-
Notifications
You must be signed in to change notification settings - Fork 11
/
utils.go
executable file
·1346 lines (1195 loc) · 41.3 KB
/
utils.go
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
package utils
import (
"context"
"crypto/sha256"
"encoding/hex"
"encoding/json"
"fmt"
"io/ioutil"
"math"
"os"
"path/filepath"
"sort"
"strings"
"github.com/pquerna/ffjson/ffjson"
log "github.com/sirupsen/logrus"
"helm.sh/helm/v3/pkg/releaseutil"
appsv1 "k8s.io/api/apps/v1"
batchv1 "k8s.io/api/batch/v1"
batchv1beta1 "k8s.io/api/batch/v1beta1"
corev1 "k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/api/resource"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/runtime"
"k8s.io/apimachinery/pkg/runtime/schema"
"k8s.io/apimachinery/pkg/util/rand"
"k8s.io/apimachinery/pkg/util/uuid"
clientset "k8s.io/client-go/kubernetes"
externalclientset "k8s.io/client-go/kubernetes"
"k8s.io/client-go/kubernetes/scheme"
restclient "k8s.io/client-go/rest"
"k8s.io/client-go/tools/clientcmd"
resourcehelper "k8s.io/kubectl/pkg/util/resource"
api "k8s.io/kubernetes/pkg/apis/core"
apiv1 "k8s.io/kubernetes/pkg/apis/core/v1"
"k8s.io/kubernetes/pkg/apis/core/validation"
"k8s.io/kubernetes/pkg/controller/daemon"
"k8s.io/kubernetes/pkg/scheduler/framework"
schedulerutil "k8s.io/kubernetes/pkg/scheduler/util"
localcache "github.com/alibaba/open-local/pkg/scheduler/algorithm/cache"
localutils "github.com/alibaba/open-local/pkg/utils"
simontype "github.com/hkust-adsl/kubernetes-scheduler-simulator/pkg/type"
gpusharecache "github.com/hkust-adsl/kubernetes-scheduler-simulator/pkg/type/open-gpu-share/cache"
"github.com/hkust-adsl/kubernetes-scheduler-simulator/pkg/type/open-gpu-share/utils"
gpushareutils "github.com/hkust-adsl/kubernetes-scheduler-simulator/pkg/type/open-gpu-share/utils"
)
var nameDelimiter = "/"
func GeneratePodKey(pod *corev1.Pod) string {
return GeneratePodKeyByName(pod.Namespace, pod.Name)
}
func GeneratePodKeyByName(namespace, name string) string {
return namespace + nameDelimiter + name
}
// ParseFilePath converts recursively directory path to a slice of file paths
func ParseFilePath(path string) (filePaths []string, err error) {
fi, err := os.Stat(path)
if err != nil {
return nil, fmt.Errorf("failed to parse path(%s): %v ", path, err)
}
switch mode := fi.Mode(); {
case mode.IsDir():
files, err := ioutil.ReadDir(path)
if err != nil {
return nil, fmt.Errorf("failed to read directory(%s): %v ", path, err)
}
for _, f := range files {
p := filepath.Join(path, f.Name())
subFiles, err := ParseFilePath(p)
if err != nil {
return nil, err
}
filePaths = append(filePaths, subFiles...)
}
case mode.IsRegular():
filePaths = append(filePaths, path)
default:
return nil, fmt.Errorf("invalid path: %s ", path)
}
return filePaths, nil
}
// DecodeYamlContent captures the yml or yaml file, and decodes it
func DecodeYamlContent(yamlRes []byte) ([]runtime.Object, error) {
objects := make([]runtime.Object, 0)
yamls := releaseutil.SplitManifests(string(yamlRes))
for _, yaml := range yamls {
decode := scheme.Codecs.UniversalDeserializer().Decode
obj, _, err := decode([]byte(yaml), nil, nil)
if err != nil {
return nil, fmt.Errorf("failed to decode yaml content: \n%s\n%v", yaml, err)
}
objects = append(objects, obj)
}
return objects, nil
}
func ReadYamlFile(path string) []byte {
fileExtension := filepath.Ext(path)
if fileExtension != ".yaml" && fileExtension != ".yml" {
return nil
}
yamlFile, err := ioutil.ReadFile(path)
if err != nil {
log.Errorf("Error while read file %s: %s\n", path, err.Error())
os.Exit(1)
}
return yamlFile
}
func ReadJsonFile(path string) []byte {
pathExtension := filepath.Ext(path)
if pathExtension != ".json" {
return nil
}
jsonFile, err := ioutil.ReadFile(path)
if err != nil {
log.Errorf("Error while read file %s: %s\n", path, err.Error())
os.Exit(1)
}
return jsonFile
}
// GetYamlContentFromDirectory returns the yaml content and ignores other content
func GetYamlContentFromDirectory(dir string) ([]string, error) {
var ymlStr []string
filePaths, err := ParseFilePath(dir)
if err != nil {
return ymlStr, err
}
for _, filePath := range filePaths {
if yml := ReadYamlFile(filePath); yml != nil {
ymlStr = append(ymlStr, string(yml))
}
}
return ymlStr, nil
}
func MakeValidPodsByDeployment(deploy *appsv1.Deployment) ([]*corev1.Pod, error) {
deploy.UID = uuid.NewUUID()
return MakeValidPodsByReplicaSet(generateReplicaSetFromDeployment(deploy))
}
func MakeValidPodsByReplicaSet(rs *appsv1.ReplicaSet) ([]*corev1.Pod, error) {
var pods []*corev1.Pod
if rs.UID == "" {
rs.UID = uuid.NewUUID()
}
if rs.Spec.Replicas == nil {
var replica int32 = 1
rs.Spec.Replicas = &replica
}
for ordinal := 0; ordinal < int(*rs.Spec.Replicas); ordinal++ {
pod := &corev1.Pod{
ObjectMeta: SetObjectMetaFromObject(rs, true),
Spec: rs.Spec.Template.Spec,
}
validPod, err := MakeValidPod(pod)
if err != nil {
return nil, err
}
validPod = AddWorkloadInfoToPod(validPod, simontype.ReplicaSet, rs.Name, rs.Namespace)
pods = append(pods, validPod)
}
return pods, nil
}
func MakeValidPodsByReplicationController(rc *corev1.ReplicationController) ([]*corev1.Pod, error) {
var pods []*corev1.Pod
rc.UID = uuid.NewUUID()
if rc.Spec.Replicas == nil {
var replica int32 = 1
rc.Spec.Replicas = &replica
}
for ordinal := 0; ordinal < int(*rc.Spec.Replicas); ordinal++ {
pod := &corev1.Pod{
ObjectMeta: SetObjectMetaFromObject(rc, true),
Spec: rc.Spec.Template.Spec,
}
validPod, err := MakeValidPod(pod)
if err != nil {
return nil, err
}
validPod = AddWorkloadInfoToPod(validPod, simontype.ReplicationController, rc.Name, rc.Namespace)
pods = append(pods, validPod)
}
return pods, nil
}
func generateReplicaSetFromDeployment(deploy *appsv1.Deployment) *appsv1.ReplicaSet {
return &appsv1.ReplicaSet{
TypeMeta: metav1.TypeMeta{APIVersion: appsv1.SchemeGroupVersion.String(), Kind: simontype.ReplicaSet},
ObjectMeta: SetObjectMetaFromObject(deploy, false),
Spec: appsv1.ReplicaSetSpec{
Selector: deploy.Spec.Selector,
Replicas: deploy.Spec.Replicas,
Template: deploy.Spec.Template,
},
}
}
func MakeValidPodByCronJob(cronjob *batchv1beta1.CronJob) ([]*corev1.Pod, error) {
cronjob.UID = uuid.NewUUID()
return MakeValidPodByJob(generateJobFromCronJob(cronjob))
}
func MakeValidPodByJob(job *batchv1.Job) ([]*corev1.Pod, error) {
var pods []*corev1.Pod
if job.UID == "" {
job.UID = uuid.NewUUID()
}
job.UID = uuid.NewUUID()
if job.Spec.Completions == nil {
var completions int32 = 1
job.Spec.Completions = &completions
}
for ordinal := 0; ordinal < int(*job.Spec.Completions); ordinal++ {
pod := &corev1.Pod{
ObjectMeta: SetObjectMetaFromObject(job, true),
Spec: job.Spec.Template.Spec,
}
validPod, err := MakeValidPod(pod)
if err != nil {
return nil, err
}
validPod = AddWorkloadInfoToPod(validPod, simontype.Job, job.Name, job.Namespace)
pods = append(pods, validPod)
}
return pods, nil
}
func generateJobFromCronJob(cronJob *batchv1beta1.CronJob) *batchv1.Job {
annotations := make(map[string]string)
annotations["cronjob.kubernetes.io/instantiate"] = "manual"
for k, v := range cronJob.Spec.JobTemplate.Annotations {
annotations[k] = v
}
return &batchv1.Job{
TypeMeta: metav1.TypeMeta{APIVersion: batchv1.SchemeGroupVersion.String(), Kind: simontype.Job},
ObjectMeta: SetObjectMetaFromObject(cronJob, false),
Spec: cronJob.Spec.JobTemplate.Spec,
}
}
func MakeValidPodsByStatefulSet(ss *appsv1.StatefulSet) ([]*corev1.Pod, error) {
var pods []*corev1.Pod
ss.UID = uuid.NewUUID()
if ss.Spec.Replicas == nil {
var replica int32 = 1
ss.Spec.Replicas = &replica
}
for ordinal := 0; ordinal < int(*ss.Spec.Replicas); ordinal++ {
pod := &corev1.Pod{
ObjectMeta: SetObjectMetaFromObject(ss, true),
Spec: ss.Spec.Template.Spec,
}
validPod, err := MakeValidPod(pod)
if err != nil {
return nil, err
}
validPod.Name = fmt.Sprintf("%s-%d", ss.Name, ordinal)
validPod = AddWorkloadInfoToPod(validPod, simontype.StatefulSet, ss.Name, ss.Namespace)
pods = append(pods, validPod)
}
if err := SetStorageAnnotationOnPods(pods, ss.Spec.VolumeClaimTemplates, ss.Name); err != nil {
return nil, err
}
return pods, nil
}
func SetStorageAnnotationOnPods(pods []*corev1.Pod, volumeClaimTemplates []corev1.PersistentVolumeClaim, stsName string) error {
var volumes VolumeRequest
volumes.Volumes = make([]Volume, 0)
for _, pvc := range volumeClaimTemplates {
if pvc.Spec.StorageClassName != nil {
if *pvc.Spec.StorageClassName == OpenLocalSCNameLVM || *pvc.Spec.StorageClassName == YodaSCNameLVM {
volume := Volume{
Size: localutils.GetPVCRequested(&pvc),
Kind: "LVM",
StorageClassName: *pvc.Spec.StorageClassName,
}
volumes.Volumes = append(volumes.Volumes, volume)
} else if *pvc.Spec.StorageClassName == OpenLocalSCNameDeviceSSD || *pvc.Spec.StorageClassName == OpenLocalSCNameMountPointSSD || *pvc.Spec.StorageClassName == YodaSCNameMountPointSSD || *pvc.Spec.StorageClassName == YodaSCNameDeviceSSD {
volume := Volume{
Size: localutils.GetPVCRequested(&pvc),
Kind: "SSD",
StorageClassName: *pvc.Spec.StorageClassName,
}
volumes.Volumes = append(volumes.Volumes, volume)
} else if *pvc.Spec.StorageClassName == OpenLocalSCNameDeviceHDD || *pvc.Spec.StorageClassName == OpenLocalSCNameMountPointHDD || *pvc.Spec.StorageClassName == YodaSCNameMountPointHDD || *pvc.Spec.StorageClassName == YodaSCNameDeviceHDD {
volume := Volume{
Size: localutils.GetPVCRequested(&pvc),
Kind: "HDD",
StorageClassName: *pvc.Spec.StorageClassName,
}
volumes.Volumes = append(volumes.Volumes, volume)
} else {
log.Errorf("unsupported storage class: %s", *pvc.Spec.StorageClassName)
}
} else {
log.Errorf("empty storageClassName in volumeTemplate of statefulset %s is not supported", stsName)
}
}
for _, pod := range pods {
b, err := json.Marshal(volumes)
if err != nil {
return err
}
metav1.SetMetaDataAnnotation(&pod.ObjectMeta, simontype.AnnoPodLocalStorage, string(b))
}
return nil
}
func SetObjectMetaFromObject(owner metav1.Object, genPod bool) metav1.ObjectMeta {
var controllerKind schema.GroupVersionKind
switch owner.(type) {
case *appsv1.Deployment:
controllerKind = appsv1.SchemeGroupVersion.WithKind(simontype.Deployment)
case *appsv1.ReplicaSet:
controllerKind = appsv1.SchemeGroupVersion.WithKind(simontype.ReplicaSet)
case *appsv1.StatefulSet:
controllerKind = appsv1.SchemeGroupVersion.WithKind(simontype.StatefulSet)
case *appsv1.DaemonSet:
controllerKind = appsv1.SchemeGroupVersion.WithKind(simontype.DaemonSet)
case *corev1.ReplicationController:
controllerKind = corev1.SchemeGroupVersion.WithKind(simontype.ReplicationController)
case *batchv1beta1.CronJob:
controllerKind = batchv1beta1.SchemeGroupVersion.WithKind(simontype.CronJob)
case *batchv1.Job:
controllerKind = batchv1.SchemeGroupVersion.WithKind(simontype.Job)
}
return metav1.ObjectMeta{
Name: owner.GetName() + simontype.SeparateSymbol + GetSHA256HashCode([]byte(rand.String(10)), GetObjectHashCodeDigit(genPod)),
Namespace: owner.GetNamespace(),
UID: uuid.NewUUID(),
Annotations: owner.GetAnnotations(),
GenerateName: owner.GetName(),
Labels: owner.GetLabels(),
OwnerReferences: []metav1.OwnerReference{
*metav1.NewControllerRef(owner, controllerKind),
},
}
}
func GetObjectHashCodeDigit(isPod bool) int {
if isPod {
return simontype.PodHashCodeDigit
}
return simontype.WorkLoadHashCodeDigit
}
func MakeValidPodsByDaemonset(ds *appsv1.DaemonSet, nodes []*corev1.Node) ([]*corev1.Pod, error) {
var pods []*corev1.Pod
ds.UID = uuid.NewUUID()
for _, node := range nodes {
pod, err := NewDaemonPod(ds, node.Name)
if err != nil {
return nil, err
}
shouldRun := NodeShouldRunPod(node, pod)
if shouldRun {
pods = append(pods, pod)
}
}
return pods, nil
}
func NewDaemonPod(ds *appsv1.DaemonSet, nodeName string) (*corev1.Pod, error) {
pod := &corev1.Pod{
ObjectMeta: SetObjectMetaFromObject(ds, true),
Spec: ds.Spec.Template.Spec,
}
pod.Spec.Affinity = SetDaemonSetPodNodeNameByNodeAffinity(pod.Spec.Affinity, nodeName)
validPod, err := MakeValidPod(pod)
if err != nil {
return nil, err
}
validPod = AddWorkloadInfoToPod(validPod, simontype.DaemonSet, ds.Name, ds.Namespace)
return validPod, nil
}
// NodeShouldRunPod determines whether a node should run a pod according to scheduling rules
func NodeShouldRunPod(node *corev1.Node, pod *corev1.Pod) bool {
taints := node.Spec.Taints
fitsNodeName, fitsNodeAffinity, fitsTaints := daemon.Predicates(pod, node, taints)
if !fitsNodeName || !fitsNodeAffinity || !fitsTaints {
return false
}
return true
}
func MakeValidPodByPod(pod *corev1.Pod) (*corev1.Pod, error) {
pod.UID = uuid.NewUUID()
newPod, err := MakeValidPod(pod)
if err != nil {
return nil, err
}
return newPod, nil
}
// MakeValidPod make pod valid, so we can handle it
func MakeValidPod(oldPod *corev1.Pod) (*corev1.Pod, error) {
newPod := oldPod.DeepCopy()
if newPod.Labels == nil {
newPod.Labels = make(map[string]string)
}
// ObjectMeta
if newPod.ObjectMeta.Namespace == "" {
newPod.ObjectMeta.Namespace = corev1.NamespaceDefault
}
if newPod.ObjectMeta.Annotations == nil {
newPod.ObjectMeta.Annotations = map[string]string{}
}
// Spec
if newPod.Spec.DNSPolicy == "" {
newPod.Spec.DNSPolicy = corev1.DNSClusterFirst
}
if newPod.Spec.RestartPolicy == "" {
newPod.Spec.RestartPolicy = corev1.RestartPolicyAlways
}
if newPod.Spec.SchedulerName == "" {
newPod.Spec.SchedulerName = simontype.DefaultSchedulerName
}
newPod.Spec.ImagePullSecrets = nil
if newPod.Spec.InitContainers != nil {
for i := range newPod.Spec.InitContainers {
if newPod.Spec.InitContainers[i].TerminationMessagePolicy == "" {
newPod.Spec.InitContainers[i].TerminationMessagePolicy = corev1.TerminationMessageFallbackToLogsOnError
}
if newPod.Spec.InitContainers[i].ImagePullPolicy == "" {
newPod.Spec.InitContainers[i].ImagePullPolicy = corev1.PullIfNotPresent
}
if newPod.Spec.InitContainers[i].SecurityContext != nil && newPod.Spec.InitContainers[i].SecurityContext.Privileged != nil {
var priv = false
newPod.Spec.InitContainers[i].SecurityContext.Privileged = &priv
}
newPod.Spec.InitContainers[i].VolumeMounts = nil
newPod.Spec.InitContainers[i].Env = nil
}
}
if newPod.Spec.Containers != nil {
for i := range newPod.Spec.Containers {
if newPod.Spec.Containers[i].TerminationMessagePolicy == "" {
newPod.Spec.Containers[i].TerminationMessagePolicy = corev1.TerminationMessageFallbackToLogsOnError
}
if newPod.Spec.Containers[i].ImagePullPolicy == "" {
newPod.Spec.Containers[i].ImagePullPolicy = corev1.PullIfNotPresent
}
if newPod.Spec.Containers[i].SecurityContext != nil && newPod.Spec.Containers[i].SecurityContext.Privileged != nil {
var priv = false
newPod.Spec.Containers[i].SecurityContext.Privileged = &priv
}
newPod.Spec.Containers[i].VolumeMounts = nil
newPod.Spec.Containers[i].Env = nil
// Probe
newPod.Spec.Containers[i].LivenessProbe = nil
newPod.Spec.Containers[i].ReadinessProbe = nil
newPod.Spec.Containers[i].StartupProbe = nil
}
}
// handle volume
// notice: open-local volume should not used in deployment, cause it is not nas storage.
if newPod.Spec.Volumes != nil {
for i := range newPod.Spec.Volumes {
if newPod.Spec.Volumes[i].PersistentVolumeClaim != nil {
newPod.Spec.Volumes[i].HostPath = new(corev1.HostPathVolumeSource)
newPod.Spec.Volumes[i].HostPath.Path = "/tmp"
newPod.Spec.Volumes[i].PersistentVolumeClaim = nil
}
}
}
// todo: handle pvc
if err := ValidatePod(newPod); err != nil {
return nil, err
}
return newPod, nil
}
// AddWorkloadInfoToPod add annotation in pod for simulating later
func AddWorkloadInfoToPod(pod *corev1.Pod, kind string, name string, namespace string) *corev1.Pod {
pod.ObjectMeta.Annotations[simontype.AnnoWorkloadKind] = kind
pod.ObjectMeta.Annotations[simontype.AnnoWorkloadName] = name
pod.ObjectMeta.Annotations[simontype.AnnoWorkloadNamespace] = namespace
return pod
}
func MakeValidNodeByNode(node *corev1.Node, nodename string) (*corev1.Node, error) {
node.ObjectMeta.Name = nodename
node.ObjectMeta.UID = uuid.NewUUID()
if node.ObjectMeta.Labels == nil {
node.ObjectMeta.Labels = map[string]string{}
}
node.ObjectMeta.Labels[corev1.LabelHostname] = nodename
if node.ObjectMeta.Annotations == nil {
node.ObjectMeta.Annotations = map[string]string{}
}
if err := ValidateNode(node); err != nil {
return nil, err
}
return node, nil
}
// ValidatePod check if pod is valid
func ValidatePod(pod *corev1.Pod) error {
internalPod := &api.Pod{}
if err := apiv1.Convert_v1_Pod_To_core_Pod(pod, internalPod, nil); err != nil {
return fmt.Errorf("unable to convert to internal version: %#v ", err)
}
if errs := validation.ValidatePodCreate(internalPod, validation.PodValidationOptions{}); len(errs) > 0 {
var errStrs []string
for _, err := range errs {
errStrs = append(errStrs, fmt.Sprintf("%v", err))
}
return fmt.Errorf("invalid pod: %#v ", strings.Join(errStrs, "\n"))
}
return nil
}
func GetSHA256HashCode(message []byte, num int) string {
hash := sha256.New()
hash.Write(message)
hashCode := hex.EncodeToString(hash.Sum(nil))
return hashCode[:num]
}
type NodeStorage struct {
VGs []localcache.SharedResource `json:"vgs"`
Devices []localcache.ExclusiveResource `json:"devices"`
}
type Volume struct {
Size int64 `json:"size,string"`
// Kind 可以是 LVM 或 HDD 或 SSD
// HDD 和 SSD 均指代独占盘
Kind string `json:"kind"`
StorageClassName string `json:"scName"`
}
type VolumeRequest struct {
Volumes []Volume `json:"volumes"`
}
func GetNodeStorage(node *corev1.Node) (*NodeStorage, error) {
nodeStorageStr, exist := node.Annotations[simontype.AnnoNodeLocalStorage]
if !exist {
return nil, nil
}
nodeStorage := new(NodeStorage)
if err := ffjson.Unmarshal([]byte(nodeStorageStr), nodeStorage); err != nil {
return nil, fmt.Errorf("failed to unmarshal info of node %s: %s ", node.Name, err.Error())
}
return nodeStorage, nil
}
func GetNodeCache(node *corev1.Node) (*localcache.NodeCache, error) {
nodeStorage, err := GetNodeStorage(node)
if err != nil {
return nil, err
} else if nodeStorage == nil {
return nil, nil
}
nc := localcache.NewNodeCache(node.Name)
var vgRes map[localcache.ResourceName]localcache.SharedResource = make(map[localcache.ResourceName]localcache.SharedResource)
for _, vg := range nodeStorage.VGs {
vgRes[localcache.ResourceName(vg.Name)] = vg
}
nc.VGs = vgRes
var deviceRes map[localcache.ResourceName]localcache.ExclusiveResource = make(map[localcache.ResourceName]localcache.ExclusiveResource)
for _, device := range nodeStorage.Devices {
deviceRes[localcache.ResourceName(device.Device)] = device
}
nc.Devices = deviceRes
return nc, nil
}
func GetPodStorage(pod *corev1.Pod) *VolumeRequest {
podStorageStr, exist := pod.Annotations[simontype.AnnoPodLocalStorage]
if !exist {
return nil
}
podStorage := new(VolumeRequest)
if err := ffjson.Unmarshal([]byte(podStorageStr), &podStorage); err != nil {
log.Errorf("unmarshal volume info of pod %s/%s failed: %s", pod.Namespace, pod.Name, err.Error())
return nil
}
return podStorage
}
func GetPodLocalPVCs(pod *corev1.Pod) ([]*corev1.PersistentVolumeClaim, []*corev1.PersistentVolumeClaim) {
podStorage := GetPodStorage(pod)
if podStorage == nil {
return nil, nil
}
lvmPVCs := make([]*corev1.PersistentVolumeClaim, 0)
devicePVCs := make([]*corev1.PersistentVolumeClaim, 0)
for i, volume := range podStorage.Volumes {
scName := ""
if volume.Kind == "LVM" || volume.Kind == "HDD" || volume.Kind == "SSD" {
scName = volume.StorageClassName
} else {
log.Errorf("unsupported volume kind: %s", volume.Kind)
continue
}
volumeQuantity := resource.NewQuantity(volume.Size, resource.BinarySI)
pvc := &corev1.PersistentVolumeClaim{
ObjectMeta: metav1.ObjectMeta{
Name: fmt.Sprintf("pvc-%s-%d", pod.Name, i),
Namespace: pod.Namespace,
},
Spec: corev1.PersistentVolumeClaimSpec{
AccessModes: []corev1.PersistentVolumeAccessMode{corev1.ReadWriteOnce},
StorageClassName: &scName,
Resources: corev1.ResourceRequirements{
Requests: corev1.ResourceList{
corev1.ResourceName(corev1.ResourceStorage): *volumeQuantity,
},
},
},
Status: corev1.PersistentVolumeClaimStatus{
Phase: corev1.ClaimPending,
},
}
if scName == OpenLocalSCNameLVM || scName == YodaSCNameLVM {
lvmPVCs = append(lvmPVCs, pvc)
} else {
devicePVCs = append(devicePVCs, pvc)
}
}
return lvmPVCs, devicePVCs
}
// ValidateNode check if node is valid
func ValidateNode(node *corev1.Node) error {
internalNode := &api.Node{}
if err := apiv1.Convert_v1_Node_To_core_Node(node, internalNode, nil); err != nil {
return fmt.Errorf("unable to convert to internal version: %#v ", err)
}
if errs := validation.ValidateNode(internalNode); len(errs) > 0 {
var errStrs []string
for _, err := range errs {
errStrs = append(errStrs, fmt.Sprintf("%v", err))
}
return fmt.Errorf("invalid node: %#v ", strings.Join(errStrs, "\n"))
}
return nil
}
func GetPodsTotalRequestsAndLimitsByNodeName(pods []*corev1.Pod, nodeName string) (map[corev1.ResourceName]resource.Quantity, map[corev1.ResourceName]resource.Quantity) {
reqs, limits := make(map[corev1.ResourceName]resource.Quantity), make(map[corev1.ResourceName]resource.Quantity)
for _, pod := range pods {
if pod.Spec.NodeName != nodeName {
continue
}
podReqs, podLimits := resourcehelper.PodRequestsAndLimits(pod)
for podReqName, podReqValue := range podReqs {
if value, ok := reqs[podReqName]; !ok {
reqs[podReqName] = podReqValue.DeepCopy()
} else {
value.Add(podReqValue)
reqs[podReqName] = value
}
}
for podLimitName, podLimitValue := range podLimits {
if value, ok := limits[podLimitName]; !ok {
limits[podLimitName] = podLimitValue.DeepCopy()
} else {
value.Add(podLimitValue)
limits[podLimitName] = value
}
}
}
return reqs, limits
}
func AdjustWorkloads(workloads map[string][]string) {
if workloads == nil {
return
}
for name, nodes := range workloads {
workloads[name] = AdjustNodesOrder(nodes)
}
}
func AdjustNodesOrder(nodes []string) []string {
queue := NewNodeQueue(nodes)
sort.Sort(queue)
return queue.nodes
}
type NodeQueue struct {
nodes []string
}
func NewNodeQueue(nodes []string) *NodeQueue {
return &NodeQueue{
nodes: nodes,
}
}
func (queue *NodeQueue) Len() int { return len(queue.nodes) }
func (queue *NodeQueue) Swap(i, j int) {
queue.nodes[i], queue.nodes[j] = queue.nodes[j], queue.nodes[i]
}
func (queue *NodeQueue) Less(i, j int) bool {
if strings.Contains(queue.nodes[i], simontype.NewNodeNamePrefix) && strings.Contains(queue.nodes[j], simontype.NewNodeNamePrefix) {
return queue.nodes[i] < queue.nodes[j]
}
if !strings.Contains(queue.nodes[i], simontype.NewNodeNamePrefix) && !strings.Contains(queue.nodes[j], simontype.NewNodeNamePrefix) {
return queue.nodes[i] < queue.nodes[j]
}
if !strings.Contains(queue.nodes[i], simontype.NewNodeNamePrefix) {
return true
}
if !strings.Contains(queue.nodes[j], simontype.NewNodeNamePrefix) {
return false
}
return true
}
// MultiplyMilliQuant scales quantity by factor
func MultiplyMilliQuant(quant resource.Quantity, factor float64) resource.Quantity {
milliValue := quant.MilliValue()
newMilliValue := int64(float64(milliValue) * factor)
newQuant := resource.NewMilliQuantity(newMilliValue, quant.Format)
return *newQuant
}
// MultiplyQuant scales quantity by factor
func MultiplyQuant(quant resource.Quantity, factor float64) resource.Quantity {
value := quant.Value()
newValue := int64(float64(value) * factor)
newQuant := resource.NewQuantity(newValue, quant.Format)
return *newQuant
}
func GetNodeAllocatable(node *corev1.Node) (resource.Quantity, resource.Quantity) {
nodeAllocatable := node.Status.Allocatable.DeepCopy()
return *nodeAllocatable.Cpu(), *nodeAllocatable.Memory()
}
func GetNodeAllocatableCpuGpu(node *corev1.Node) []float64 {
milliCpu := node.Status.Allocatable.Cpu().MilliValue()
milliGpu := gpushareutils.GetGpuCountOfNode(node) * gpushareutils.MILLI
return []float64{float64(milliCpu), float64(milliGpu)}
}
func MeetResourceRequests(node *corev1.Node, pod *corev1.Pod, daemonSets []*appsv1.DaemonSet) (bool, error) {
// CPU and Memory
totalResource := map[corev1.ResourceName]*resource.Quantity{
corev1.ResourceCPU: resource.NewQuantity(0, resource.DecimalSI),
corev1.ResourceMemory: resource.NewQuantity(0, resource.DecimalSI),
}
for _, item := range daemonSets {
newItem := item
daemonPod, err := NewDaemonPod(newItem, simontype.NewNodeNamePrefix)
if err != nil {
return false, err
}
if NodeShouldRunPod(node, daemonPod) {
for _, container := range daemonPod.Spec.Containers {
totalResource[corev1.ResourceCPU].Add(*container.Resources.Requests.Cpu())
totalResource[corev1.ResourceMemory].Add(*container.Resources.Requests.Memory())
}
}
}
for _, container := range pod.Spec.Containers {
totalResource[corev1.ResourceCPU].Add(*container.Resources.Requests.Cpu())
totalResource[corev1.ResourceMemory].Add(*container.Resources.Requests.Memory())
}
if totalResource[corev1.ResourceCPU].Cmp(*node.Status.Allocatable.Cpu()) == 1 ||
totalResource[corev1.ResourceMemory].Cmp(*node.Status.Allocatable.Memory()) == 1 {
return false, nil
}
// Local Storage
nodeStorage, err := GetNodeStorage(node)
if err != nil {
return false, err
} else if nodeStorage == nil {
return true, nil
}
var nodeVGMax int64 = 0
for _, vg := range nodeStorage.VGs {
if vg.Capacity > int64(nodeVGMax) {
nodeVGMax = vg.Capacity
}
}
lvmPVCs, _ := GetPodLocalPVCs(pod)
var pvcSum int64 = 0
for _, pvc := range lvmPVCs {
pvcSum += localutils.GetPVCRequested(pvc)
}
return pvcSum <= nodeVGMax, nil
}
func CreateKubeClient(kubeconfig string) (*clientset.Clientset, error) {
if len(kubeconfig) == 0 {
return nil, nil
}
var err error
var cfg *restclient.Config
master, err := GetMasterFromKubeConfig(kubeconfig)
if err != nil {
return nil, fmt.Errorf("failed to parse kubeclient file: %v ", err)
}
cfg, err = clientcmd.BuildConfigFromFlags(master, kubeconfig)
if err != nil {
return nil, fmt.Errorf("unable to build config: %v ", err)
}
kubeClient, err := clientset.NewForConfig(cfg)
if err != nil {
return nil, err
}
return kubeClient, nil
}
func GetMasterFromKubeConfig(filename string) (string, error) {
config, err := clientcmd.LoadFromFile(filename)
if err != nil {
return "", fmt.Errorf("can not load kubeconfig file: %v", err)
}
varContext, ok := config.Contexts[config.CurrentContext]
if !ok {
return "", fmt.Errorf("failed to get master address from kubeconfig")
}
if val, ok := config.Clusters[varContext.Cluster]; ok {
return val.Server, nil
}
return "", fmt.Errorf("failed to get master address from kubeconfig")
}
func SetDaemonSetPodNodeNameByNodeAffinity(affinity *corev1.Affinity, nodename string) *corev1.Affinity {
nodeSelReq := corev1.NodeSelectorRequirement{
Key: api.ObjectNameField,
Operator: corev1.NodeSelectorOpIn,
Values: []string{nodename},
}
nodeSelector := &corev1.NodeSelector{
NodeSelectorTerms: []corev1.NodeSelectorTerm{
{
MatchFields: []corev1.NodeSelectorRequirement{nodeSelReq},
},
},
}
if affinity == nil {
return &corev1.Affinity{
NodeAffinity: &corev1.NodeAffinity{
RequiredDuringSchedulingIgnoredDuringExecution: nodeSelector,
},
}
}
if affinity.NodeAffinity == nil {
affinity.NodeAffinity = &corev1.NodeAffinity{
RequiredDuringSchedulingIgnoredDuringExecution: nodeSelector,
}
return affinity
}
if affinity.NodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution == nil {
affinity.NodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution = nodeSelector
return affinity
}
// Replace node selector with the new one.
nodeSelectorTerms := affinity.NodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution.NodeSelectorTerms
for i, item := range nodeSelectorTerms {
item.MatchFields = []corev1.NodeSelectorRequirement{nodeSelReq}
nodeSelectorTerms[i] = item
}
affinity.NodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution.NodeSelectorTerms = nodeSelectorTerms
return affinity
}
func GetGpuNodeInfoFromAnnotation(node *corev1.Node) (*gpusharecache.GpuNodeInfoStr, error) {
gpuNodeInfoString, exist := node.Annotations[simontype.AnnoNodeGpuShare]
if !exist {
return nil, nil
}
gpuNodeInfoStr := new(gpusharecache.GpuNodeInfoStr)
if err := ffjson.Unmarshal([]byte(gpuNodeInfoString), gpuNodeInfoStr); err != nil {
return nil, fmt.Errorf("failed to unmarshal GPU info of node %s: %s ", node.Name, err.Error())
}
return gpuNodeInfoStr, nil
}
func GetAllocatablePodList(clientset externalclientset.Interface) []corev1.Pod {
podList, _ := clientset.CoreV1().Pods(corev1.NamespaceAll).List(context.Background(), metav1.ListOptions{})
return podList.Items
}
func IsNodeAccessibleToPod(nodeRes simontype.NodeResource, podRes simontype.PodResource) bool {
pt := podRes.GpuType
nt := nodeRes.GpuType
return IsNodeAccessibleToPodByType(nt, pt)
}
func IsNodeAccessibleToPodByType(nodeGpuType string, podGpuType string) bool {
if len(podGpuType) == 0 {
return true
}
if len(nodeGpuType) == 0 {
return false // i.e., CPU node
}
//pm, ok := gpushareutils.MapGpuTypeMemoryMiB[podGpuType]
//if !ok {
// log.Errorf("Pod GPU Type: %s not in Map", podGpuType)
// return false
//}
//
//nm, ok := gpushareutils.MapGpuTypeMemoryMiB[nodeGpuType]
//if !ok {
// log.Errorf("Node GPU Type: %s not in Map", nodeGpuType)
// return false
//}
//
//if pm > nm {
// return false
//}
podGpuTypeList := strings.Split(podGpuType, "|")
cnt := 0
for _, gpuType := range podGpuTypeList {
if len(gpuType) == 0 {
continue
}
cnt++
if gpuType == nodeGpuType {
return true
}
}
if cnt > 0 { // pod requests at least one specific GPU type but node doesn't match
return false
} else { // pod actually doesn't request any specific GPU type
return true
}
}
func GetPodResource(pod *corev1.Pod) simontype.PodResource {
gpuNumber := gpushareutils.GetGpuCountFromPodAnnotation(pod)
gpuMilli := gpushareutils.GetGpuMilliFromPodAnnotation(pod)
gpuType := gpushareutils.GetGpuModelFromPodAnnotation(pod)