/
scheduling_queue.go
977 lines (873 loc) · 34.4 KB
/
scheduling_queue.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
/*
Copyright 2017 The Kubernetes Authors.
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.
*/
// This file contains structures that implement scheduling queue types.
// Scheduling queues hold pods waiting to be scheduled. This file implements a/
// priority queue which has two sub queues. One sub-queue holds pods that are
// being considered for scheduling. This is called activeQ. Another queue holds
// pods that are already tried and are determined to be unschedulable. The latter
// is called unschedulableQ.
package queue
import (
"fmt"
"reflect"
"sync"
"time"
"k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/labels"
"k8s.io/apimachinery/pkg/types"
"k8s.io/apimachinery/pkg/util/sets"
"k8s.io/apimachinery/pkg/util/wait"
utilfeature "k8s.io/apiserver/pkg/util/feature"
"k8s.io/client-go/informers"
listersv1 "k8s.io/client-go/listers/core/v1"
"k8s.io/client-go/tools/cache"
"k8s.io/klog/v2"
"k8s.io/kubernetes/pkg/features"
"k8s.io/kubernetes/pkg/scheduler/framework"
"k8s.io/kubernetes/pkg/scheduler/framework/plugins/interpodaffinity"
"k8s.io/kubernetes/pkg/scheduler/internal/heap"
"k8s.io/kubernetes/pkg/scheduler/metrics"
"k8s.io/kubernetes/pkg/scheduler/util"
)
const (
// If the pod stays in unschedulableQ longer than the unschedulableQTimeInterval,
// the pod will be moved from unschedulableQ to activeQ.
unschedulableQTimeInterval = 60 * time.Second
queueClosed = "scheduling queue is closed"
)
const (
// DefaultPodInitialBackoffDuration is the default value for the initial backoff duration
// for unschedulable pods. To change the default podInitialBackoffDurationSeconds used by the
// scheduler, update the ComponentConfig value in defaults.go
DefaultPodInitialBackoffDuration time.Duration = 1 * time.Second
// DefaultPodMaxBackoffDuration is the default value for the max backoff duration
// for unschedulable pods. To change the default podMaxBackoffDurationSeconds used by the
// scheduler, update the ComponentConfig value in defaults.go
DefaultPodMaxBackoffDuration time.Duration = 10 * time.Second
)
// PreEnqueueCheck is a function type. It's used to build functions that
// run against a Pod and the caller can choose to enqueue or skip the Pod
// by the checking result.
type PreEnqueueCheck func(pod *v1.Pod) bool
// SchedulingQueue is an interface for a queue to store pods waiting to be scheduled.
// The interface follows a pattern similar to cache.FIFO and cache.Heap and
// makes it easy to use those data structures as a SchedulingQueue.
type SchedulingQueue interface {
framework.PodNominator
Add(pod *v1.Pod) error
// Activate moves the given pods to activeQ iff they're in unschedulableQ or backoffQ.
// The passed-in pods are originally compiled from plugins that want to activate Pods,
// by injecting the pods through a reserved CycleState struct (PodsToActivate).
Activate(pods map[string]*v1.Pod)
// AddUnschedulableIfNotPresent adds an unschedulable pod back to scheduling queue.
// The podSchedulingCycle represents the current scheduling cycle number which can be
// returned by calling SchedulingCycle().
AddUnschedulableIfNotPresent(pod *framework.QueuedPodInfo, podSchedulingCycle int64) error
// SchedulingCycle returns the current number of scheduling cycle which is
// cached by scheduling queue. Normally, incrementing this number whenever
// a pod is popped (e.g. called Pop()) is enough.
SchedulingCycle() int64
// Pop removes the head of the queue and returns it. It blocks if the
// queue is empty and waits until a new item is added to the queue.
Pop() (*framework.QueuedPodInfo, error)
Update(oldPod, newPod *v1.Pod) error
Delete(pod *v1.Pod) error
MoveAllToActiveOrBackoffQueue(event framework.ClusterEvent, preCheck PreEnqueueCheck)
AssignedPodAdded(pod *v1.Pod)
AssignedPodUpdated(pod *v1.Pod)
PendingPods() []*v1.Pod
// Close closes the SchedulingQueue so that the goroutine which is
// waiting to pop items can exit gracefully.
Close()
// NumUnschedulablePods returns the number of unschedulable pods exist in the SchedulingQueue.
NumUnschedulablePods() int
// Run starts the goroutines managing the queue.
Run()
}
// NewSchedulingQueue initializes a priority queue as a new scheduling queue.
func NewSchedulingQueue(
lessFn framework.LessFunc,
informerFactory informers.SharedInformerFactory,
opts ...Option) SchedulingQueue {
return NewPriorityQueue(lessFn, informerFactory, opts...)
}
// NominatedNodeName returns nominated node name of a Pod.
func NominatedNodeName(pod *v1.Pod) string {
return pod.Status.NominatedNodeName
}
// PriorityQueue implements a scheduling queue.
// The head of PriorityQueue is the highest priority pending pod. This structure
// has three sub queues. One sub-queue holds pods that are being considered for
// scheduling. This is called activeQ and is a Heap. Another queue holds
// pods that are already tried and are determined to be unschedulable. The latter
// is called unschedulableQ. The third queue holds pods that are moved from
// unschedulable queues and will be moved to active queue when backoff are completed.
type PriorityQueue struct {
// PodNominator abstracts the operations to maintain nominated Pods.
framework.PodNominator
stop chan struct{}
clock util.Clock
// pod initial backoff duration.
podInitialBackoffDuration time.Duration
// pod maximum backoff duration.
podMaxBackoffDuration time.Duration
lock sync.RWMutex
cond sync.Cond
// activeQ is heap structure that scheduler actively looks at to find pods to
// schedule. Head of heap is the highest priority pod.
activeQ *heap.Heap
// podBackoffQ is a heap ordered by backoff expiry. Pods which have completed backoff
// are popped from this heap before the scheduler looks at activeQ
podBackoffQ *heap.Heap
// unschedulableQ holds pods that have been tried and determined unschedulable.
unschedulableQ *UnschedulablePodsMap
// schedulingCycle represents sequence number of scheduling cycle and is incremented
// when a pod is popped.
schedulingCycle int64
// moveRequestCycle caches the sequence number of scheduling cycle when we
// received a move request. Unschedulable pods in and before this scheduling
// cycle will be put back to activeQueue if we were trying to schedule them
// when we received move request.
moveRequestCycle int64
clusterEventMap map[framework.ClusterEvent]sets.String
// closed indicates that the queue is closed.
// It is mainly used to let Pop() exit its control loop while waiting for an item.
closed bool
nsLister listersv1.NamespaceLister
}
type priorityQueueOptions struct {
clock util.Clock
podInitialBackoffDuration time.Duration
podMaxBackoffDuration time.Duration
podNominator framework.PodNominator
clusterEventMap map[framework.ClusterEvent]sets.String
}
// Option configures a PriorityQueue
type Option func(*priorityQueueOptions)
// WithClock sets clock for PriorityQueue, the default clock is util.RealClock.
func WithClock(clock util.Clock) Option {
return func(o *priorityQueueOptions) {
o.clock = clock
}
}
// WithPodInitialBackoffDuration sets pod initial backoff duration for PriorityQueue.
func WithPodInitialBackoffDuration(duration time.Duration) Option {
return func(o *priorityQueueOptions) {
o.podInitialBackoffDuration = duration
}
}
// WithPodMaxBackoffDuration sets pod max backoff duration for PriorityQueue.
func WithPodMaxBackoffDuration(duration time.Duration) Option {
return func(o *priorityQueueOptions) {
o.podMaxBackoffDuration = duration
}
}
// WithPodNominator sets pod nominator for PriorityQueue.
func WithPodNominator(pn framework.PodNominator) Option {
return func(o *priorityQueueOptions) {
o.podNominator = pn
}
}
// WithClusterEventMap sets clusterEventMap for PriorityQueue.
func WithClusterEventMap(m map[framework.ClusterEvent]sets.String) Option {
return func(o *priorityQueueOptions) {
o.clusterEventMap = m
}
}
var defaultPriorityQueueOptions = priorityQueueOptions{
clock: util.RealClock{},
podInitialBackoffDuration: DefaultPodInitialBackoffDuration,
podMaxBackoffDuration: DefaultPodMaxBackoffDuration,
}
// Making sure that PriorityQueue implements SchedulingQueue.
var _ SchedulingQueue = &PriorityQueue{}
// newQueuedPodInfoForLookup builds a QueuedPodInfo object for a lookup in the queue.
func newQueuedPodInfoForLookup(pod *v1.Pod, plugins ...string) *framework.QueuedPodInfo {
// Since this is only used for a lookup in the queue, we only need to set the Pod,
// and so we avoid creating a full PodInfo, which is expensive to instantiate frequently.
return &framework.QueuedPodInfo{
PodInfo: &framework.PodInfo{Pod: pod},
UnschedulablePlugins: sets.NewString(plugins...),
}
}
// NewPriorityQueue creates a PriorityQueue object.
func NewPriorityQueue(
lessFn framework.LessFunc,
informerFactory informers.SharedInformerFactory,
opts ...Option,
) *PriorityQueue {
options := defaultPriorityQueueOptions
for _, opt := range opts {
opt(&options)
}
comp := func(podInfo1, podInfo2 interface{}) bool {
pInfo1 := podInfo1.(*framework.QueuedPodInfo)
pInfo2 := podInfo2.(*framework.QueuedPodInfo)
return lessFn(pInfo1, pInfo2)
}
if options.podNominator == nil {
options.podNominator = NewPodNominator(informerFactory.Core().V1().Pods().Lister())
}
pq := &PriorityQueue{
PodNominator: options.podNominator,
clock: options.clock,
stop: make(chan struct{}),
podInitialBackoffDuration: options.podInitialBackoffDuration,
podMaxBackoffDuration: options.podMaxBackoffDuration,
activeQ: heap.NewWithRecorder(podInfoKeyFunc, comp, metrics.NewActivePodsRecorder()),
unschedulableQ: newUnschedulablePodsMap(metrics.NewUnschedulablePodsRecorder()),
moveRequestCycle: -1,
clusterEventMap: options.clusterEventMap,
}
pq.cond.L = &pq.lock
pq.podBackoffQ = heap.NewWithRecorder(podInfoKeyFunc, pq.podsCompareBackoffCompleted, metrics.NewBackoffPodsRecorder())
if utilfeature.DefaultFeatureGate.Enabled(features.PodAffinityNamespaceSelector) {
pq.nsLister = informerFactory.Core().V1().Namespaces().Lister()
}
return pq
}
// Run starts the goroutine to pump from podBackoffQ to activeQ
func (p *PriorityQueue) Run() {
go wait.Until(p.flushBackoffQCompleted, 1.0*time.Second, p.stop)
go wait.Until(p.flushUnschedulableQLeftover, 30*time.Second, p.stop)
}
// Add adds a pod to the active queue. It should be called only when a new pod
// is added so there is no chance the pod is already in active/unschedulable/backoff queues
func (p *PriorityQueue) Add(pod *v1.Pod) error {
p.lock.Lock()
defer p.lock.Unlock()
pInfo := p.newQueuedPodInfo(pod)
if err := p.activeQ.Add(pInfo); err != nil {
klog.ErrorS(err, "Error adding pod to the scheduling queue", "pod", klog.KObj(pod))
return err
}
if p.unschedulableQ.get(pod) != nil {
klog.ErrorS(nil, "Error: pod is already in the unschedulable queue", "pod", klog.KObj(pod))
p.unschedulableQ.delete(pod)
}
// Delete pod from backoffQ if it is backing off
if err := p.podBackoffQ.Delete(pInfo); err == nil {
klog.ErrorS(nil, "Error: pod is already in the podBackoff queue", "pod", klog.KObj(pod))
}
metrics.SchedulerQueueIncomingPods.WithLabelValues("active", PodAdd).Inc()
p.PodNominator.AddNominatedPod(pInfo.PodInfo, "")
p.cond.Broadcast()
return nil
}
// Activate moves the given pods to activeQ iff they're in unschedulableQ or backoffQ.
func (p *PriorityQueue) Activate(pods map[string]*v1.Pod) {
p.lock.Lock()
defer p.lock.Unlock()
activated := false
for _, pod := range pods {
if p.activate(pod) {
activated = true
}
}
if activated {
p.cond.Broadcast()
}
}
func (p *PriorityQueue) activate(pod *v1.Pod) bool {
// Verify if the pod is present in activeQ.
if _, exists, _ := p.activeQ.Get(newQueuedPodInfoForLookup(pod)); exists {
// No need to activate if it's already present in activeQ.
return false
}
var pInfo *framework.QueuedPodInfo
// Verify if the pod is present in unschedulableQ or backoffQ.
if pInfo = p.unschedulableQ.get(pod); pInfo == nil {
// If the pod doesn't belong to unschedulableQ or backoffQ, don't activate it.
if obj, exists, _ := p.podBackoffQ.Get(newQueuedPodInfoForLookup(pod)); !exists {
klog.ErrorS(nil, "To-activate pod does not exist in unschedulableQ or backoffQ", "pod", klog.KObj(pod))
return false
} else {
pInfo = obj.(*framework.QueuedPodInfo)
}
}
if pInfo == nil {
// Redundant safe check. We shouldn't reach here.
klog.ErrorS(nil, "Internal error: cannot obtain pInfo")
return false
}
if err := p.activeQ.Add(pInfo); err != nil {
klog.ErrorS(err, "Error adding pod to the scheduling queue", "pod", klog.KObj(pod))
return false
}
p.unschedulableQ.delete(pod)
p.podBackoffQ.Delete(pInfo)
metrics.SchedulerQueueIncomingPods.WithLabelValues("active", ForceActivate).Inc()
p.PodNominator.AddNominatedPod(pInfo.PodInfo, "")
return true
}
// isPodBackingoff returns true if a pod is still waiting for its backoff timer.
// If this returns true, the pod should not be re-tried.
func (p *PriorityQueue) isPodBackingoff(podInfo *framework.QueuedPodInfo) bool {
boTime := p.getBackoffTime(podInfo)
return boTime.After(p.clock.Now())
}
// SchedulingCycle returns current scheduling cycle.
func (p *PriorityQueue) SchedulingCycle() int64 {
p.lock.RLock()
defer p.lock.RUnlock()
return p.schedulingCycle
}
// AddUnschedulableIfNotPresent inserts a pod that cannot be scheduled into
// the queue, unless it is already in the queue. Normally, PriorityQueue puts
// unschedulable pods in `unschedulableQ`. But if there has been a recent move
// request, then the pod is put in `podBackoffQ`.
func (p *PriorityQueue) AddUnschedulableIfNotPresent(pInfo *framework.QueuedPodInfo, podSchedulingCycle int64) error {
p.lock.Lock()
defer p.lock.Unlock()
pod := pInfo.Pod
if p.unschedulableQ.get(pod) != nil {
return fmt.Errorf("Pod %v is already present in unschedulable queue", klog.KObj(pod))
}
if _, exists, _ := p.activeQ.Get(pInfo); exists {
return fmt.Errorf("Pod %v is already present in the active queue", klog.KObj(pod))
}
if _, exists, _ := p.podBackoffQ.Get(pInfo); exists {
return fmt.Errorf("Pod %v is already present in the backoff queue", klog.KObj(pod))
}
// Refresh the timestamp since the pod is re-added.
pInfo.Timestamp = p.clock.Now()
// If a move request has been received, move it to the BackoffQ, otherwise move
// it to unschedulableQ.
if p.moveRequestCycle >= podSchedulingCycle {
if err := p.podBackoffQ.Add(pInfo); err != nil {
return fmt.Errorf("error adding pod %v to the backoff queue: %v", pod.Name, err)
}
metrics.SchedulerQueueIncomingPods.WithLabelValues("backoff", ScheduleAttemptFailure).Inc()
} else {
p.unschedulableQ.addOrUpdate(pInfo)
metrics.SchedulerQueueIncomingPods.WithLabelValues("unschedulable", ScheduleAttemptFailure).Inc()
}
p.PodNominator.AddNominatedPod(pInfo.PodInfo, "")
return nil
}
// flushBackoffQCompleted Moves all pods from backoffQ which have completed backoff in to activeQ
func (p *PriorityQueue) flushBackoffQCompleted() {
p.lock.Lock()
defer p.lock.Unlock()
for {
rawPodInfo := p.podBackoffQ.Peek()
if rawPodInfo == nil {
return
}
pod := rawPodInfo.(*framework.QueuedPodInfo).Pod
boTime := p.getBackoffTime(rawPodInfo.(*framework.QueuedPodInfo))
if boTime.After(p.clock.Now()) {
return
}
_, err := p.podBackoffQ.Pop()
if err != nil {
klog.ErrorS(err, "Unable to pop pod from backoff queue despite backoff completion", "pod", klog.KObj(pod))
return
}
p.activeQ.Add(rawPodInfo)
metrics.SchedulerQueueIncomingPods.WithLabelValues("active", BackoffComplete).Inc()
defer p.cond.Broadcast()
}
}
// flushUnschedulableQLeftover moves pod which stays in unschedulableQ longer than the unschedulableQTimeInterval
// to activeQ.
func (p *PriorityQueue) flushUnschedulableQLeftover() {
p.lock.Lock()
defer p.lock.Unlock()
var podsToMove []*framework.QueuedPodInfo
currentTime := p.clock.Now()
for _, pInfo := range p.unschedulableQ.podInfoMap {
lastScheduleTime := pInfo.Timestamp
if currentTime.Sub(lastScheduleTime) > unschedulableQTimeInterval {
podsToMove = append(podsToMove, pInfo)
}
}
if len(podsToMove) > 0 {
p.movePodsToActiveOrBackoffQueue(podsToMove, UnschedulableTimeout)
}
}
// Pop removes the head of the active queue and returns it. It blocks if the
// activeQ is empty and waits until a new item is added to the queue. It
// increments scheduling cycle when a pod is popped.
func (p *PriorityQueue) Pop() (*framework.QueuedPodInfo, error) {
p.lock.Lock()
defer p.lock.Unlock()
for p.activeQ.Len() == 0 {
// When the queue is empty, invocation of Pop() is blocked until new item is enqueued.
// When Close() is called, the p.closed is set and the condition is broadcast,
// which causes this loop to continue and return from the Pop().
if p.closed {
return nil, fmt.Errorf(queueClosed)
}
p.cond.Wait()
}
obj, err := p.activeQ.Pop()
if err != nil {
return nil, err
}
pInfo := obj.(*framework.QueuedPodInfo)
pInfo.Attempts++
p.schedulingCycle++
return pInfo, err
}
// isPodUpdated checks if the pod is updated in a way that it may have become
// schedulable. It drops status of the pod and compares it with old version.
func isPodUpdated(oldPod, newPod *v1.Pod) bool {
strip := func(pod *v1.Pod) *v1.Pod {
p := pod.DeepCopy()
p.ResourceVersion = ""
p.Generation = 0
p.Status = v1.PodStatus{}
p.ManagedFields = nil
p.Finalizers = nil
return p
}
return !reflect.DeepEqual(strip(oldPod), strip(newPod))
}
// Update updates a pod in the active or backoff queue if present. Otherwise, it removes
// the item from the unschedulable queue if pod is updated in a way that it may
// become schedulable and adds the updated one to the active queue.
// If pod is not present in any of the queues, it is added to the active queue.
func (p *PriorityQueue) Update(oldPod, newPod *v1.Pod) error {
p.lock.Lock()
defer p.lock.Unlock()
if oldPod != nil {
oldPodInfo := newQueuedPodInfoForLookup(oldPod)
// If the pod is already in the active queue, just update it there.
if oldPodInfo, exists, _ := p.activeQ.Get(oldPodInfo); exists {
pInfo := updatePod(oldPodInfo, newPod)
p.PodNominator.UpdateNominatedPod(oldPod, pInfo.PodInfo)
return p.activeQ.Update(pInfo)
}
// If the pod is in the backoff queue, update it there.
if oldPodInfo, exists, _ := p.podBackoffQ.Get(oldPodInfo); exists {
pInfo := updatePod(oldPodInfo, newPod)
p.PodNominator.UpdateNominatedPod(oldPod, pInfo.PodInfo)
return p.podBackoffQ.Update(pInfo)
}
}
// If the pod is in the unschedulable queue, updating it may make it schedulable.
if usPodInfo := p.unschedulableQ.get(newPod); usPodInfo != nil {
pInfo := updatePod(usPodInfo, newPod)
p.PodNominator.UpdateNominatedPod(oldPod, pInfo.PodInfo)
if isPodUpdated(oldPod, newPod) {
if p.isPodBackingoff(usPodInfo) {
if err := p.podBackoffQ.Add(pInfo); err != nil {
return err
}
p.unschedulableQ.delete(usPodInfo.Pod)
} else {
if err := p.activeQ.Add(pInfo); err != nil {
return err
}
p.unschedulableQ.delete(usPodInfo.Pod)
p.cond.Broadcast()
}
} else {
// Pod update didn't make it schedulable, keep it in the unschedulable queue.
p.unschedulableQ.addOrUpdate(pInfo)
}
return nil
}
// If pod is not in any of the queues, we put it in the active queue.
pInfo := p.newQueuedPodInfo(newPod)
if err := p.activeQ.Add(pInfo); err != nil {
return err
}
p.PodNominator.AddNominatedPod(pInfo.PodInfo, "")
p.cond.Broadcast()
return nil
}
// Delete deletes the item from either of the two queues. It assumes the pod is
// only in one queue.
func (p *PriorityQueue) Delete(pod *v1.Pod) error {
p.lock.Lock()
defer p.lock.Unlock()
p.PodNominator.DeleteNominatedPodIfExists(pod)
if err := p.activeQ.Delete(newQueuedPodInfoForLookup(pod)); err != nil {
// The item was probably not found in the activeQ.
p.podBackoffQ.Delete(newQueuedPodInfoForLookup(pod))
p.unschedulableQ.delete(pod)
}
return nil
}
// AssignedPodAdded is called when a bound pod is added. Creation of this pod
// may make pending pods with matching affinity terms schedulable.
func (p *PriorityQueue) AssignedPodAdded(pod *v1.Pod) {
p.lock.Lock()
p.movePodsToActiveOrBackoffQueue(p.getUnschedulablePodsWithMatchingAffinityTerm(pod), AssignedPodAdd)
p.lock.Unlock()
}
// AssignedPodUpdated is called when a bound pod is updated. Change of labels
// may make pending pods with matching affinity terms schedulable.
func (p *PriorityQueue) AssignedPodUpdated(pod *v1.Pod) {
p.lock.Lock()
p.movePodsToActiveOrBackoffQueue(p.getUnschedulablePodsWithMatchingAffinityTerm(pod), AssignedPodUpdate)
p.lock.Unlock()
}
// MoveAllToActiveOrBackoffQueue moves all pods from unschedulableQ to activeQ or backoffQ.
// This function adds all pods and then signals the condition variable to ensure that
// if Pop() is waiting for an item, it receives it after all the pods are in the
// queue and the head is the highest priority pod.
func (p *PriorityQueue) MoveAllToActiveOrBackoffQueue(event framework.ClusterEvent, preCheck PreEnqueueCheck) {
p.lock.Lock()
defer p.lock.Unlock()
unschedulablePods := make([]*framework.QueuedPodInfo, 0, len(p.unschedulableQ.podInfoMap))
for _, pInfo := range p.unschedulableQ.podInfoMap {
if preCheck == nil || preCheck(pInfo.Pod) {
unschedulablePods = append(unschedulablePods, pInfo)
}
}
p.movePodsToActiveOrBackoffQueue(unschedulablePods, event)
}
// NOTE: this function assumes lock has been acquired in caller
func (p *PriorityQueue) movePodsToActiveOrBackoffQueue(podInfoList []*framework.QueuedPodInfo, event framework.ClusterEvent) {
moved := false
for _, pInfo := range podInfoList {
// If the event doesn't help making the Pod schedulable, continue.
// Note: we don't run the check if pInfo.UnschedulablePlugins is nil, which denotes
// either there is some abnormal error, or scheduling the pod failed by plugins other than PreFilter, Filter and Permit.
// In that case, it's desired to move it anyways.
if len(pInfo.UnschedulablePlugins) != 0 && !p.podMatchesEvent(pInfo, event) {
continue
}
moved = true
pod := pInfo.Pod
if p.isPodBackingoff(pInfo) {
if err := p.podBackoffQ.Add(pInfo); err != nil {
klog.ErrorS(err, "Error adding pod to the backoff queue", "pod", klog.KObj(pod))
} else {
metrics.SchedulerQueueIncomingPods.WithLabelValues("backoff", event.Label).Inc()
p.unschedulableQ.delete(pod)
}
} else {
if err := p.activeQ.Add(pInfo); err != nil {
klog.ErrorS(err, "Error adding pod to the scheduling queue", "pod", klog.KObj(pod))
} else {
metrics.SchedulerQueueIncomingPods.WithLabelValues("active", event.Label).Inc()
p.unschedulableQ.delete(pod)
}
}
}
p.moveRequestCycle = p.schedulingCycle
if moved {
p.cond.Broadcast()
}
}
// getUnschedulablePodsWithMatchingAffinityTerm returns unschedulable pods which have
// any affinity term that matches "pod".
// NOTE: this function assumes lock has been acquired in caller.
func (p *PriorityQueue) getUnschedulablePodsWithMatchingAffinityTerm(pod *v1.Pod) []*framework.QueuedPodInfo {
nsSelectorEnabled := p.nsLister != nil
var nsLabels labels.Set
if nsSelectorEnabled {
nsLabels = interpodaffinity.GetNamespaceLabelsSnapshot(pod.Namespace, p.nsLister)
}
var podsToMove []*framework.QueuedPodInfo
for _, pInfo := range p.unschedulableQ.podInfoMap {
for _, term := range pInfo.RequiredAffinityTerms {
if term.Matches(pod, nsLabels, nsSelectorEnabled) {
podsToMove = append(podsToMove, pInfo)
break
}
}
}
return podsToMove
}
// PendingPods returns all the pending pods in the queue. This function is
// used for debugging purposes in the scheduler cache dumper and comparer.
func (p *PriorityQueue) PendingPods() []*v1.Pod {
p.lock.RLock()
defer p.lock.RUnlock()
var result []*v1.Pod
for _, pInfo := range p.activeQ.List() {
result = append(result, pInfo.(*framework.QueuedPodInfo).Pod)
}
for _, pInfo := range p.podBackoffQ.List() {
result = append(result, pInfo.(*framework.QueuedPodInfo).Pod)
}
for _, pInfo := range p.unschedulableQ.podInfoMap {
result = append(result, pInfo.Pod)
}
return result
}
// Close closes the priority queue.
func (p *PriorityQueue) Close() {
p.lock.Lock()
defer p.lock.Unlock()
close(p.stop)
p.closed = true
p.cond.Broadcast()
}
// DeleteNominatedPodIfExists deletes <pod> from nominatedPods.
func (npm *nominator) DeleteNominatedPodIfExists(pod *v1.Pod) {
npm.Lock()
npm.delete(pod)
npm.Unlock()
}
// AddNominatedPod adds a pod to the nominated pods of the given node.
// This is called during the preemption process after a node is nominated to run
// the pod. We update the structure before sending a request to update the pod
// object to avoid races with the following scheduling cycles.
func (npm *nominator) AddNominatedPod(pi *framework.PodInfo, nodeName string) {
npm.Lock()
npm.add(pi, nodeName)
npm.Unlock()
}
// NominatedPodsForNode returns pods that are nominated to run on the given node,
// but they are waiting for other pods to be removed from the node.
func (npm *nominator) NominatedPodsForNode(nodeName string) []*framework.PodInfo {
npm.RLock()
defer npm.RUnlock()
// TODO: we may need to return a copy of []*Pods to avoid modification
// on the caller side.
return npm.nominatedPods[nodeName]
}
func (p *PriorityQueue) podsCompareBackoffCompleted(podInfo1, podInfo2 interface{}) bool {
pInfo1 := podInfo1.(*framework.QueuedPodInfo)
pInfo2 := podInfo2.(*framework.QueuedPodInfo)
bo1 := p.getBackoffTime(pInfo1)
bo2 := p.getBackoffTime(pInfo2)
return bo1.Before(bo2)
}
// NumUnschedulablePods returns the number of unschedulable pods exist in the SchedulingQueue.
func (p *PriorityQueue) NumUnschedulablePods() int {
p.lock.RLock()
defer p.lock.RUnlock()
return len(p.unschedulableQ.podInfoMap)
}
// newQueuedPodInfo builds a QueuedPodInfo object.
func (p *PriorityQueue) newQueuedPodInfo(pod *v1.Pod, plugins ...string) *framework.QueuedPodInfo {
now := p.clock.Now()
return &framework.QueuedPodInfo{
PodInfo: framework.NewPodInfo(pod),
Timestamp: now,
InitialAttemptTimestamp: now,
UnschedulablePlugins: sets.NewString(plugins...),
}
}
// getBackoffTime returns the time that podInfo completes backoff
func (p *PriorityQueue) getBackoffTime(podInfo *framework.QueuedPodInfo) time.Time {
duration := p.calculateBackoffDuration(podInfo)
backoffTime := podInfo.Timestamp.Add(duration)
return backoffTime
}
// calculateBackoffDuration is a helper function for calculating the backoffDuration
// based on the number of attempts the pod has made.
func (p *PriorityQueue) calculateBackoffDuration(podInfo *framework.QueuedPodInfo) time.Duration {
duration := p.podInitialBackoffDuration
for i := 1; i < podInfo.Attempts; i++ {
duration = duration * 2
if duration > p.podMaxBackoffDuration {
return p.podMaxBackoffDuration
}
}
return duration
}
func updatePod(oldPodInfo interface{}, newPod *v1.Pod) *framework.QueuedPodInfo {
pInfo := oldPodInfo.(*framework.QueuedPodInfo)
pInfo.Update(newPod)
return pInfo
}
// UnschedulablePodsMap holds pods that cannot be scheduled. This data structure
// is used to implement unschedulableQ.
type UnschedulablePodsMap struct {
// podInfoMap is a map key by a pod's full-name and the value is a pointer to the QueuedPodInfo.
podInfoMap map[string]*framework.QueuedPodInfo
keyFunc func(*v1.Pod) string
// metricRecorder updates the counter when elements of an unschedulablePodsMap
// get added or removed, and it does nothing if it's nil
metricRecorder metrics.MetricRecorder
}
// Add adds a pod to the unschedulable podInfoMap.
func (u *UnschedulablePodsMap) addOrUpdate(pInfo *framework.QueuedPodInfo) {
podID := u.keyFunc(pInfo.Pod)
if _, exists := u.podInfoMap[podID]; !exists && u.metricRecorder != nil {
u.metricRecorder.Inc()
}
u.podInfoMap[podID] = pInfo
}
// Delete deletes a pod from the unschedulable podInfoMap.
func (u *UnschedulablePodsMap) delete(pod *v1.Pod) {
podID := u.keyFunc(pod)
if _, exists := u.podInfoMap[podID]; exists && u.metricRecorder != nil {
u.metricRecorder.Dec()
}
delete(u.podInfoMap, podID)
}
// Get returns the QueuedPodInfo if a pod with the same key as the key of the given "pod"
// is found in the map. It returns nil otherwise.
func (u *UnschedulablePodsMap) get(pod *v1.Pod) *framework.QueuedPodInfo {
podKey := u.keyFunc(pod)
if pInfo, exists := u.podInfoMap[podKey]; exists {
return pInfo
}
return nil
}
// Clear removes all the entries from the unschedulable podInfoMap.
func (u *UnschedulablePodsMap) clear() {
u.podInfoMap = make(map[string]*framework.QueuedPodInfo)
if u.metricRecorder != nil {
u.metricRecorder.Clear()
}
}
// newUnschedulablePodsMap initializes a new object of UnschedulablePodsMap.
func newUnschedulablePodsMap(metricRecorder metrics.MetricRecorder) *UnschedulablePodsMap {
return &UnschedulablePodsMap{
podInfoMap: make(map[string]*framework.QueuedPodInfo),
keyFunc: util.GetPodFullName,
metricRecorder: metricRecorder,
}
}
// nominator is a structure that stores pods nominated to run on nodes.
// It exists because nominatedNodeName of pod objects stored in the structure
// may be different than what scheduler has here. We should be able to find pods
// by their UID and update/delete them.
type nominator struct {
// podLister is used to verify if the given pod is alive.
podLister listersv1.PodLister
// nominatedPods is a map keyed by a node name and the value is a list of
// pods which are nominated to run on the node. These are pods which can be in
// the activeQ or unschedulableQ.
nominatedPods map[string][]*framework.PodInfo
// nominatedPodToNode is map keyed by a Pod UID to the node name where it is
// nominated.
nominatedPodToNode map[types.UID]string
sync.RWMutex
}
func (npm *nominator) add(pi *framework.PodInfo, nodeName string) {
// always delete the pod if it already exist, to ensure we never store more than
// one instance of the pod.
npm.delete(pi.Pod)
nnn := nodeName
if len(nnn) == 0 {
nnn = NominatedNodeName(pi.Pod)
if len(nnn) == 0 {
return
}
}
if npm.podLister != nil {
// If the pod is not alive, don't contain it.
if _, err := npm.podLister.Pods(pi.Pod.Namespace).Get(pi.Pod.Name); err != nil {
klog.V(4).InfoS("Pod doesn't exist in podLister, aborting adding it to the nominator", "pod", klog.KObj(pi.Pod))
return
}
}
npm.nominatedPodToNode[pi.Pod.UID] = nnn
for _, npi := range npm.nominatedPods[nnn] {
if npi.Pod.UID == pi.Pod.UID {
klog.V(4).InfoS("Pod already exists in the nominator", "pod", klog.KObj(npi.Pod))
return
}
}
npm.nominatedPods[nnn] = append(npm.nominatedPods[nnn], pi)
}
func (npm *nominator) delete(p *v1.Pod) {
nnn, ok := npm.nominatedPodToNode[p.UID]
if !ok {
return
}
for i, np := range npm.nominatedPods[nnn] {
if np.Pod.UID == p.UID {
npm.nominatedPods[nnn] = append(npm.nominatedPods[nnn][:i], npm.nominatedPods[nnn][i+1:]...)
if len(npm.nominatedPods[nnn]) == 0 {
delete(npm.nominatedPods, nnn)
}
break
}
}
delete(npm.nominatedPodToNode, p.UID)
}
// UpdateNominatedPod updates the <oldPod> with <newPod>.
func (npm *nominator) UpdateNominatedPod(oldPod *v1.Pod, newPodInfo *framework.PodInfo) {
npm.Lock()
defer npm.Unlock()
// In some cases, an Update event with no "NominatedNode" present is received right
// after a node("NominatedNode") is reserved for this pod in memory.
// In this case, we need to keep reserving the NominatedNode when updating the pod pointer.
nodeName := ""
// We won't fall into below `if` block if the Update event represents:
// (1) NominatedNode info is added
// (2) NominatedNode info is updated
// (3) NominatedNode info is removed
if NominatedNodeName(oldPod) == "" && NominatedNodeName(newPodInfo.Pod) == "" {
if nnn, ok := npm.nominatedPodToNode[oldPod.UID]; ok {
// This is the only case we should continue reserving the NominatedNode
nodeName = nnn
}
}
// We update irrespective of the nominatedNodeName changed or not, to ensure
// that pod pointer is updated.
npm.delete(oldPod)
npm.add(newPodInfo, nodeName)
}
// NewPodNominator creates a nominator as a backing of framework.PodNominator.
// A podLister is passed in so as to check if the pod exists
// before adding its nominatedNode info.
func NewPodNominator(podLister listersv1.PodLister) framework.PodNominator {
return &nominator{
podLister: podLister,
nominatedPods: make(map[string][]*framework.PodInfo),
nominatedPodToNode: make(map[types.UID]string),
}
}
// MakeNextPodFunc returns a function to retrieve the next pod from a given
// scheduling queue
func MakeNextPodFunc(queue SchedulingQueue) func() *framework.QueuedPodInfo {
return func() *framework.QueuedPodInfo {
podInfo, err := queue.Pop()
if err == nil {
klog.V(4).InfoS("About to try and schedule pod", "pod", klog.KObj(podInfo.Pod))
return podInfo
}
klog.ErrorS(err, "Error while retrieving next pod from scheduling queue")
return nil
}
}
func podInfoKeyFunc(obj interface{}) (string, error) {
return cache.MetaNamespaceKeyFunc(obj.(*framework.QueuedPodInfo).Pod)
}
// Checks if the Pod may become schedulable upon the event.
// This is achieved by looking up the global clusterEventMap registry.
func (p *PriorityQueue) podMatchesEvent(podInfo *framework.QueuedPodInfo, clusterEvent framework.ClusterEvent) bool {
if clusterEvent.IsWildCard() {
return true
}
for evt, nameSet := range p.clusterEventMap {
// Firstly verify if the two ClusterEvents match:
// - either the registered event from plugin side is a WildCardEvent,
// - or the two events have identical Resource fields and *compatible* ActionType.
// Note the ActionTypes don't need to be *identical*. We check if the ANDed value
// is zero or not. In this way, it's easy to tell Update&Delete is not compatible,
// but Update&All is.
evtMatch := evt.IsWildCard() ||
(evt.Resource == clusterEvent.Resource && evt.ActionType&clusterEvent.ActionType != 0)
// Secondly verify the plugin name matches.
// Note that if it doesn't match, we shouldn't continue to search.
if evtMatch && intersect(nameSet, podInfo.UnschedulablePlugins) {
return true
}
}
return false
}
func intersect(x, y sets.String) bool {
if len(x) > len(y) {
x, y = y, x
}
for v := range x {
if y.Has(v) {
return true
}
}
return false
}