbInstanceCond = k8s.FindCondition(inst.Status.Conditions, k8s.DatabaseInstanceReady)
// Load default preferences (aka "config") if provided by a customer.
config, err := r.loadConfig(ctx, req.NamespacedName.Namespace)
if err != nil {
return ctrl.Result{}, err
}
images := CloneMap(r.Images)
if err := r.overrideDefaultImages(config, images, &inst, log); err != nil {
return ctrl.Result{}, err
}
applyOpts := []client.PatchOption{client.ForceOwnership, client.FieldOwner("instance-controller")}
cm, err := controllers.NewConfigMap(&inst, r.Scheme(), fmt.Sprintf(controllers.CmName, inst.Name))
if err != nil {
log.Error(err, "failed to create a ConfigMap", "cm", cm)
return ctrl.Result{}, err
}
if err := r.Patch(ctx, cm, client.Apply, applyOpts...); err != nil {
return ctrl.Result{}, err
}
// Create a StatefulSet if needed.
sp := controllers.StsParams{
Inst: &inst,
Scheme: r.Scheme(),
Namespace: req.NamespacedName.Namespace,
Images: images,
SvcName: fmt.Sprintf(controllers.SvcName, inst.Name),
StsName: fmt.Sprintf(controllers.StsName, inst.Name),
PrivEscalation: false,
ConfigMap: cm,
Disks: controllers.DiskSpecs(&inst, config),
Config: config,
Log: log,
Services: enabledServices,
}
if IsPatchingStateMachineEntryCondition(inst.Spec.Services, inst.Status.ActiveImages, sp.Images, inst.Status.LastFailedImages, instanceReadyCond, dbInstanceCond) ||
inst.Status.CurrentActiveStateMachine == "PatchingStateMachine" {
databasePatchingTimeout := DefaultStsPatchingTimeout
if inst.Spec.DatabasePatchingTimeout != nil {
databasePatchingTimeout = inst.Spec.DatabasePatchingTimeout.Duration
}
result, err, done := r.patchingStateMachine(req, instanceReadyCond, dbInstanceCond, &inst, ctx, &sp, config, databasePatchingTimeout, log)
if err != nil {
log.Error(err, "patchingStateMachine failed")
}
if done {
return result, err
}
}
// If there is a Restore section in the spec the reconciliation will be handled
// by restore state machine until the Spec.Restore section is removed again.
if inst.Spec.Restore != nil {
// Ask the restore state machine to reconcile
result, err := r.restoreStateMachine(req, instanceReadyCond, dbInstanceCond, &inst, ctx, sp, log)
if err != nil {
log.Error(err, "restoreStateMachine failed")
return result, err
}
if !result.IsZero() {
return result, err
}
// No error and no result - state machine is done, proceed with main reconciler
}
//if we return something we have to requeue
res, err := r.handleResize(ctx, &inst, instanceReadyCond, dbInstanceCond, sp, applyOpts, log)
if err != nil {
return ctrl.Result{}, err
} else if !res.IsZero() {
return res, nil
}
if k8s.ConditionStatusEquals(instanceReadyCond, v1.ConditionTrue) && k8s.ConditionStatusEquals(dbInstanceCond, v1.ConditionTrue) {
log.Info("instance has already been provisioned and ready")
if res, err := r.reconcileMonitoring(ctx, &inst, log, images); err != nil || res.RequeueAfter > 0 {
return res, err
}
return ctrl.Result{}, r.updateDatabaseIncarnationStatus(ctx, &inst, r.Log)
}
if result, err := r.createStatefulSet(ctx, &inst, sp, applyOpts, log); err != nil {
return result, err
}
dbLoadBalancer, err := r.createDBLoadBalancer(ctx, &inst, applyOpts)
if err != nil {
return ctrl.Result{}, err
}
_, _, err = r.createDataplaneServices(ctx, inst, applyOpts)
if err != nil {
return ctrl.Result{}, err
}
if instanceReadyCond == nil {
instanceReadyCond = k8s.InstanceUpsertCondition(&inst.Status, k8s.Ready, v1.ConditionFalse, k8s.CreateInProgress, "")
}
inst.Status.Endpoint = fmt.Sprintf(controllers.SvcEndpoint, fmt.Sprintf(controllers.SvcName, inst.Name), inst.Namespace)
inst.Status.URL = commonutils.LoadBalancerURL(dbLoadBalancer, consts.SecureListenerPort)
// RequeueAfter 30 seconds to avoid constantly reconcile errors before statefulSet is ready.
// Update status when the Service is ready (for the initial provisioning).
// Also confirm that the StatefulSet is up and running.
if k8s.ConditionReasonEquals(instanceReadyCond, k8s.CreateInProgress) {
elapsed := k8s.ElapsedTimeFromLastTransitionTime(instanceReadyCond, time.Second)
if elapsed > InstanceReadyTimeout {
r.Recorder.Eventf(&inst, corev1.EventTypeWarning, "InstanceReady", fmt.Sprintf("Instance provision timed out after %v", InstanceReadyTimeout))
msg := fmt.Sprintf("Instance provision timed out. Elapsed Time: %v", elapsed)
log.Info(msg)
k8s.InstanceUpsertCondition(&inst.Status, k8s.Ready, v1.ConditionFalse, k8s.CreateInProgress, msg)
return ctrl.Result{}, nil
}
if !r.updateProgressCondition(ctx, inst, req.NamespacedName.Namespace, controllers.CreateInProgress, log) {
log.Info("requeue after 30 seconds")
return ctrl.Result{RequeueAfter: 30 * time.Second}, nil
}
if inst.Status.URL != "" {
if !k8s.ConditionReasonEquals(instanceReadyCond, k8s.CreateComplete) {
r.Recorder.Eventf(&inst, corev1.EventTypeNormal, "InstanceReady", "Instance has been created successfully. Elapsed Time: %v", elapsed)
k8s.InstanceUpsertCondition(&inst.Status, k8s.Ready, v1.ConditionTrue, k8s.CreateComplete, "")
inst.Status.ActiveImages = CloneMap(sp.Images)
return ctrl.Result{Requeue: true}, nil
}
}
}
if inst.Labels == nil {
inst.Labels = map[string]string{"instance": inst.Name}
if err := r.Update(ctx, &inst); err != nil {
log.Error(err, "failed to update the Instance spec (set labels)")
return ctrl.Result{}, err
}
}
if isStandbyDR(&inst) {
k8s.InstanceUpsertCondition(&inst.Status, k8s.Ready, v1.ConditionFalse, k8s.StandbyDRInProgress, "standby DR in progress")
return ctrl.Result{Requeue: true}, nil
}
// When we reach here, the instance should be ready.
if inst.Spec.Mode == commonv1alpha1.ManuallySetUpStandby {
log.Info("reconciling instance for manually set up standby: DONE")
// the code will return here, so we can rely on defer function to update database status.
k8s.InstanceUpsertCondition(&inst.Status, k8s.Ready, v1.ConditionFalse, k8s.ManuallySetUpStandbyInProgress, fmt.Sprintf("Setting up standby database in progress, remove spec.mode %v to promote the instance", inst.Spec.Mode))
k8s.InstanceUpsertCondition(&inst.Status, k8s.StandbyReady, v1.ConditionTrue, k8s.CreateComplete, fmt.Sprintf("standby instance creation complete, ready to set up standby database in the instance"))
return ctrl.Result{}, nil
}
if k8s.ConditionStatusEquals(k8s.FindCondition(inst.Status.Conditions, k8s.StandbyReady), v1.ConditionTrue) {
// promote the standby instance, bootstrap is part of promotion.
r.Recorder.Eventf(&inst, corev1.EventTypeNormal, k8s.PromoteStandbyInProgress, "")
if err := r.bootstrapStandby(ctx, &inst); err != nil {
r.Recorder.Eventf(&inst, corev1.EventTypeWarning, k8s.PromoteStandbyFailed, fmt.Sprintf("Error promoting standby: %v", err))
return ctrl.Result{}, err
}
// the standby instance has been successfully promoted, set ready condition
// to true and standby ready to false. Promotion need to be idempotent to
// ensure the correctness under retry.
r.Recorder.Eventf(&inst, corev1.EventTypeNormal, k8s.PromoteStandbyComplete, "")
k8s.InstanceUpsertCondition(&inst.Status, k8s.Ready, v1.ConditionTrue, k8s.CreateComplete, "")
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionTrue, k8s.CreateComplete, "")
k8s.InstanceUpsertCondition(&inst.Status, k8s.StandbyReady, v1.ConditionFalse, k8s.PromoteStandbyComplete, "")
return ctrl.Result{Requeue: true}, err
}
var dbs v1alpha1.DatabaseList
if err := r.List(ctx, &dbs, client.InNamespace(req.Namespace)); err != nil {
log.V(1).Info("failed to list databases for instance", "inst.Name", inst.Name)
} else {
log.Info("list of queried databases", "dbs", dbs)
}
log.Info("instance status", "instanceReadyCond", instanceReadyCond, "endpoint", inst.Status.Endpoint,
"url", inst.Status.URL, "databases", inst.Status.DatabaseNames)
log.Info("reconciling instance: DONE")
result, err := r.reconcileDatabaseInstance(ctx, &inst, r.Log, images)
log.Info("reconciling database instance: DONE", "result", result, "err", err)
if err != nil {
return result, err
}
return result, nil
}
// Create a name for the createCDB LRO operation based on instance GUID.
func lroCreateCDBOperationID(instance v1alpha1.Instance) string {
return fmt.Sprintf("CreateCDB_%s", instance.GetUID())
}
// Create a name for the bootstrapCDB LRO operation based on instance GUID.
func lroBootstrapCDBOperationID(instance v1alpha1.Instance) string {
return fmt.Sprintf("BootstrapCDB_%s", instance.GetUID())
}
func (r *InstanceReconciler) reconcileInstanceDeletion(ctx context.Context, req ctrl.Request, log logr.Logger) (ctrl.Result, error) {
log.Info("Deleting Instance...", "InstanceName", req.NamespacedName.Name)
// NOTE: must be kept in sync with reconcileMonitoring
var monitor appsv1.Deployment
if err := r.Get(ctx, client.ObjectKey{Namespace: req.Namespace, Name: fmt.Sprintf("%s-monitor", req.Name)}, &monitor); err == nil {
if err := r.Delete(ctx, &monitor); err != nil {
log.Error(err, "failed to delete monitoring deployment", "InstanceName", req.Name, "MonitorDeployment", monitor.Name)
return ctrl.Result{}, err
}
return ctrl.Result{Requeue: true}, nil
} else if !apierrors.IsNotFound(err) { // retry on other errors.
return ctrl.Result{}, err
}
var inst v1alpha1.Instance
if err := r.Get(ctx, req.NamespacedName, &inst); err != nil {
return ctrl.Result{}, client.IgnoreNotFound(err)
}
if len(inst.Status.DatabaseNames) == 0 {
controllerutil.RemoveFinalizer(&inst, controllers.FinalizerName)
if err := r.Update(ctx, &inst); err != nil {
log.Error(err, "failed to remove a finalizer from an Instance", "InstanceName", inst.Name, "FinalizerName", controllers.FinalizerName)
return ctrl.Result{}, err
}
return ctrl.Result{}, nil
}
for _, dbName := range inst.Status.DatabaseNames {
var db v1alpha1.Database
if err := r.Get(ctx, types.NamespacedName{Namespace: req.Namespace, Name: dbName}, &db); err != nil {
return ctrl.Result{}, client.IgnoreNotFound(err)
}
log.Info("instance deletion in progress. deleting an attached Database(PDB) first", "InstanceName", inst.Name, "DatabaseName", db.Name)
if err := r.Delete(ctx, &db); err != nil {
log.Error(err, "failed to delete database(PDB) attached to Instance", "InstanceName", inst.Name, "DatabaseName", db.Name)
return ctrl.Result{}, err
}
}
return ctrl.Result{RequeueAfter: 10 * time.Second}, nil
}
func (r *InstanceReconciler) reconcileInstanceStop(ctx context.Context, req ctrl.Request, log logr.Logger) (ctrl.Result, error) {
log.Info("Stopping Instance...", "InstanceName", req.NamespacedName.Name)
var inst v1alpha1.Instance
if err := r.Get(ctx, req.NamespacedName, &inst); err != nil {
return ctrl.Result{}, client.IgnoreNotFound(err)
}
r.recordEventAndUpdateStatus(ctx, &inst, v1.ConditionFalse, k8s.InstanceStoppingInProgress, "Instance is being stopped", log)
if _, err := r.stopDBStatefulset(ctx, req, log); err != nil {
return ctrl.Result{}, err
}
//delete the monitoring pod
var monitor appsv1.Deployment
if err := r.Get(ctx, client.ObjectKey{Namespace: req.Namespace, Name: fmt.Sprintf("%s-monitor", req.Name)}, &monitor); err == nil {
if err := r.Delete(ctx, &monitor); err != nil {
log.Error(err, "failed to delete monitoring deployment", "InstanceName", req.Name, "MonitorDeployment", monitor.Name)
return ctrl.Result{}, err
}
} else if !apierrors.IsNotFound(err) { // retry on other errors.
return ctrl.Result{}, err
}
if _, err := r.deleteDBLoadBalancer(ctx, &inst, log); err != nil {
return ctrl.Result{}, err
}
if _, err := r.deleteDBDSVC(ctx, &inst, log); err != nil {
return ctrl.Result{}, err
}
if _, err := r.deleteAgentSVC(ctx, &inst, log); err != nil {
return ctrl.Result{}, err
}
if err := r.Get(ctx, req.NamespacedName, &inst); err != nil {
return ctrl.Result{}, client.IgnoreNotFound(err)
}
r.recordEventAndUpdateStatus(ctx, &inst, v1.ConditionTrue, k8s.InstanceStopped, "Instance has been stopped", log)
return ctrl.Result{Requeue: false}, nil
}
func (r *InstanceReconciler) shutdownDB(ctx context.Context, inst v1alpha1.Instance) error {
dbClient, closeConn, err := r.DatabaseClientFactory.New(ctx, r, inst.GetNamespace(), inst.Name)
if err != nil {
return err
}
defer closeConn()
_, err = dbClient.BounceDatabase(ctx, &dbdpb.BounceDatabaseRequest{
Operation: dbdpb.BounceDatabaseRequest_SHUTDOWN,
DatabaseName: inst.Spec.CDBName,
Option: "immediate",
})
if err != nil {
return fmt.Errorf("BounceDatabase: error while shutting db: %v", err)
}
return nil
}
// reconcileDatabaseInstance reconciling the underlying database instance.
// Successful state transition for seeded instance:
// nil->BootstrapPending->BootstrapInProgress->CreateFailed/CreateComplete
// Successful state transition for unseeded instance:
// nil->CreatePending->CreateInProgress->BootstrapPending->BootstrapInProgress->ReconcileServices->CreateFailed/CreateComplete
// Successful state transition for instance with restoreSpec:
// nil->RestorePending->CreateComplete
func (r *InstanceReconciler) reconcileDatabaseInstance(ctx context.Context, inst *v1alpha1.Instance, log logr.Logger, images map[string]string) (ctrl.Result, error) {
instanceReadyCond := k8s.FindCondition(inst.Status.Conditions, k8s.Ready)
dbInstanceCond := k8s.FindCondition(inst.Status.Conditions, k8s.DatabaseInstanceReady)
log.Info("reconciling database instance: ", "instanceReadyCond", instanceReadyCond, "dbInstanceCond", dbInstanceCond)
// reconcile database only when instance is ready, but requeue.
if !k8s.ConditionStatusEquals(instanceReadyCond, v1.ConditionTrue) {
return ctrl.Result{Requeue: true}, nil
}
isImageSeeded, err := r.isImageSeeded(ctx, inst, log)
if err != nil {
log.Error(err, "unable to determine image type")
return ctrl.Result{}, err
}
if dbInstanceCond == nil {
if inst.Spec.Restore != nil {
log.Info("Skip bootstrap CDB database, waiting to be restored")
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionFalse, k8s.RestorePending, "Awaiting restore CDB")
} else if !isImageSeeded {
log.Info("Unseeded image used, waiting to be created")
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionFalse, k8s.CreatePending, "Awaiting create CDB")
} else {
log.Info("Seeded image used, waiting to be bootstrapped")
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionFalse, k8s.BootstrapPending, "Awaiting bootstrap CDB")
}
return ctrl.Result{Requeue: true}, r.Status().Update(ctx, inst)
}
// Check for timeout
if !k8s.ConditionStatusEquals(dbInstanceCond, v1.ConditionTrue) {
elapsed := k8s.ElapsedTimeFromLastTransitionTime(dbInstanceCond, time.Second)
createDatabaseInstanceTimeout := DatabaseInstanceReadyTimeoutSeeded
if !isImageSeeded {
createDatabaseInstanceTimeout = DatabaseInstanceReadyTimeoutUnseeded
}
if elapsed < createDatabaseInstanceTimeout {
log.Info(fmt.Sprintf("database instance creation in progress for %v", elapsed))
} else {
log.Info(fmt.Sprintf("database instance creation timed out. Elapsed Time: %v", elapsed))
if !strings.Contains(dbInstanceCond.Message, "Warning") { // so that we would create only one database instance timeout event
r.Recorder.Eventf(inst, corev1.EventTypeWarning, k8s.DatabaseInstanceTimeout, "DatabaseInstance has been in progress for over %v, please try delete and recreate.", createDatabaseInstanceTimeout)
}
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionFalse, dbInstanceCond.Reason, "Warning: db instance is taking too long to start up - please try delete and recreate")
return ctrl.Result{}, nil
}
}
switch dbInstanceCond.Reason {
case k8s.CreatePending:
// Launch the CreateCDB LRO
req := &controllers.CreateCDBRequest{
Sid: inst.Spec.CDBName,
DbUniqueName: inst.Spec.DBUniqueName,
DbDomain: controllers.GetDBDomain(inst),
CharacterSet: inst.Spec.CharacterSet,
MemoryPercent: int32(inst.Spec.MemoryPercent),
//DBCA expects the parameters in the following string array format
// ["key1=val1", "key2=val2","key3=val3"]
AdditionalParams: mapsToStringArray(inst.Spec.Parameters),
LroInput: &controllers.LROInput{OperationId: lroCreateCDBOperationID(*inst)},
}
_, err = controllers.CreateCDB(ctx, r, r.DatabaseClientFactory, inst.GetNamespace(), inst.GetName(), *req)
if err != nil {
if !controllers.IsAlreadyExistsError(err) {
log.Error(err, "CreateCDB failed")
return ctrl.Result{}, err
}
} else {
log.Info("CreateCDB started")
}
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionFalse, k8s.CreateInProgress, "Database creation in progress")
return ctrl.Result{Requeue: true}, r.Status().Update(ctx, inst)
case k8s.CreateInProgress:
id := lroCreateCDBOperationID(*inst)
done, err := controllers.IsLROOperationDone(ctx, r.DatabaseClientFactory, r.Client, id, inst.GetNamespace(), inst.GetName())
if !done {
log.Info("CreateCDB still in progress, waiting")
return ctrl.Result{RequeueAfter: 15 * time.Second}, nil
}
controllers.DeleteLROOperation(ctx, r.DatabaseClientFactory, r.Client, id, inst.Namespace, inst.Name)
if err != nil {
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionFalse, k8s.CreateFailed, "CreateCDB LRO returned error")
log.Error(err, "CreateCDB LRO returned error")
return ctrl.Result{Requeue: true}, r.Status().Update(ctx, inst)
}
log.Info("CreateCDB done successfully")
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionFalse, k8s.BootstrapPending, "")
// Reconcile again
return ctrl.Result{Requeue: true}, r.Status().Update(ctx, inst)
case k8s.BootstrapPending:
// Launch the BootstrapDatabase LRO
bootstrapMode := controllers.BootstrapDatabaseRequest_ProvisionUnseeded
if isImageSeeded {
bootstrapMode = controllers.BootstrapDatabaseRequest_ProvisionSeeded
}
req := &controllers.BootstrapDatabaseRequest{
CdbName: inst.Spec.CDBName,
DbUniqueName: inst.Spec.DBUniqueName,
Dbdomain: controllers.GetDBDomain(inst),
Mode: bootstrapMode,
LroInput: &controllers.LROInput{OperationId: lroBootstrapCDBOperationID(*inst)},
}
lro, err := controllers.BootstrapDatabase(ctx, r, r.DatabaseClientFactory, inst.Namespace, inst.Name, *req)
if err != nil {
if !controllers.IsAlreadyExistsError(err) {
log.Error(err, "BootstrapDatabase failed")
return ctrl.Result{}, err
}
} else if lro.GetDone() {
// handle synchronous version of BootstrapDatabase
r.Log.Info("encountered synchronous version of BootstrapDatabase")
r.Log.Info("BootstrapDatabase DONE")
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionFalse, k8s.ReconcileServices, "Services starting")
return ctrl.Result{Requeue: true}, r.Status().Update(ctx, inst)
}
log.Info("BootstrapDatabase started")
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionFalse, k8s.BootstrapInProgress, "Database bootstrap in progress")
return ctrl.Result{Requeue: true}, r.Status().Update(ctx, inst)
case k8s.BootstrapInProgress:
id := lroBootstrapCDBOperationID(*inst)
done, err := controllers.IsLROOperationDone(ctx, r.DatabaseClientFactory, r.Client, id, inst.GetNamespace(), inst.GetName())
if !done {
log.Info("BootstrapDatabase still in progress, waiting")
return ctrl.Result{RequeueAfter: 15 * time.Second}, nil
}
controllers.DeleteLROOperation(ctx, r.DatabaseClientFactory, r.Client, id, inst.Namespace, inst.Name)
if err != nil {
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionFalse, k8s.CreateFailed, "BootstrapDatabase LRO returned error")
log.Error(err, "BootstrapDatabase LRO returned error")
return ctrl.Result{Requeue: true}, r.Status().Update(ctx, inst)
}
log.Info("BootstrapDatabase done successfully")
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionFalse, k8s.ReconcileServices, "Services starting")
// Reconcile again
return ctrl.Result{Requeue: true}, r.Status().Update(ctx, inst)
case k8s.ReconcileServices:
res, err := r.reconcileMonitoring(ctx, inst, log, images)
if err == nil && res.RequeueAfter == 0 {
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionTrue, k8s.CreateComplete, "")
return ctrl.Result{Requeue: true}, r.Status().Update(ctx, inst)
}
return res, err
case k8s.RestorePending:
if k8s.ConditionReasonEquals(instanceReadyCond, k8s.RestoreComplete) {
k8s.InstanceUpsertCondition(&inst.Status, k8s.DatabaseInstanceReady, v1.ConditionTrue, k8s.CreateComplete, "")
return ctrl.Result{Requeue: true}, r.Status().Update(ctx, inst)
}
default:
r.Log.Info("reconcileDatabaseInstance: no action needed, proceed with main reconciliation")
}
return ctrl.Result{}, nil
}
// setDnfs enables dNFS protocol in Oracle database.
func (r *InstanceReconciler) setDnfs(ctx context.Context, inst v1alpha1.Instance, enable bool) error {
dbClient, closeConn, err := r.DatabaseClientFactory.New(ctx, r, inst.GetNamespace(), inst.Name)
if err != nil {
return err
}
defer closeConn()
if _, err := dbClient.SetDnfsState(ctx, &dbdpb.SetDnfsStateRequest{
Enable: enable,
}); err != nil {
return fmt.Errorf("error while enabling dNFS: %v", err)
}
_, err = dbClient.BounceDatabase(ctx, &dbdpb.BounceDatabaseRequest{
Operation: dbdpb.BounceDatabaseRequest_SHUTDOWN,
DatabaseName: inst.Spec.CDBName,
Option: "immediate",
})
if err != nil {
return fmt.Errorf("BounceDatabase: error while shutting db: %v", err)
}
_, err = dbClient.BounceDatabase(ctx, &dbdpb.BounceDatabaseRequest{
Operation: dbdpb.BounceDatabaseRequest_STARTUP,
DatabaseName: inst.Spec.CDBName,
AvoidConfigBackup: false,
})
if err != nil {
return fmt.Errorf("dbClient/BounceDatabase: error while starting db: %v", err)
}
return nil
}
func (r *InstanceReconciler) SetupWithManager(mgr ctrl.Manager) error {
r.Log.V(1).Info("SetupWithManager", "images", r.Images)
// configPredicate is used to determine if watched events should cause
// all instances in the namespace to be reconciled. Right now only
// Images in the config affect running instances and can cause Patching
// so we only trigger instance reconciliation on Create/Delete and
// Update when Images changes.
configPredicate := predicate.Funcs{
UpdateFunc: func(e event.UpdateEvent) bool {
oldConfig, ok := e.ObjectOld.(*v1alpha1.Config)
if !ok {
return false
}
newConfig, ok := e.ObjectNew.(*v1alpha1.Config)
if !ok {
return false
}
return !cmp.Equal(oldConfig.Spec.Images, newConfig.Spec.Images)
},
CreateFunc: func(e event.CreateEvent) bool { return true },
DeleteFunc: func(e event.DeleteEvent) bool { return true },
GenericFunc: func(e event.GenericEvent) bool { return false },
}
instancesForConfig := func(obj client.Object) []ctrl.Request {
var requests []ctrl.Request
var insts v1alpha1.InstanceList
if err := r.List(context.Background(), &insts, client.InNamespace(obj.GetNamespace())); err != nil {
return nil
}
for _, inst := range insts.Items {
requests = append(requests, reconcile.Request{
NamespacedName: types.NamespacedName{
Name: inst.Name,
Namespace: inst.Namespace,
}})
}
r.Log.Info("Config event triggered instance reconcile ", "requests", requests)
return requests
}
return ctrl.NewControllerManagedBy(mgr).
For(&v1alpha1.Instance{}).
Owns(&corev1.Service{}).
Owns(&appsv1.StatefulSet{}).
Owns(&corev1.PersistentVolumeClaim{}).
Owns(&corev1.ConfigMap{}).
Watches(
&source.Kind{Type: &v1alpha1.Config{}},
handler.EnqueueRequestsFromMapFunc(instancesForConfig),
builder.WithPredicates(configPredicate),
).
WithOptions(controller.Options{MaxConcurrentReconciles: 4}).
Complete(r)
}
func lroOperationID(opType string, instance *v1alpha1.Instance) string {
switch opType {
case physicalRestore:
return fmt.Sprintf("%s_%s_%s", opType, instance.GetUID(), instance.Status.LastRestoreTime.Format(time.RFC3339))
default:
return fmt.Sprintf("%s_%s", opType, instance.GetUID())
}
}
func IsStopped(instance *v1alpha1.Instance) bool {
return instance.InstanceSpec().IsStopped != nil && *instance.InstanceSpec().IsStopped == true
}
func IsDeleting(instance *v1alpha1.Instance) bool {
return !instance.GetDeletionTimestamp().IsZero()
}