Skip to content
Navigation Menu
{{ message }}
-
Notifications
You must be signed in to change notification settings - Fork 181
Expand file tree
/
Copy pathcommon.go
More file actions
1458 lines (1250 loc) · 57 KB
/
Copy pathcommon.go
File metadata and controls
1458 lines (1250 loc) · 57 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
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
//go:build test_e2e || test_e2e_vm || sql_integration || compliance || destructive || externalbackups || test_compatibility
package tests
import (
"bytes"
"context"
"encoding/json"
"errors"
"fmt"
"io"
"math"
"os"
"os/exec"
"strings"
"testing"
"time"
v1 "github.com/stackrox/rox/generated/api/v1"
"github.com/stackrox/rox/generated/storage"
"github.com/stackrox/rox/pkg/docker/config"
"github.com/stackrox/rox/pkg/pointers"
"github.com/stackrox/rox/pkg/protocompat"
"github.com/stackrox/rox/pkg/retry"
"github.com/stackrox/rox/pkg/retryablehttp"
"github.com/stackrox/rox/pkg/search"
"github.com/stackrox/rox/pkg/testutils"
"github.com/stackrox/rox/pkg/testutils/centralgrpc"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
"github.com/stretchr/testify/suite"
"google.golang.org/grpc"
appsV1 "k8s.io/api/apps/v1"
coreV1 "k8s.io/api/core/v1"
apiErrors "k8s.io/apimachinery/pkg/api/errors"
metaV1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/labels"
"k8s.io/apimachinery/pkg/types"
"k8s.io/apimachinery/pkg/util/intstr"
"k8s.io/apimachinery/pkg/util/yaml"
"k8s.io/client-go/kubernetes"
_ "k8s.io/client-go/plugin/pkg/client/auth"
"k8s.io/client-go/rest"
"k8s.io/client-go/tools/clientcmd"
)
//lint:file-ignore U1000 since common.go is included in several different go:build tags but not every function is used
// within every tag, hence the linter incorrectly shows most functions in this file as unused.
const (
nginxDeploymentName = `nginx`
expectedLatestTagPolicy = `Latest tag`
sensorDeployment = "sensor"
sensorContainer = "sensor"
waitTimeout = 5 * time.Minute
timeout = 30 * time.Second
)
var (
sensorPodLabels = map[string]string{"app": "sensor"}
)
// testutilsLogger adapts testutils.T to retryablehttp.Logger interface.
//
// WHY THIS EXISTS (and why it's unfortunate):
// The retryablehttp library requires a logger implementing its Logger interface with Printf method.
// Go's *testing.T has Logf but NOT Printf (different method names, same functionality).
// We cannot add Printf to testutils.T because that would break compatibility with *testing.T,
// which is used throughout the codebase (e.g., centralgrpc.GRPCConnectionToCentral(t testutils.T)).
//
// CLEANER ALTERNATIVE:
// Accept *testing.T directly in helper functions and create testutils.T wrappers locally where needed
// (specifically for the retry mechanism). This would avoid the interface constraint propagating everywhere.
// However, this would be a larger refactor affecting many test helper functions.
//
// CURRENT COMPROMISE:
// Use this tiny adapter ONLY where retryablehttp requires it. Everywhere else uses testutils.T naturally.
type testutilsLogger struct{ testutils.T }
func (l testutilsLogger) Printf(format string, v ...interface{}) { l.Logf(format, v...) }
// logf logs using the testing logger, prefixing a high-resolution timestamp.
// Using testing.T.Logf means that the output is hidden unless the test fails or verbose logging is enabled with -v.
func logf(t *testing.T, format string, args ...any) {
t.Logf(time.Now().Format(time.StampMilli)+" "+format, args...)
}
// testContexts returns a couple of contexts for the given test: with a timeout for the testing logic and another
// with an additional longer timeout for debug artifact gathering and cleanup.
func testContexts(t *testing.T, name string, timeout time.Duration) (testCtx context.Context, overallCtx context.Context, cancel func()) {
var (
overallCancel func()
testCancel func()
)
cleanupTimeout := 10 * time.Minute
logf(t, "Running %s with a timeout of %s plus %s for cleanup", name, timeout, cleanupTimeout)
overallTimeout := timeout + cleanupTimeout
overallErr := fmt.Errorf("overall %s test+cleanup timeout of %s reached", name, overallTimeout)
testErr := fmt.Errorf("%s test timeout of %s reached", name, timeout)
overallCtx, overallCancel = context.WithTimeoutCause(context.Background(), overallTimeout, overallErr)
testCtx, testCancel = context.WithTimeoutCause(overallCtx, timeout, testErr)
cancel = func() {
testCancel()
overallCancel()
}
return
}
// mustGetEnv calls os.GetEnv and fails the test if result is empty.
func mustGetEnv(t *testing.T, varName string) string {
val := os.Getenv(varName)
require.NotEmptyf(t, val, "Environment variable %q must be set.", varName)
return val
}
func retrieveDeployment(service v1.DeploymentServiceClient, deploymentID string) (*storage.Deployment, error) {
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
return service.GetDeployment(ctx, &v1.ResourceByID{Id: deploymentID})
}
func retrieveDeployments(service v1.DeploymentServiceClient, deps []*storage.ListDeployment) ([]*storage.Deployment, error) {
deployments := make([]*storage.Deployment, 0, len(deps))
for _, d := range deps {
deployment, err := retrieveDeployment(service, d.GetId())
if err != nil {
return nil, err
}
deployments = append(deployments, deployment)
}
return deployments, nil
}
func waitForDeploymentCountInCentral(t testutils.T, query string, count int) {
conn := centralgrpc.GRPCConnectionToCentral(t)
service := v1.NewDeploymentServiceClient(conn)
ticker := time.NewTicker(time.Second)
defer ticker.Stop()
timer := time.NewTimer(waitTimeout)
defer timer.Stop()
for {
select {
case <-ticker.C:
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
deploymentCount, err := service.CountDeployments(ctx, &v1.RawQuery{Query: query})
cancel()
if err != nil {
t.Logf("Error listing deployments: %s", err)
continue
}
if deploymentCount.GetCount() == int32(count) {
return
}
case <-timer.C:
t.Fatalf("Timed out waiting for deployments %q", query)
}
}
}
func waitForDeploymentReadyInK8s(t testutils.T, deploymentName, namespace string) {
client := createK8sClient(t)
ticker := time.NewTicker(2 * time.Second)
defer ticker.Stop()
timer := time.NewTimer(waitTimeout)
defer timer.Stop()
ctx := context.Background()
t.Logf("Waiting for deployment %q in namespace %q to be ready in Kubernetes", deploymentName, namespace)
for {
select {
case <-ticker.C:
deploy, err := client.AppsV1().Deployments(namespace).Get(ctx, deploymentName, metaV1.GetOptions{})
if err != nil {
if apiErrors.IsNotFound(err) {
t.Logf("Deployment %q in namespace %q not found yet, waiting...", deploymentName, namespace)
continue
}
t.Logf("Error getting deployment %q from namespace %q: %v", deploymentName, namespace, err)
continue
}
// Check if generation matches observed generation.
// Practical Example: generation: 5, observedGeneration: 4
// This tells you: "Someone just updated the deployment (5th change),
// but the controller is still working on the 4th revision—be patient, reconciliation is in progress."
// Without these numbers, you'd just see repeated "NOT ready" messages without understanding the root cause.
if deploy.GetGeneration() != deploy.Status.ObservedGeneration {
t.Logf("Deployment %q in namespace %q NOT ready: generation %d != observed generation %d",
deploymentName, namespace, deploy.GetGeneration(), deploy.Status.ObservedGeneration)
continue
}
// Check if all replicas are ready
if deploy.Status.Replicas == 0 || deploy.Status.Replicas != deploy.Status.ReadyReplicas {
t.Logf("Deployment %q in namespace %q NOT ready: %d/%d ready replicas",
deploymentName, namespace, deploy.Status.ReadyReplicas, deploy.Status.Replicas)
continue
}
// Ensure all pods are from the current generation (no old pods during rollout).
if deploy.Spec.Replicas != nil && deploy.Status.UpdatedReplicas < *deploy.Spec.Replicas {
t.Logf("Deployment %q in namespace %q NOT ready: rollout incomplete (%d/%d updated replicas)",
deploymentName, namespace, deploy.Status.UpdatedReplicas, *deploy.Spec.Replicas)
continue
}
if deploy.Status.Replicas > deploy.Status.UpdatedReplicas {
t.Logf("Deployment %q in namespace %q NOT ready: old replicas pending termination (%d total, %d updated)",
deploymentName, namespace, deploy.Status.Replicas, deploy.Status.UpdatedReplicas)
continue
}
if deploy.Status.AvailableReplicas < deploy.Status.UpdatedReplicas {
t.Logf("Deployment %q in namespace %q NOT ready: updated replicas not yet available (%d/%d available)",
deploymentName, namespace, deploy.Status.AvailableReplicas, deploy.Status.UpdatedReplicas)
continue
}
// Check conditions for additional insights
printDeploymentConditions(t, deploy.Status.Conditions, deploymentName, namespace)
t.Logf("Deployment %q in namespace %q READY in Kubernetes (%d/%d ready replicas)",
deploymentName, namespace, deploy.Status.ReadyReplicas, deploy.Status.Replicas)
return
case <-timer.C:
// Get final status for error message
deploy, err := client.AppsV1().Deployments(namespace).Get(ctx, deploymentName, metaV1.GetOptions{})
if err != nil {
require.NoError(t, err, "Timed out waiting for deployment %q in namespace %q. Failed to get final status: %v", deploymentName, namespace, err)
}
require.Failf(t, "Timed out waiting for deployment in Kubernetes",
"Deployment %q in namespace %q did not become ready within %v.\nStatus: Replicas=%d, ReadyReplicas=%d, UpdatedReplicas=%d, AvailableReplicas=%d, UnavailableReplicas=%d\nGeneration=%d, ObservedGeneration=%d",
deploymentName, namespace, waitTimeout,
deploy.Status.Replicas, deploy.Status.ReadyReplicas, deploy.Status.UpdatedReplicas,
deploy.Status.AvailableReplicas, deploy.Status.UnavailableReplicas,
deploy.GetGeneration(), deploy.Status.ObservedGeneration)
}
}
}
func printDeploymentConditions(t testutils.T, conditions []appsV1.DeploymentCondition, deploymentName, namespace string) {
for _, cond := range conditions {
if cond.Type == appsV1.DeploymentAvailable && cond.Status != coreV1.ConditionTrue {
t.Logf("Deployment %q in namespace %q NOT ready: Available condition is %s: %s",
deploymentName, namespace, cond.Status, cond.Message)
}
if cond.Type == appsV1.DeploymentProgressing && cond.Status != coreV1.ConditionTrue {
t.Logf("Deployment %q in namespace %q NOT ready: Progressing condition is %s: %s",
deploymentName, namespace, cond.Status, cond.Message)
}
}
}
func waitForDeploymentInCentral(t testutils.T, deploymentName string) {
conn := centralgrpc.GRPCConnectionToCentral(t)
service := v1.NewDeploymentServiceClient(conn)
ticker := time.NewTicker(time.Second)
defer ticker.Stop()
timer := time.NewTimer(waitTimeout)
defer timer.Stop()
qb := search.NewQueryBuilder().AddExactMatches(search.DeploymentName, deploymentName)
for {
select {
case <-ticker.C:
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
listDeployments, err := service.ListDeployments(ctx, &v1.RawQuery{
Query: qb.Query(),
},
)
cancel()
if err != nil {
t.Logf("Error listing deployments: %s", err)
continue
}
deployments, err := retrieveDeployments(service, listDeployments.GetDeployments())
if err != nil {
t.Logf("Error retrieving deployments: %s", err)
continue
}
if len(deployments) > 0 {
d := deployments[0]
if len(d.GetContainers()) > 0 && d.GetContainers()[0].GetImage().GetId() != "" {
return
}
}
case <-timer.C:
t.Fatalf("Timed out waiting for deployment %s", deploymentName)
}
}
}
// waitForAlert waits for the desired number of alerts to appear in Central.
// It polls the AlertService every 2 seconds for up to 90 seconds (45 attempts).
func waitForAlert(t *testing.T, service v1.AlertServiceClient, req *v1.ListAlertsRequest, desired int) {
var alerts []*storage.ListAlert
// Retry until desired alert count is reached when sensor(s) resync
for i := 0; i < 45; i++ {
ctx, cancel := context.WithTimeout(context.Background(), time.Minute)
resp, err := service.ListAlerts(ctx, req)
cancel()
require.NoError(t, err)
alerts = resp.GetAlerts()
if len(alerts) == desired {
return
}
time.Sleep(2 * time.Second)
}
var alertStrings strings.Builder
for _, alert := range alerts {
alertStrings.WriteString(protocompat.MarshalTextString(alert))
alertStrings.WriteString("\n")
}
t.Logf("Received alerts:\n%s", alertStrings.String())
require.Fail(t, fmt.Sprintf("Failed to have %d alerts, instead received %d alerts", desired, len(alerts)))
}
func waitForTermination(t testutils.T, deploymentName string) {
conn := centralgrpc.GRPCConnectionToCentral(t)
service := v1.NewDeploymentServiceClient(conn)
ticker := time.NewTicker(time.Second)
defer ticker.Stop()
timer := time.NewTimer(waitTimeout)
defer timer.Stop()
query := search.NewQueryBuilder().AddStrings(search.DeploymentName, deploymentName).Query()
for {
select {
case <-ticker.C:
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
listDeployments, err := service.ListDeployments(ctx, &v1.RawQuery{
Query: query,
})
cancel()
if err != nil {
t.Logf("Error listing deployments: %v", err)
continue
}
if len(listDeployments.GetDeployments()) == 0 {
return
}
case <-timer.C:
t.Fatalf("Timed out waiting for deployment %s to stop", deploymentName)
}
}
}
func getPodFromFile(t testutils.T, path string) *coreV1.Pod {
data, err := os.ReadFile(path)
require.NoError(t, err)
decoder := yaml.NewYAMLOrJSONDecoder(bytes.NewBuffer(data), len(data))
var pod coreV1.Pod
err = decoder.Decode(&pod)
require.NoError(t, err)
return &pod
}
func setupDeploymentInNamespace(t *testing.T, image, deploymentName, namespace string) {
setupDeploymentWithReplicasInNamespace(t, image, deploymentName, 1, namespace)
}
func setupDeploymentWithReplicas(t *testing.T, image, deploymentName string, replicas int) {
setupDeploymentWithReplicasInNamespace(t, image, deploymentName, replicas, "default")
}
func setupDeploymentWithReplicasInNamespace(t *testing.T, image, deploymentName string, replicas int, namespace string) {
setupDeploymentNoWaitInNamespace(t, image, deploymentName, replicas, namespace)
waitForDeploymentReadyInK8s(t, deploymentName, namespace)
waitForDeploymentInCentral(t, deploymentName)
}
func setupDeploymentNoWait(t *testing.T, image, deploymentName string, replicas int) {
setupDeploymentNoWaitInNamespace(t, image, deploymentName, replicas, "default")
}
func setupDeploymentNoWaitInNamespace(t *testing.T, image, deploymentName string, replicas int, namespace string) {
require.NoError(t, createDeploymentViaAPI(t, image, deploymentName, replicas, namespace))
}
// buildContainer constructs a container spec.
func buildContainer(name, image string, pullPolicy coreV1.PullPolicy) coreV1.Container {
container := coreV1.Container{
Name: name,
Image: image,
ImagePullPolicy: pullPolicy,
Resources: coreV1.ResourceRequirements{}, // Match kubectl behavior
}
return container
}
// createDeploymentViaAPI creates a Kubernetes deployment using the K8s API client.
// Mirrors qa-tests-backend/src/main/groovy/orchestratormanager/Kubernetes.groovy:2316-2318
// to support IMAGE_PULL_POLICY_FOR_QUAY_IO for prefetched images.
func createDeploymentViaAPI(t *testing.T, image, deploymentName string, replicas int, namespace string) error {
client := createK8sClient(t)
t.Logf("Creating deployment %q in namespace %q with image %q and %d replicas", deploymentName, namespace, image, replicas)
// Determine imagePullPolicy - allow override ONLY for actual quay.io/ images.
// NOTE: This intentionally does NOT apply to mirrored images (e.g., icsp.invalid, idms.invalid)
// as those are used to test mirroring functionality and should use their own pull behavior.
pullPolicy := coreV1.PullIfNotPresent
if policy := os.Getenv("IMAGE_PULL_POLICY_FOR_QUAY_IO"); policy != "" && strings.HasPrefix(image, "quay.io/") {
pullPolicy = coreV1.PullPolicy(policy)
t.Logf("Setting imagePullPolicy=%s for quay.io image (IMAGE_PULL_POLICY_FOR_QUAY_IO)", policy)
}
// Build deployment object
deployment := &appsV1.Deployment{
ObjectMeta: metaV1.ObjectMeta{
Name: deploymentName,
Namespace: namespace,
Labels: map[string]string{
"app": deploymentName,
},
},
Spec: appsV1.DeploymentSpec{
Replicas: pointers.Int32(int32(replicas)),
Selector: &metaV1.LabelSelector{
MatchLabels: map[string]string{"app": deploymentName},
},
Template: coreV1.PodTemplateSpec{
ObjectMeta: metaV1.ObjectMeta{
Labels: map[string]string{"app": deploymentName},
},
Spec: coreV1.PodSpec{
Containers: []coreV1.Container{buildContainer(deploymentName, image, pullPolicy)},
},
},
},
}
t.Logf("Deployment object created: name=%s, namespace=%s, replicas=%d, image=%s, labels=%v",
deployment.Name, deployment.Namespace, *deployment.Spec.Replicas, image, deployment.Labels)
// Create the deployment with timeout
ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
defer cancel()
t.Logf("Calling K8s API to create deployment %q in namespace %q...", deploymentName, namespace)
createdDeployment, err := client.AppsV1().Deployments(namespace).Create(ctx, deployment, metaV1.CreateOptions{})
if err != nil {
// Detailed error logging
if apiErrors.IsAlreadyExists(err) {
t.Logf("ERROR: Deployment %q already exists in namespace %q: %v", deploymentName, namespace, err)
} else if apiErrors.IsInvalid(err) {
t.Logf("ERROR: Deployment %q spec is invalid: %v", deploymentName, err)
} else if apiErrors.IsForbidden(err) {
t.Logf("ERROR: Permission denied creating deployment %q in namespace %q: %v", deploymentName, namespace, err)
} else if apiErrors.IsTimeout(err) {
t.Logf("ERROR: Timeout creating deployment %q in namespace %q after 30s: %v", deploymentName, namespace, err)
} else if apiErrors.IsServerTimeout(err) {
t.Logf("ERROR: Server timeout creating deployment %q in namespace %q: %v", deploymentName, namespace, err)
} else if apiErrors.IsServiceUnavailable(err) {
t.Logf("ERROR: K8s API service unavailable when creating deployment %q: %v", deploymentName, err)
} else {
t.Logf("ERROR: Unexpected error creating deployment %q in namespace %q: %v (type: %T)", deploymentName, namespace, err, err)
}
// Log deployment conditions only if deployment was partially created (useful for debugging failures)
if createdDeployment != nil && len(createdDeployment.Status.Conditions) > 0 {
t.Logf("Deployment %q has %d status conditions:", deploymentName, len(createdDeployment.Status.Conditions))
for i, cond := range createdDeployment.Status.Conditions {
t.Logf(" Condition[%d]: Type=%s, Status=%s, Reason=%s, Message=%q, LastUpdateTime=%v",
i, cond.Type, cond.Status, cond.Reason, cond.Message, cond.LastUpdateTime)
}
}
return fmt.Errorf("failed to create deployment %q: %w", deploymentName, err)
}
t.Logf("Deployment %q successfully created in namespace %q", deploymentName, namespace)
t.Logf("Deployment UID: %s, ResourceVersion: %s, Generation: %d",
createdDeployment.UID, createdDeployment.ResourceVersion, createdDeployment.Generation)
t.Logf("Deployment status: Replicas=%d, UpdatedReplicas=%d, ReadyReplicas=%d, AvailableReplicas=%d, UnavailableReplicas=%d",
createdDeployment.Status.Replicas,
createdDeployment.Status.UpdatedReplicas,
createdDeployment.Status.ReadyReplicas,
createdDeployment.Status.AvailableReplicas,
createdDeployment.Status.UnavailableReplicas)
t.Logf("Deployment %q creation completed successfully", deploymentName)
return nil
}
func setImage(t *testing.T, deploymentName string, deploymentID string, containerName string, image string) {
cmd := exec.Command(`kubectl`, `set`, `image`, fmt.Sprintf("deployment/%s", deploymentName), fmt.Sprintf("%s=%s", containerName, image))
output, err := cmd.CombinedOutput()
require.NoError(t, err, string(output))
conn := centralgrpc.GRPCConnectionToCentral(t)
service := v1.NewDeploymentServiceClient(conn)
waitForCondition(t, func() bool {
deployment, err := retrieveDeployment(service, deploymentID)
if err != nil {
t.Logf("Error retrieving deployment: %v", err)
return false
}
containers := deployment.GetContainers()
for _, container := range containers {
if container.GetImage().GetName().GetFullName() != image {
return false
}
}
t.Logf("Image set to %s for deployment %s(%s) container %s", image, deploymentName, deploymentID, containerName)
return true
}, "image updated", time.Minute, 5*time.Second)
}
// ensurePodExists creates a pod in Kubernetes. If the pod already exists, this is a no-op.
// This makes the function idempotent and safe to retry.
func ensurePodExists(t testutils.T, client kubernetes.Interface, pod *coreV1.Pod) {
ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
defer cancel()
t.Logf("Ensuring pod %s %s exists", pod.GetNamespace(), pod.GetName())
_, err := client.CoreV1().Pods(pod.GetNamespace()).Create(ctx, pod, metaV1.CreateOptions{})
if err != nil && !apiErrors.IsAlreadyExists(err) {
require.NoError(t, err)
}
if apiErrors.IsAlreadyExists(err) {
t.Logf("Pod %s already exists, continuing", pod.GetName())
}
}
// waitForPodRunning waits for a Kubernetes pod to be in Running phase with all containers ready.
// It polls the pod status with retries and provides detailed error messages about pod and container states.
// Timeout is set to 3 minutes to handle slow CI environments (image pull, scheduling, etc.).
func waitForPodRunning(t testutils.T, client kubernetes.Interface, podNamespace, podName string) *coreV1.Pod {
// Increased timeout to 3 minutes to handle slow CI environments (image pull, scheduling, etc.)
ctx, cancel := context.WithTimeout(context.Background(), 3*time.Minute)
defer cancel()
var k8sPod *coreV1.Pod
// Increased from 30×2s (60s) to 60×3s (180s) to account for slower pod startup in CI
testutils.Retry(t, 60, 3*time.Second, func(waitT testutils.T) {
var err error
k8sPod, err = client.CoreV1().Pods(podNamespace).Get(ctx, podName, metaV1.GetOptions{})
require.NoError(waitT, err, "failed to get pod %s", podName)
// Log pod and container status for debugging
// Note: ImagePullBackOff, ErrImagePull, etc. appear in container status, not pod status
var logMsg strings.Builder
logMsg.WriteString(fmt.Sprintf("Pod phase: %s, Reason: %q, Message: %q",
k8sPod.Status.Phase, k8sPod.Status.Reason, k8sPod.Status.Message))
var containerInfo strings.Builder
for _, status := range k8sPod.Status.ContainerStatuses {
// Build log message for non-ready containers
if !status.Ready {
if status.State.Waiting != nil {
logMsg.WriteString(fmt.Sprintf(", Container %q: %q", status.Name, status.State.Waiting.Reason))
} else if status.State.Terminated != nil {
logMsg.WriteString(fmt.Sprintf(", Container %q: Terminated (%q)", status.Name, status.State.Terminated.Reason))
}
}
// Build detailed info for error message (always, in case pod is not running)
containerInfo.WriteString(fmt.Sprintf("\n - %s: ready=%v, started=%v",
status.Name, status.Ready, status.Started != nil && *status.Started))
if status.State.Waiting != nil {
containerInfo.WriteString(fmt.Sprintf(", waiting: %s - %s",
status.State.Waiting.Reason, status.State.Waiting.Message))
}
}
waitT.Logf(logMsg.String())
// Provide detailed error message if pod is not running
if k8sPod.Status.Phase != coreV1.PodRunning {
require.Failf(waitT, "pod not in Running phase",
"Pod %s is in %s phase (expected Running)\nContainers:%s\nPod Reason: %s\nPod Message: %s",
podName, k8sPod.Status.Phase, containerInfo.String(),
k8sPod.Status.Reason, k8sPod.Status.Message)
}
// Ensure all containers are ready before checking for process events
for _, status := range k8sPod.Status.ContainerStatuses {
require.True(waitT, status.Ready, "container %s not ready (state: %+v)",
status.Name, status.State)
}
})
t.Logf("Pod %s is running with all containers ready in Kubernetes", k8sPod.Name)
return k8sPod
}
func teardownPod(t testutils.T, client kubernetes.Interface, pod *coreV1.Pod) {
ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
defer cancel()
err := client.CoreV1().Pods(pod.GetNamespace()).Delete(ctx, pod.GetName(), metaV1.DeleteOptions{GracePeriodSeconds: pointers.Int64(0)})
require.NoError(t, err)
waitForTermination(t, pod.GetName())
}
// teardownDeploymentInternal handles deployment deletion with configurable verification.
// When waitForCompletion is true, it retries deletion and waits for both K8s and Central cleanup.
// When false, it only issues the delete command without waiting or failing on errors.
func teardownDeploymentInternal(t *testing.T, deploymentName string, namespace string, waitForCompletion bool) {
client := createK8sClient(t)
deletePolicy := metaV1.DeletePropagationForeground
gracePeriod := int64(1)
deleteFunc := func() error {
ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
defer cancel()
logf(t, "Deleting deployment %q in namespace %q via API with propagation policy %s", deploymentName, namespace, deletePolicy)
err := client.AppsV1().Deployments(namespace).Delete(ctx, deploymentName, metaV1.DeleteOptions{
GracePeriodSeconds: &gracePeriod,
PropagationPolicy: &deletePolicy,
})
if err != nil {
if apiErrors.IsNotFound(err) {
logf(t, "Deployment %q not found in namespace %q (already deleted)", deploymentName, namespace)
return nil
}
return fmt.Errorf("failed to delete deployment %q in namespace %q: %w", deploymentName, namespace, err)
}
logf(t, "Successfully initiated deletion of deployment %q in namespace %q", deploymentName, namespace)
return nil
}
if waitForCompletion {
// Retry the delete + wait sequence if the deployment doesn't delete within 15 seconds
err := retry.WithRetry(
func() error {
if err := deleteFunc(); err != nil {
return retry.MakeRetryable(err)
}
// Verify deployment is actually deleted from Kubernetes
return waitForK8sDeploymentDeletion(t, client, deploymentName, namespace, 15*time.Second)
},
retry.Tries(4), // Try up to 4 times (1 initial + 3 retries)
retry.OnlyRetryableErrors(), // Only retry on retriable errors
retry.BetweenAttempts(func(int) {
logf(t, "Retrying deployment %q deletion in namespace %q", deploymentName, namespace)
}),
)
require.NoError(t, err, "Failed to delete deployment %q in namespace %q after retries", deploymentName, namespace)
// Wait for Central to recognize the deletion
waitForTermination(t, deploymentName)
} else {
// Fire and forget - don't fail the test if deletion fails
if err := deleteFunc(); err != nil {
logf(t, "Deployment %q deletion in namespace %q failed (non-fatal): %v", deploymentName, namespace, err)
}
}
}
func teardownDeployment(t *testing.T, deploymentName string, namespace string) {
teardownDeploymentInternal(t, deploymentName, namespace, true)
}
// waitForK8sDeploymentDeletion polls Kubernetes to verify the deployment is actually gone.
// Returns an error if the deployment still exists after the timeout, which triggers a retry
// of the entire delete operation.
func waitForK8sDeploymentDeletion(t *testing.T, client kubernetes.Interface, deploymentName string, namespace string, timeout time.Duration) error {
ticker := time.NewTicker(time.Second)
defer ticker.Stop()
timer := time.NewTimer(timeout)
defer timer.Stop()
attempt := 0
for {
select {
case <-ticker.C:
attempt++
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
_, err := client.AppsV1().Deployments(namespace).Get(ctx, deploymentName, metaV1.GetOptions{})
cancel()
if apiErrors.IsNotFound(err) {
logf(t, "Verified: deployment %q deleted from namespace %q after %d attempt(s)", deploymentName, namespace, attempt)
return nil
}
if err != nil {
logf(t, "Error checking deployment %q status in namespace %q (attempt %d): %v", deploymentName, namespace, attempt, err)
continue
}
logf(t, "Deployment %q still exists in namespace %q (attempt %d)", deploymentName, namespace, attempt)
case <-timer.C:
return retry.MakeRetryable(fmt.Errorf("deployment %s in namespace %s still exists in Kubernetes after %v", deploymentName, namespace, timeout))
}
}
}
func teardownDeploymentWithoutCheck(t *testing.T, deploymentName string, namespace string) {
// In cases where deployment will not impact other tests,
// we can trigger deletion and assume that it will be deleted eventually.
teardownDeploymentInternal(t, deploymentName, namespace, false)
}
func getConfig(t testutils.T) *rest.Config {
config, err := clientcmd.NewDefaultClientConfigLoadingRules().Load()
require.NoError(t, err, "could not load default Kubernetes client config")
restCfg, err := clientcmd.NewDefaultClientConfig(*config, &clientcmd.ConfigOverrides{}).ClientConfig()
require.NoError(t, err, "could not get REST client config from kubernetes config")
configureRetryableTransport(t, restCfg)
return restCfg
}
// configureRetryableTransport configures a rest.Config to use retryable HTTP client
// for network resilience. This adds automatic retry logic for transient network errors.
func configureRetryableTransport(t testutils.T, restCfg *rest.Config) {
if restCfg.Timeout == 0 {
restCfg.Timeout = 30 * time.Second
}
retryablehttp.ConfigureRESTConfig(restCfg,
retryablehttp.WithLogger(&testutilsLogger{t}),
)
}
func createK8sClient(t testutils.T) kubernetes.Interface {
return createK8sClientWithConfig(t, getConfig(t))
}
func createK8sClientWithConfig(t testutils.T, restCfg *rest.Config) kubernetes.Interface {
k8sClient, err := kubernetes.NewForConfig(restCfg)
require.NoError(t, err, "creating Kubernetes client from REST config")
return k8sClient
}
// getClusterID returns the ID of the cluster in Central.
// Fails if there is not exactly one cluster (E2E tests assume single-cluster environment).
func getClusterID(t *testing.T) string {
conn := centralgrpc.GRPCConnectionToCentral(t)
clusterService := v1.NewClustersServiceClient(conn)
ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
defer cancel()
resp, err := clusterService.GetClusters(ctx, &v1.GetClustersRequest{})
require.NoError(t, err, "getting clusters")
require.Len(t, resp.GetClusters(), 1, "expected exactly one cluster, found %d", len(resp.GetClusters()))
return resp.GetClusters()[0].GetId()
}
// setClusterLabels sets cluster labels using Central's Cluster API.
// Pass nil to remove all labels.
// Works for both operator-managed and helm-managed deployments.
func setClusterLabels(t *testing.T, clusterID string, labels map[string]string) {
conn := centralgrpc.GRPCConnectionToCentral(t)
clusterService := v1.NewClustersServiceClient(conn)
ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
defer cancel()
// Get current cluster
resp, err := clusterService.GetCluster(ctx, &v1.ResourceByID{Id: clusterID})
require.NoError(t, err, "getting cluster %s", clusterID)
cluster := resp.GetCluster()
require.NotNil(t, cluster, "GetCluster returned nil cluster for ID %s", clusterID)
// Update labels
cluster.Labels = labels
// Put updated cluster
ctx2, cancel2 := context.WithTimeout(context.Background(), 30*time.Second)
defer cancel2()
_, err = clusterService.PutCluster(ctx2, cluster)
require.NoError(t, err, "updating cluster %s with labels %v", clusterID, labels)
if labels == nil {
t.Logf("Removed cluster labels")
} else {
t.Logf("Set cluster labels: %v", labels)
}
}
// createNamespaceWithLabels creates a namespace with the specified labels.
func createNamespaceWithLabels(t *testing.T, name string, labels map[string]string) {
client := createK8sClient(t)
namespace := &coreV1.Namespace{
ObjectMeta: metaV1.ObjectMeta{
Name: name,
Labels: labels,
},
}
ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
defer cancel()
_, err := client.CoreV1().Namespaces().Create(ctx, namespace, metaV1.CreateOptions{})
require.NoError(t, err, "creating namespace %q with labels %v", name, labels)
t.Logf("Created namespace %q with labels: %v", name, labels)
}
// deleteNamespace deletes a namespace.
func deleteNamespace(t *testing.T, name string) {
client := createK8sClient(t)
ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
defer cancel()
err := client.CoreV1().Namespaces().Delete(ctx, name, metaV1.DeleteOptions{})
if err != nil && !apiErrors.IsNotFound(err) {
t.Logf("Failed to delete namespace %q: %v", name, err)
} else {
t.Logf("Deleted namespace %q", name)
}
}
// execInDeployment executes a command in a pod from the given deployment.
// Assumes deployment pods have label app=<deploymentName> (set by privilegedDeploymentSpec).
func execInDeployment(t *testing.T, client kubernetes.Interface, deploymentName, namespace string, command ...string) {
ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
defer cancel()
podList, err := client.CoreV1().Pods(namespace).List(ctx, metaV1.ListOptions{
LabelSelector: fmt.Sprintf("app=%s", deploymentName),
})
require.NoError(t, err, "listing pods for deployment %q", deploymentName)
require.NotEmpty(t, podList.Items, "no pods found for deployment %q", deploymentName)
podName := podList.Items[0].Name
args := make([]string, 0, 5+len(command))
args = append(args, "exec", "-n", namespace, podName, "--")
args = append(args, command...)
cmd := exec.CommandContext(ctx, "kubectl", args...)
output, err := cmd.CombinedOutput()
if err != nil {
t.Logf("kubectl exec output: %s", string(output))
}
require.NoError(t, err, "executing command %v in pod %q", command, podName)
t.Logf("Executed command %v in pod %q (deployment %q)", command, podName, deploymentName)
}
func waitForCondition(t testutils.T, condition func() bool, desc string, timeout time.Duration, frequency time.Duration) {
ticker := time.NewTicker(frequency)
defer ticker.Stop()
timer := time.NewTimer(timeout)
defer timer.Stop()
for {
select {
case <-ticker.C:
if condition() {
return
}
case <-timer.C:
t.Fatalf("Timed out waiting for condition %s to stop", desc)
}
}
}
type KubernetesSuite struct {
suite.Suite
k8s kubernetes.Interface
}
func (ks *KubernetesSuite) SetupSuite() {
ks.k8s = createK8sClient(ks.T())
}
func (ks *KubernetesSuite) logf(format string, args ...any) {
logf(ks.T(), format, args...)
}
type logMatcher interface {
Match(reader io.ReadSeeker) (bool, error)
fmt.Stringer
}
// waitUntilLog waits until ctx expires or logs of container in all pods matching podLabels satisfy all logMatchers.
func (ks *KubernetesSuite) waitUntilLog(ctx context.Context, namespace string, podLabels map[string]string, container string, description string, logMatchers ...logMatcher) {
ls := labels.SelectorFromSet(podLabels).String()
checkLogs := ks.checkLogsClosure(ctx, namespace, ls, container, logMatchers...)
ks.logf("Waiting until %q pods logs %s: %s", ls, description, logMatchers)
mustEventually(ks.T(), ctx, checkLogs, 10*time.Second, fmt.Sprintf("Not all %q pods logs %s", ls, description))
}
// checkLogsMatch will check if the logs of container in all pods matching pod labels satisfy all log matchers.
func (ks *KubernetesSuite) checkLogsMatch(ctx context.Context, namespace string, podLabels map[string]string, container string, description string, logMatchers ...logMatcher) {
ls := labels.SelectorFromSet(podLabels).String()
ks.logf("Checking %q pods logs %s: %s", ls, description, logMatchers)
err := ks.checkLogsClosure(ctx, namespace, ls, container, logMatchers...)()
require.NoError(ks.T(), err, fmt.Sprintf("%q was untrue", description))
}
// checkLogsClosure returns a function that checks if the logs of container in all pods matching pod labels satisfy all the log matchers.
func (ks *KubernetesSuite) checkLogsClosure(ctx context.Context, namespace, labelSelector, container string, logMatchers ...logMatcher) func() error {
return func() error {
podList, err := ks.k8s.CoreV1().Pods(namespace).List(ctx, metaV1.ListOptions{LabelSelector: labelSelector})
if err != nil {
return fmt.Errorf("could not list pods matching %q in namespace %q: %w", labelSelector, namespace, err)
}
if len(podList.Items) == 0 {
if ok, err := allMatch(strings.NewReader(""), logMatchers...); ok {
return nil
} else if err != nil {
return fmt.Errorf("empty list of pods caused failure: %w", err)
}
return errors.New("empty list of pods does not satisfy the condition")
}
for _, pod := range podList.Items {
resp := ks.k8s.CoreV1().Pods(namespace).GetLogs(pod.GetName(), &coreV1.PodLogOptions{Container: container}).Do(ctx)
log, err := resp.Raw()
if err != nil {
return fmt.Errorf("retrieving logs of pod %q in namespace %q failed: %w", pod.GetName(), namespace, err)
}
if ok, err := allMatch(bytes.NewReader(log), logMatchers...); ok {
continue
} else if err != nil {
return fmt.Errorf("log of pod %q in namespace %q caused failure: %w", pod.GetName(), namespace, err)
}
return fmt.Errorf("log of pod %q in namespace %q does not satisfy the condition", pod.GetName(), namespace)
}
return nil
}
}
// getLastLogBytePos will return the position of the last log byte for a particular pod. This is intended to be used with
// the log matchers to ensure a log from previous tests isn't matched against.
func (ks *KubernetesSuite) getLastLogBytePos(ctx context.Context, namespace, pod, container string) (int64, error) {
resp := ks.k8s.CoreV1().Pods(namespace).GetLogs(pod, &coreV1.PodLogOptions{Container: container}).Do(ctx)
logB, err := resp.Raw()
if err != nil {
return 0, fmt.Errorf("retrieving logs of pod %q in namespace %q failed: %w", pod, namespace, err)
}
return int64(len(logB)), nil
}
// getSensorPod is a convenience method to get details of the Sensor pod.
func (ks *KubernetesSuite) getSensorPod(ctx context.Context, namespace string) (*coreV1.Pod, error) {
ls := labels.SelectorFromSet(sensorPodLabels).String()
podList, err := ks.k8s.CoreV1().Pods(namespace).List(ctx, metaV1.ListOptions{LabelSelector: ls})
if err != nil {
return nil, fmt.Errorf("could not list pods matching %q in namespace %q: %w", sensorPodLabels, namespace, err)
}
if len(podList.Items) == 0 {
return nil, errors.New("empty list of pods does not satisfy the condition")
}
if len(podList.Items) > 1 {
return nil, errors.New("more than one sensor pod running")
}
return &podList.Items[0], nil
}
// waitUntilK8sDeploymentReady waits until k8s reports all pods for a deployment are ready or context done.
func (ks *KubernetesSuite) waitUntilK8sDeploymentReady(ctx context.Context, namespace string, deploymentName string) {
ticker := time.NewTicker(2 * time.Second)
defer ticker.Stop()
timer := time.NewTimer(waitTimeout)
ks.logf("Waiting for deployment %q in namespace %q to be ready", deploymentName, namespace)
for {
select {
case <-ctx.Done():
require.NoError(ks.T(), ctx.Err())
case <-ticker.C:
deploy, err := ks.k8s.AppsV1().Deployments(namespace).Get(ctx, deploymentName, metaV1.GetOptions{})
require.NoError(ks.T(), err, "getting deployment %q from namespace %q", deploymentName, namespace)
if deploy.GetGeneration() != deploy.Status.ObservedGeneration {
ks.logf("deployment %q in namespace %q NOT ready, generation %d, observed generation %d", deploymentName, namespace, deploy.GetGeneration(), deploy.Status.ObservedGeneration)
continue
}
You can’t perform that action at this time.