Repository navigation
Expand file tree
/
Copy pathopenstackdataplanedeployment_controller.go
More file actions
893 lines (800 loc) · 31.9 KB
/
Copy pathopenstackdataplanedeployment_controller.go
File metadata and controls
893 lines (800 loc) · 31.9 KB
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
/*
Copyright 2023.
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 dataplane contains the OpenStackDataPlaneDeployment controller implementation
package dataplane
import (
"context"
"errors"
"fmt"
"math"
"slices"
"time"
"github.com/go-playground/validator/v10"
"github.com/iancoleman/strcase"
batchv1 "k8s.io/api/batch/v1"
corev1 "k8s.io/api/core/v1"
k8s_errors "k8s.io/apimachinery/pkg/api/errors"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/runtime"
"k8s.io/apimachinery/pkg/types"
"k8s.io/client-go/kubernetes"
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/handler"
"sigs.k8s.io/controller-runtime/pkg/log"
"sigs.k8s.io/controller-runtime/pkg/predicate"
"sigs.k8s.io/controller-runtime/pkg/reconcile"
certmgrv1 "github.com/cert-manager/cert-manager/pkg/apis/certmanager/v1"
"github.com/go-logr/logr"
condition "github.com/openstack-k8s-operators/lib-common/modules/common/condition"
"github.com/openstack-k8s-operators/lib-common/modules/common/helper"
"github.com/openstack-k8s-operators/lib-common/modules/common/util"
dataplanev1 "github.com/openstack-k8s-operators/openstack-operator/api/dataplane/v1beta1"
deployment "github.com/openstack-k8s-operators/openstack-operator/internal/dataplane"
dataplaneutil "github.com/openstack-k8s-operators/openstack-operator/internal/dataplane/util"
)
// OpenStackDataPlaneDeploymentReconciler reconciles a OpenStackDataPlaneDeployment object
type OpenStackDataPlaneDeploymentReconciler struct {
client.Client
Kclient kubernetes.Interface
Scheme *runtime.Scheme
}
// GetLogger returns a logger object with a prefix of "controller.name" and additional controller context fields
func (r *OpenStackDataPlaneDeploymentReconciler) GetLogger(ctx context.Context) logr.Logger {
return log.FromContext(ctx).WithName("Controllers").WithName("OpenStackDataPlaneDeployment")
}
// +kubebuilder:rbac:groups=dataplane.openstack.org,resources=openstackdataplanedeployments,verbs=get;list;watch;create;delete
// +kubebuilder:rbac:groups=dataplane.openstack.org,resources=openstackdataplanedeployments/status,verbs=get;update;patch
// +kubebuilder:rbac:groups=dataplane.openstack.org,resources=openstackdataplanedeployments/finalizers,verbs=update;patch
// +kubebuilder:rbac:groups=dataplane.openstack.org,resources=openstackdataplanenodesets,verbs=get;list;watch
// +kubebuilder:rbac:groups=dataplane.openstack.org,resources=openstackdataplaneservices,verbs=get;list;watch
// +kubebuilder:rbac:groups="",resources=pods,verbs=list;watch
// +kubebuilder:rbac:groups=discovery.k8s.io,resources=endpointslices,verbs=get;list;watch;create;update;patch;delete;
// +kubebuilder:rbac:groups=cert-manager.io,resources=issuers,verbs=get;list;watch;
// +kubebuilder:rbac:groups=cert-manager.io,resources=certificates,verbs=get;list;watch;create;update;patch;delete;
// +kubebuilder:rbac:groups=core.openstack.org,resources=openstackcontrolplanes,verbs=get;list;watch;
// Reconcile is part of the main kubernetes reconciliation loop which aims to
// move the current state of the cluster closer to the desired state.
func (r *OpenStackDataPlaneDeploymentReconciler) Reconcile(ctx context.Context, req ctrl.Request) (result ctrl.Result, _err error) {
Log := r.GetLogger(ctx)
Log.Info("Reconciling Deployment")
// Check if deployment name matches RFC1123 for use in labels
validate := validator.New()
if err := validate.Var(req.Name, "hostname_rfc1123"); err != nil {
Log.Error(err, "error validating OpenStackDataPlaneDeployment name, the name must follow RFC1123")
return ctrl.Result{}, err
}
// Fetch the OpenStackDataPlaneDeployment instance
instance := &dataplanev1.OpenStackDataPlaneDeployment{}
err := r.Get(ctx, req.NamespacedName, instance)
if err != nil {
if k8s_errors.IsNotFound(err) {
// Request object not found, could have been deleted after reconcile request.
// Owned objects are automatically garbage collected.
// For additional cleanup logic use finalizers. Return and don't requeue.
return ctrl.Result{}, nil
}
// Error reading the object - requeue the request.
return ctrl.Result{}, err
}
helper, _ := helper.NewHelper(
instance,
r.Client,
r.Kclient,
r.Scheme,
Log,
)
// Whether Job cleanup may be scheduled: a verdict freshly computed by this
// reconcile is only trustworthy once the status patch below persists it.
finishedOnEntry := instance.IsFinished()
statusPersisted := false
// Registered ahead of the status patch defer so it runs after the verdict persists.
defer func() {
if err := r.reconcileJobTTL(ctx, helper, instance, finishedOnEntry || statusPersisted); err != nil {
Log.Error(err, "Failed to reconcile AnsibleEE job cleanup", "deployment", instance.Name)
if _err == nil {
_err = err
}
}
}()
// If the deploy is already done, return immediately.
if instance.Status.Deployed {
Log.Info("Already deployed", "instance.Status.Deployed", instance.Status.Deployed)
return ctrl.Result{}, nil
}
// If the deployment is in a terminal state, do not reconcile even if
// nodesets or other watched resources change. The deployment should not
// be retried.
if terminalFailure(instance) {
Log.Info("Deployment is terminal, skipping reconciliation",
"deployment", instance.Name)
return ctrl.Result{}, nil
}
// initialize status if Conditions is nil, but do not reset if it already
// exists
isNewInstance := instance.Status.Conditions == nil
if isNewInstance {
instance.Status.Conditions = condition.Conditions{}
}
// Save a copy of the conditions so that we can restore the LastTransitionTime
// when a condition's state doesn't change.
savedConditions := instance.Status.Conditions.DeepCopy()
// Reset all conditions to Unknown as the state is not yet known for
// this reconcile loop.
instance.InitConditions()
instance.InitHashesAndImages()
// Set ObservedGeneration since we've reset conditions
instance.Status.ObservedGeneration = instance.Generation
// Always patch the instance status when exiting this function so we can persist any changes.
defer func() { // update the Ready condition based on the sub conditions
// Don't update the status, if reconciler Panics
if r := recover(); r != nil {
Log.Info("panic during reconcile", "panic", r)
panic(r)
}
if instance.Status.Conditions.AllSubConditionIsTrue() {
instance.Status.Conditions.MarkTrue(
condition.ReadyCondition, condition.ReadyMessage)
} else if instance.Status.Conditions.IsUnknown(condition.ReadyCondition) {
// Recalculate ReadyCondition based on the state of the rest of the conditions
instance.Status.Conditions.Set(
instance.Status.Conditions.Mirror(condition.ReadyCondition))
}
condition.RestoreLastTransitionTimes(
&instance.Status.Conditions, savedConditions)
err := helper.PatchInstance(ctx, instance)
if err != nil {
Log.Error(err, "Error updating instance status conditions")
_err = err
return
}
statusPersisted = true
}()
// Ensure NodeSets
nodeSets, err := r.listNodeSets(ctx, instance)
if err != nil {
if k8s_errors.IsNotFound(err) {
return ctrl.Result{RequeueAfter: time.Second * time.Duration(instance.Spec.DeploymentRequeueTime)}, nil
}
return ctrl.Result{}, err
}
globalInventorySecrets := map[string]string{}
globalSSHKeySecrets := map[string]string{}
// Check that all NodeSets are ready and get TLS certs
for _, nodeSet := range nodeSets.Items {
// Add inventory secret to list of inventories for global services
globalInventorySecrets[nodeSet.Name] = fmt.Sprintf("dataplanenodeset-%s", nodeSet.Name)
globalSSHKeySecrets[nodeSet.Name] = nodeSet.Spec.NodeTemplate.AnsibleSSHPrivateKeySecret
if !nodeSet.Status.Conditions.IsTrue(dataplanev1.SetupReadyCondition) {
Log.Info("NodeSet SetupReadyCondition is not True", "NodeSet", nodeSet.Name)
return ctrl.Result{RequeueAfter: time.Second * time.Duration(instance.Spec.DeploymentRequeueTime)}, nil
}
if nodeSet.Spec.TLSEnabled {
var services []string
if len(instance.Spec.ServicesOverride) != 0 {
services = instance.Spec.ServicesOverride
} else {
services = nodeSet.Spec.Services
}
nsConditions := instance.Status.NodeSetConditions[nodeSet.Name]
for _, serviceName := range services {
service, err := r.GetService(ctx, helper, instance, nsConditions, serviceName)
if err != nil {
return ctrl.Result{}, err
}
// If there is a different EDPMServiceType set and TLSCerts is
// not also set, get the service referenced by EDPMServiceType
// instead and use its cert data.
if serviceName != service.Spec.EDPMServiceType && service.Spec.TLSCerts == nil {
service, err = r.GetService(ctx, helper, instance, nsConditions, service.Spec.EDPMServiceType)
if err != nil {
return ctrl.Result{}, err
}
}
if service.Spec.TLSCerts != nil {
for certKey := range service.Spec.TLSCerts {
result, err := deployment.EnsureTLSCerts(ctx, helper, &nodeSet,
nodeSet.Status.AllHostnames, nodeSet.Status.AllIPs, service, certKey)
if err != nil {
instance.Status.Conditions.MarkFalse(
condition.InputReadyCondition,
condition.ErrorReason,
condition.SeverityError,
condition.TLSInputErrorMessage,
err.Error())
nsConditions.MarkFalse(
dataplanev1.NodeSetDeploymentReadyCondition,
condition.ErrorReason,
condition.SeverityError,
condition.TLSInputErrorMessage,
err.Error())
return ctrl.Result{}, err
} else if (*result != ctrl.Result{}) {
return *result, nil // requeue here
}
}
}
}
}
}
// All nodeSets successfully fetched.
// Mark InputReadyCondition=True
instance.Status.Conditions.MarkTrue(condition.InputReadyCondition, condition.InputReadyMessage)
shouldRequeue := false
haveError := false
deploymentErrMsg := ""
var servicePlan *deployment.ServicePlan
var finishedNodeSets []dataplanev1.OpenStackDataPlaneNodeSet
if instance.Status.FailedNodeSets == nil {
instance.Status.FailedNodeSets = map[string]string{}
}
if servicePlan, err = deployment.DedupeServices(ctx, helper, nodeSets.Items,
instance.Spec.ServicesOverride); err != nil {
util.LogErrorForObject(helper, err, "OpenStackDeployment error for deployment", instance)
var missingServiceErr *deployment.MissingServiceError
if errors.As(err, &missingServiceErr) {
nsConditions := instance.Status.NodeSetConditions[missingServiceErr.NodeSet]
nodeSetServiceErrorMessage := fmt.Sprintf(
dataplanev1.NodeSetServiceDeploymentErrorMessage,
missingServiceErr.Service) + " error %s"
nsConditions.MarkFalse(
dataplanev1.NodeSetDeploymentReadyCondition,
condition.ErrorReason,
condition.SeverityError,
nodeSetServiceErrorMessage,
err.Error())
serviceCondition := condition.Type(fmt.Sprintf(
"Service%sDeploymentReady",
strcase.ToCamel(missingServiceErr.Service)))
nsConditions.MarkFalse(
serviceCondition,
condition.ErrorReason,
condition.SeverityError,
nodeSetServiceErrorMessage,
err.Error())
instance.Status.NodeSetConditions[missingServiceErr.NodeSet] = nsConditions
instance.Status.Conditions.MarkFalse(
condition.DeploymentReadyCondition,
condition.ErrorReason,
condition.SeverityWarning,
condition.DeploymentReadyErrorMessage,
fmt.Sprintf("nodeSet: %s error: %s", missingServiceErr.NodeSet, err.Error()))
return ctrl.Result{}, err
}
instance.Status.Conditions.MarkFalse(
condition.DeploymentReadyCondition,
condition.ErrorReason,
condition.SeverityError,
dataplanev1.ServiceErrorMessage,
err.Error())
return ctrl.Result{}, err
}
version, err := dataplaneutil.GetVersion(ctx, helper, instance.Namespace)
if err != nil {
return ctrl.Result{}, err
}
// Built up front: the deployment starts its deployment scoped services itself,
// independent of which nodeset reconciles first.
deployers := make(map[string]*deployment.Deployer, len(nodeSets.Items))
serviceLevels := make(map[string][][]string, len(nodeSets.Items))
for i := range nodeSets.Items {
nodeSet := &nodeSets.Items[i]
ansibleEESpec := nodeSet.GetAnsibleEESpec()
ansibleEESpec.AnsibleTags = instance.Spec.AnsibleTags
ansibleEESpec.AnsibleSkipTags = instance.Spec.AnsibleSkipTags
ansibleEESpec.AnsibleLimit = instance.Spec.AnsibleLimit
ansibleEESpec.ExtraVars = instance.Spec.AnsibleExtraVars
ansibleEESpec.AnsibleEEEnvConfigMapName = instance.Spec.AnsibleEEEnvConfigMapName
if nodeSet.Status.DNSClusterAddresses != nil && nodeSet.Status.CtlplaneSearchDomain != "" {
ansibleEESpec.DNSConfig = &corev1.PodDNSConfig{
Nameservers: nodeSet.Status.DNSClusterAddresses,
Searches: []string{nodeSet.Status.CtlplaneSearchDomain},
}
}
deployers[nodeSet.Name] = &deployment.Deployer{
Ctx: ctx,
Helper: helper,
NodeSet: nodeSet,
Deployment: instance,
Status: &instance.Status,
AeeSpec: &ansibleEESpec,
InventorySecrets: globalInventorySecrets,
AnsibleSSHPrivateKeySecrets: globalSSHKeySecrets,
ServiceCache: servicePlan.Cache,
Version: version,
}
if instance.Spec.UseParallelExecution {
serviceLevels[nodeSet.Name], err = deployment.BuildServiceLevels(
ctx, helper, servicePlan.Cache, servicePlan.Services[nodeSet.Name])
if err != nil {
util.LogErrorForObject(helper, err, fmt.Sprintf("error building service dependency graph for NodeSet %s", nodeSet.Name), instance)
graphErrMsg := fmt.Sprintf("nodeSet: %s error: %s", nodeSet.Name, err.Error())
nsConditions := instance.Status.NodeSetConditions[nodeSet.Name]
nsConditions.MarkFalse(
dataplanev1.NodeSetDeploymentReadyCondition,
condition.ErrorReason,
condition.SeverityError,
dataplanev1.DataPlaneNodeSetErrorMessage,
err.Error())
instance.Status.NodeSetConditions[nodeSet.Name] = nsConditions
instance.Status.Conditions.MarkFalse(
condition.DeploymentReadyCondition,
condition.ErrorReason,
condition.SeverityError,
condition.DeploymentReadyErrorMessage,
graphErrMsg)
return ctrl.Result{}, fmt.Errorf("error building service dependency graph for nodeset %s: %w", nodeSet.Name, err)
}
} else {
serviceLevels[nodeSet.Name] = deployment.SerialServiceLevels(servicePlan.Services[nodeSet.Name])
}
deployers[nodeSet.Name].AllServices = slices.Concat(serviceLevels[nodeSet.Name]...)
}
// Start them only once every listing nodeset reached its barrier, else a
// nodeset advances while the shared Job still runs against its hosts.
if err := r.deployGlobalServices(ctx, helper, instance, servicePlan, serviceLevels, deployers); err != nil {
return ctrl.Result{}, err
}
// Deploy each nodeSet
// The loop starts and checks NodeSet deployments sequentially. However, after they
// are started, they are running in parallel, since the loop does not wait
// for the first started NodeSet to finish before starting the next.
if !instance.Spec.UseParallelExecution {
Log.Info("useParallelExecution is not set to true, services will run sequentially in list order and service dependencies (dependsOn) are ignored")
}
for i := range nodeSets.Items {
nodeSet := &nodeSets.Items[i]
if failure, ok := instance.Status.FailedNodeSets[nodeSet.Name]; ok {
Log.Info("NodeSet already terminated by a permanent failure", "nodeSet", nodeSet.Name)
nsConditions := instance.Status.NodeSetConditions[nodeSet.Name]
nsConditions.MarkFalse(
dataplanev1.NodeSetDeploymentReadyCondition,
condition.JobReasonBackoffLimitExceeded,
condition.SeverityError,
"%s", failure)
instance.Status.NodeSetConditions[nodeSet.Name] = nsConditions
haveError = true
deploymentErrMsg = appendDeploymentErrMsg(deploymentErrMsg,
fmt.Sprintf("nodeSet: %s error: %s", nodeSet.Name, failure))
continue
}
Log.Info("Deploying NodeSet", "nodeSet", nodeSet.Name)
Log.Info("Set Status.Deployed to false", "instance", instance)
instance.Status.Deployed = false
Log.Info("Set DeploymentReadyCondition false")
instance.Status.Conditions.MarkFalse(
condition.DeploymentReadyCondition, condition.RequestedReason,
condition.SeverityInfo, condition.DeploymentReadyRunningMessage)
var deployResult *ctrl.Result
deployResult, err = deployers[nodeSet.Name].Deploy(serviceLevels[nodeSet.Name])
nsConditions := instance.Status.NodeSetConditions[nodeSet.Name]
nsConditions.Set(nsConditions.Mirror(dataplanev1.NodeSetDeploymentReadyCondition))
// A barrier wait sets no service condition, so surface it on the nodeset.
if err == nil {
if blockedOn := deployers[nodeSet.Name].BlockedOn; blockedOn != "" {
nsConditions.Set(condition.FalseCondition(
dataplanev1.NodeSetDeploymentReadyCondition,
condition.RequestedReason,
condition.SeverityInfo,
"%s", fmt.Sprintf("%s: it runs once for the whole deployment",
fmt.Sprintf(dataplanev1.NodeSetServiceDeploymentReadyWaitingMessage, blockedOn))))
}
}
instance.Status.NodeSetConditions[nodeSet.Name] = nsConditions
if err != nil {
util.LogErrorForObject(helper, err, fmt.Sprintf("OpenStackDeployment error for NodeSet %s", nodeSet.Name), instance)
Log.Info("Set NodeSetDeploymentReadyCondition false", "nodeSet", nodeSet.Name)
haveError = true
deploymentErrMsg = appendDeploymentErrMsg(deploymentErrMsg,
fmt.Sprintf("nodeSet: %s error: %s", nodeSet.Name, err.Error()))
if nsConditions.Get(dataplanev1.NodeSetDeploymentReadyCondition) != nil &&
nsConditions.Get(dataplanev1.NodeSetDeploymentReadyCondition).Reason == condition.JobReasonBackoffLimitExceeded {
instance.Status.FailedNodeSets[nodeSet.Name] = err.Error()
}
}
if deployResult != nil {
shouldRequeue = true
} else {
Log.Info("OpenStackDeployment succeeded for NodeSet", "NodeSet", nodeSet.Name)
Log.Info("Set NodeSetDeploymentReadyCondition true", "nodeSet", nodeSet.Name)
nsConditions.MarkTrue(
dataplanev1.NodeSetDeploymentReadyCondition,
condition.DeploymentReadyMessage)
finishedNodeSets = append(finishedNodeSets, *nodeSet)
}
}
// A nodeSet that finished alongside a failure must still get its hashes, or the
// NodeSet controller re-deploys it because its inputs look changed.
if len(finishedNodeSets) > 0 {
err = r.setHashes(ctx, helper, instance,
dataplanev1.OpenStackDataPlaneNodeSetList{Items: finishedNodeSets})
if err != nil {
return ctrl.Result{}, err
}
}
if haveError {
settled := terminalFailure(instance)
var reason condition.Reason
reason = condition.ErrorReason
severity := condition.SeverityWarning
if settled {
reason = condition.JobReasonBackoffLimitExceeded
severity = condition.SeverityError
}
instance.Status.Conditions.MarkFalse(
condition.DeploymentReadyCondition,
reason,
severity,
condition.DeploymentReadyErrorMessage,
deploymentErrMsg)
if settled {
return ctrl.Result{}, nil
}
return ctrl.Result{}, fmt.Errorf("%s", deploymentErrMsg)
}
if shouldRequeue {
Log.Info("Not all NodeSets done for OpenStackDeployment")
return ctrl.Result{}, nil
}
Log.Info("Set DeploymentReadyCondition true")
instance.Status.Conditions.MarkTrue(condition.DeploymentReadyCondition, condition.DeploymentReadyMessage)
Log.Info("Set Status.Deployed to true", "instance", instance)
instance.Status.Deployed = true
if version != nil {
instance.Status.DeployedVersion = version.Spec.TargetVersion
}
return ctrl.Result{}, nil
}
// deployGlobalServices starts the single AnsibleEE Job of every deployment
// scoped service whose barrier has cleared.
func (r *OpenStackDataPlaneDeploymentReconciler) deployGlobalServices(
ctx context.Context,
helper *helper.Helper,
instance *dataplanev1.OpenStackDataPlaneDeployment,
servicePlan *deployment.ServicePlan,
serviceLevels map[string][][]string,
deployers map[string]*deployment.Deployer,
) error {
Log := r.GetLogger(ctx)
jobState := deployment.JobState{}
for _, serviceName := range servicePlan.GlobalServices {
gateOpen, err := deployment.GlobalServiceGateOpen(
ctx, helper, servicePlan, instance, serviceName, serviceLevels, jobState)
if err != nil {
return err
}
if !gateOpen {
Log.Info("Deployment scoped service is behind a barrier that has not cleared",
"service", serviceName, "nodeSets", servicePlan.NodesetsWithGlobalService[serviceName])
continue
}
foundService, err := servicePlan.Cache.Get(ctx, helper, serviceName)
if err != nil {
instance.Status.Conditions.MarkFalse(
condition.InputReadyCondition,
condition.ErrorReason,
condition.SeverityError,
dataplanev1.ServiceErrorMessage,
err.Error())
return err
}
listingNodeSets := servicePlan.NodesetsWithGlobalService[serviceName]
if len(listingNodeSets) == 0 {
continue
}
// Any listing nodeset works; it only supplies the Job labels and mounts.
if err := deployers[listingNodeSets[0]].DeployGlobalService(foundService); err != nil {
util.LogErrorForObject(helper, err,
fmt.Sprintf("Unable to execute deployment scoped service %s", serviceName), instance)
instance.Status.Conditions.MarkFalse(
condition.DeploymentReadyCondition,
condition.ErrorReason,
condition.SeverityError,
condition.DeploymentReadyErrorMessage,
err.Error())
return err
}
Log.Info("Started deployment scoped service", "service", serviceName, "nodeSets", listingNodeSets)
}
return nil
}
// Mirrors the delay lib-common applies to finished Jobs.
const jobCleanupDelaySeconds int32 = 10 * 60
// reconcileJobTTL defers Job cleanup until the deployment is finished: while it
// runs the Jobs are the only record of which services already ran. scheduleCleanup
// is false while the terminal verdict exists only in memory.
func (r *OpenStackDataPlaneDeploymentReconciler) reconcileJobTTL(
ctx context.Context,
helper *helper.Helper,
instance *dataplanev1.OpenStackDataPlaneDeployment,
scheduleCleanup bool,
) error {
if instance.Spec.PreserveJobs {
return nil
}
jobList := &batchv1.JobList{}
err := helper.GetClient().List(
ctx,
jobList,
client.InNamespace(instance.Namespace),
client.MatchingLabels{dataplaneutil.AnsibleEEDeploymentLabel: instance.Name},
)
if err != nil {
return err
}
finished := scheduleCleanup && instance.IsFinished()
finishedAt := instance.FinishedAt()
for i := range jobList.Items {
job := &jobList.Items[i]
if finished {
// Ageing starts at the Job's own completion time, so one schedule per Job
// is enough and the value must not be re-derived afterwards.
if job.Spec.TTLSecondsAfterFinished != nil {
continue
}
patch := client.MergeFrom(job.DeepCopy())
ttl := jobCleanupDelay(job, finishedAt)
job.Spec.TTLSecondsAfterFinished = &ttl
err := helper.GetClient().Patch(ctx, job, patch)
if err != nil && !k8s_errors.IsNotFound(err) {
return err
}
continue
}
// An older operator may have left a deadline on a running deployment.
if job.Spec.TTLSecondsAfterFinished == nil {
continue
}
patch := client.MergeFrom(job.DeepCopy())
job.Spec.TTLSecondsAfterFinished = nil
err := helper.GetClient().Patch(ctx, job, patch)
if err != nil && !k8s_errors.IsNotFound(err) {
return err
}
}
return nil
}
// jobCleanupDelay returns the TTLSecondsAfterFinished that makes the Job expire
// jobCleanupDelaySeconds after the deployment finished. Ageing starts at the
// Job's own completion time, so a Job that finished while the deployment still
// ran gets that elapsed time added back - a flat delay would delete it at once.
func jobCleanupDelay(job *batchv1.Job, finishedAt *metav1.Time) int32 {
finishTime := jobFinishTime(job)
if finishedAt == nil || finishTime == nil || !finishedAt.After(finishTime.Time) {
return jobCleanupDelaySeconds
}
delay := int64(jobCleanupDelaySeconds) + int64(math.Floor(finishedAt.Time.Sub(finishTime.Time).Seconds()))
return int32(min(delay, math.MaxInt32))
}
// jobFinishTime mirrors the TTL-after-finished controller, which measures from
// the Complete/Failed transition rather than CompletionTime.
func jobFinishTime(job *batchv1.Job) *metav1.Time {
for _, jobCondition := range job.Status.Conditions {
isFinished := (jobCondition.Type == batchv1.JobComplete || jobCondition.Type == batchv1.JobFailed) &&
jobCondition.Status == corev1.ConditionTrue
if isFinished && !jobCondition.LastTransitionTime.IsZero() {
finishTime := jobCondition.LastTransitionTime
return &finishTime
}
}
return job.Status.CompletionTime
}
// GetService retrieves a service for the OpenStackDataPlaneDeployment
func (r *OpenStackDataPlaneDeploymentReconciler) GetService(
ctx context.Context,
helper *helper.Helper,
instance *dataplanev1.OpenStackDataPlaneDeployment,
nsConditions condition.Conditions,
serviceName string,
) (dataplanev1.OpenStackDataPlaneService, error) {
service, err := deployment.GetService(ctx, helper, serviceName)
if err != nil {
instance.Status.Conditions.MarkFalse(
condition.InputReadyCondition,
condition.ErrorReason,
condition.SeverityError,
dataplanev1.ServiceErrorMessage,
err.Error())
nsConditions.MarkFalse(
dataplanev1.NodeSetDeploymentReadyCondition,
condition.ErrorReason,
condition.SeverityError,
dataplanev1.ServiceErrorMessage,
err.Error())
}
return service, err
}
func (r *OpenStackDataPlaneDeploymentReconciler) setHashes(
ctx context.Context,
helper *helper.Helper,
instance *dataplanev1.OpenStackDataPlaneDeployment,
nodeSets dataplanev1.OpenStackDataPlaneNodeSetList,
) error {
var err error
services := []string{}
if len(instance.Spec.ServicesOverride) > 0 {
services = instance.Spec.ServicesOverride
} else {
// get the union of services across nodesets
type void struct{}
var member void
s := make(map[string]void)
for _, nodeSet := range nodeSets.Items {
for _, serviceName := range nodeSet.Spec.Services {
s[serviceName] = member
}
}
for service := range s {
services = append(services, service)
}
}
for _, serviceName := range services {
err = deployment.GetDeploymentHashesForService(
ctx,
helper,
instance.Namespace,
serviceName,
instance.Status.ConfigMapHashes,
instance.Status.SecretHashes,
nodeSets)
if err != nil {
return err
}
}
for _, nodeSet := range nodeSets.Items {
err = setNodeSetAnsibleVarsFromHashes(ctx, helper, &nodeSet, instance.Status.ConfigMapHashes, instance.Status.SecretHashes)
if err != nil {
return err
}
}
// Split from the loop above: terminalFailure reads a NodeSetHashes entry as proof
// the nodeSet's input hashes are recorded, so nothing above may fail after it.
for _, nodeSet := range nodeSets.Items {
instance.Status.NodeSetHashes[nodeSet.Name] = nodeSet.Status.ConfigHash
instance.Status.BmhRefHashes[nodeSet.Name] = nodeSet.Status.BmhRefHash
}
return nil
}
// SetupWithManager sets up the controller with the Manager.
func (r *OpenStackDataPlaneDeploymentReconciler) SetupWithManager(mgr ctrl.Manager) error {
// watch for changes in certificates
certFn := func(ctx context.Context, obj client.Object) []reconcile.Request {
Log := r.GetLogger(ctx)
result := []reconcile.Request{}
objectLabelValue, ok := obj.GetLabels()[deployment.NodeSetLabel]
if !ok {
// cert doesn't have a nodeset label
return nil
}
// get all deployments in namespace
deployments := &dataplanev1.OpenStackDataPlaneDeploymentList{}
listOpts := []client.ListOption{
client.InNamespace(obj.GetNamespace()),
}
if err := r.List(context.Background(), deployments, listOpts...); err != nil {
Log.Error(err, "Unable to retrieve deployments %w")
return nil
}
for _, dep := range deployments.Items {
if dep.Status.Deployed {
continue
}
if util.StringInSlice(objectLabelValue, dep.Spec.NodeSets) {
name := client.ObjectKey{
Namespace: dep.GetNamespace(),
Name: dep.GetName(),
}
Log.Info("Cert is used by deployment", "cert", obj.GetName(), "deployment", dep.GetName())
result = append(result, reconcile.Request{NamespacedName: name})
}
}
if len(result) > 0 {
return result
}
return nil
}
return ctrl.NewControllerManagedBy(mgr).
For(&dataplanev1.OpenStackDataPlaneDeployment{},
builder.WithPredicates(predicate.Or(
predicate.GenerationChangedPredicate{},
predicate.AnnotationChangedPredicate{},
predicate.LabelChangedPredicate{}))).
Owns(&batchv1.Job{}).
Watches(&certmgrv1.Certificate{},
handler.EnqueueRequestsFromMapFunc(certFn)).
Complete(r)
}
func (r *OpenStackDataPlaneDeploymentReconciler) listNodeSets(ctx context.Context, instance *dataplanev1.OpenStackDataPlaneDeployment) (*dataplanev1.OpenStackDataPlaneNodeSetList, error) {
var nodeSets = dataplanev1.OpenStackDataPlaneNodeSetList{}
var err error
Log := r.GetLogger(ctx)
for _, nodeSet := range instance.Spec.NodeSets {
// Fetch the OpenStackDataPlaneNodeSet instance
nodeSetInstance := &dataplanev1.OpenStackDataPlaneNodeSet{}
err := r.Get(
ctx,
types.NamespacedName{
Namespace: instance.GetNamespace(),
Name: nodeSet,
},
nodeSetInstance)
if err != nil {
Log.Info("NodeSet not found", "NodeSet", nodeSet)
return &nodeSets, err
}
if err = nodeSetInstance.Spec.ValidateTLS(ctx, instance.GetNamespace(), r.Client); err != nil {
Log.Info("error while comparing TLS settings of nodeset %s with control plane: %w", nodeSet, err)
instance.Status.Conditions.MarkFalse(
dataplanev1.SetupReadyCondition,
condition.ErrorReason,
condition.SeverityError,
dataplanev1.DataPlaneNodeSetErrorMessage,
err.Error())
return &nodeSets, err
}
nodeSets.Items = append(nodeSets.Items, *nodeSetInstance)
}
return &nodeSets, err
}
// terminalFailure reports whether the deployment has a permanently failed
// nodeSet and nothing left to wait for. A nodeSet still in flight keeps this
// false even when another one already failed. Deployments that failed before
// FailedNodeSets existed carry the failure only on the top-level condition, so
// honor it as well: their backing jobs are garbage collected and must not be
// recreated.
func terminalFailure(instance *dataplanev1.OpenStackDataPlaneDeployment) bool {
if c := instance.Status.Conditions.Get(condition.DeploymentReadyCondition); c != nil &&
c.Severity == condition.SeverityError &&
c.Reason == condition.JobReasonBackoffLimitExceeded {
return true
}
if len(instance.Spec.NodeSets) == 0 || len(instance.Status.FailedNodeSets) == 0 {
return false
}
for _, nodeSet := range instance.Spec.NodeSets {
if _, failed := instance.Status.FailedNodeSets[nodeSet]; failed {
continue
}
if _, done := instance.Status.NodeSetHashes[nodeSet]; !done {
return false
}
}
return true
}
func appendDeploymentErrMsg(current string, errMsg string) string {
if len(current) == 0 {
return errMsg
}
return fmt.Sprintf("%s & %s", current, errMsg)
}
func setNodeSetAnsibleVarsFromHashes(
ctx context.Context,
helper *helper.Helper,
nodeSetInstance *dataplanev1.OpenStackDataPlaneNodeSet,
configMapHashes map[string]string, // configMapHashes from deployment.instance.configMapHashes
secretHashes map[string]string, // secretHashes from deployment.instance.secretHashes
) error {
namespace := nodeSetInstance.Namespace
// Process NodeTemplate level AnsibleVarsFrom
if err := deployment.ProcessAnsibleVarsFrom(ctx, helper, namespace, configMapHashes, secretHashes, nodeSetInstance.Spec.NodeTemplate.Ansible.AnsibleVarsFrom); err != nil {
return err
}
// Process individual Node level AnsibleVarsFrom
for _, node := range nodeSetInstance.Spec.Nodes {
if err := deployment.ProcessAnsibleVarsFrom(ctx, helper, namespace, configMapHashes, secretHashes, node.Ansible.AnsibleVarsFrom); err != nil {
return err
}
}
return nil
}