/
scheduling_queue.go
1183 lines (1060 loc) · 44.1 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
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/*
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 and a additional data structure,
// namely: activeQ, backoffQ and unschedulablePods.
// - activeQ holds pods that are being considered for scheduling.
// - backoffQ holds pods that moved from unschedulablePods and will move to
// activeQ when their backoff periods complete.
// - unschedulablePods holds pods that were already attempted for scheduling and
// are currently determined to be unschedulable.
package queue
import (
"context"
"fmt"
"math/rand"
"reflect"
"sync"
"time"
v1 "k8s.io/api/core/v1"
"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"
"k8s.io/utils/clock"
)
const (
// DefaultPodMaxInUnschedulablePodsDuration is the default value for the maximum
// time a pod can stay in unschedulablePods. If a pod stays in unschedulablePods
// for longer than this value, the pod will be moved from unschedulablePods to
// backoffQ or activeQ. If this value is empty, the default value (5min)
// will be used.
DefaultPodMaxInUnschedulablePodsDuration time.Duration = 5 * time.Minute
queueClosed = "scheduling queue is closed"
// Scheduling queue names
activeQName = "Active"
backoffQName = "Backoff"
unschedulablePods = "Unschedulable"
preEnqueue = "PreEnqueue"
)
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 unschedulablePods 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, string)
// Close closes the SchedulingQueue so that the goroutine which is
// waiting to pop items can exit gracefully.
Close()
// 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 two sub queues and a additional data structure, namely: activeQ,
// backoffQ and unschedulablePods.
// - activeQ holds pods that are being considered for scheduling.
// - backoffQ holds pods that moved from unschedulablePods and will move to
// activeQ when their backoff periods complete.
// - unschedulablePods holds pods that were already attempted for scheduling and
// are currently determined to be unschedulable.
type PriorityQueue struct {
*nominator
stop chan struct{}
clock clock.Clock
// pod initial backoff duration.
podInitialBackoffDuration time.Duration
// pod maximum backoff duration.
podMaxBackoffDuration time.Duration
// the maximum time a pod can stay in the unschedulablePods.
podMaxInUnschedulablePodsDuration time.Duration
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
// unschedulablePods holds pods that have been tried and determined unschedulable.
unschedulablePods *UnschedulablePods
// 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
// preEnqueuePluginMap is keyed with profile name, valued with registered preEnqueue plugins.
preEnqueuePluginMap map[string][]framework.PreEnqueuePlugin
// 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
metricsRecorder metrics.MetricAsyncRecorder
// pluginMetricsSamplePercent is the percentage of plugin metrics to be sampled.
pluginMetricsSamplePercent int
}
type priorityQueueOptions struct {
clock clock.Clock
podInitialBackoffDuration time.Duration
podMaxBackoffDuration time.Duration
podMaxInUnschedulablePodsDuration time.Duration
podLister listersv1.PodLister
metricsRecorder metrics.MetricAsyncRecorder
pluginMetricsSamplePercent int
clusterEventMap map[framework.ClusterEvent]sets.String
preEnqueuePluginMap map[string][]framework.PreEnqueuePlugin
}
// Option configures a PriorityQueue
type Option func(*priorityQueueOptions)
// WithClock sets clock for PriorityQueue, the default clock is clock.RealClock.
func WithClock(clock clock.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
}
}
// WithPodLister sets pod lister for PriorityQueue.
func WithPodLister(pl listersv1.PodLister) Option {
return func(o *priorityQueueOptions) {
o.podLister = pl
}
}
// WithClusterEventMap sets clusterEventMap for PriorityQueue.
func WithClusterEventMap(m map[framework.ClusterEvent]sets.String) Option {
return func(o *priorityQueueOptions) {
o.clusterEventMap = m
}
}
// WithPodMaxInUnschedulablePodsDuration sets podMaxInUnschedulablePodsDuration for PriorityQueue.
func WithPodMaxInUnschedulablePodsDuration(duration time.Duration) Option {
return func(o *priorityQueueOptions) {
o.podMaxInUnschedulablePodsDuration = duration
}
}
// WithPreEnqueuePluginMap sets preEnqueuePluginMap for PriorityQueue.
func WithPreEnqueuePluginMap(m map[string][]framework.PreEnqueuePlugin) Option {
return func(o *priorityQueueOptions) {
o.preEnqueuePluginMap = m
}
}
// WithMetricsRecorder sets metrics recorder.
func WithMetricsRecorder(recorder metrics.MetricAsyncRecorder) Option {
return func(o *priorityQueueOptions) {
o.metricsRecorder = recorder
}
}
// WithPluginMetricsSamplePercent sets the percentage of plugin metrics to be sampled.
func WithPluginMetricsSamplePercent(percent int) Option {
return func(o *priorityQueueOptions) {
o.pluginMetricsSamplePercent = percent
}
}
var defaultPriorityQueueOptions = priorityQueueOptions{
clock: clock.RealClock{},
podInitialBackoffDuration: DefaultPodInitialBackoffDuration,
podMaxBackoffDuration: DefaultPodMaxBackoffDuration,
podMaxInUnschedulablePodsDuration: DefaultPodMaxInUnschedulablePodsDuration,
}
// 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
if options.podLister == nil {
options.podLister = informerFactory.Core().V1().Pods().Lister()
}
for _, opt := range opts {
opt(&options)
}
comp := func(podInfo1, podInfo2 interface{}) bool {
pInfo1 := podInfo1.(*framework.QueuedPodInfo)
pInfo2 := podInfo2.(*framework.QueuedPodInfo)
return lessFn(pInfo1, pInfo2)
}
pq := &PriorityQueue{
nominator: newPodNominator(options.podLister),
clock: options.clock,
stop: make(chan struct{}),
podInitialBackoffDuration: options.podInitialBackoffDuration,
podMaxBackoffDuration: options.podMaxBackoffDuration,
podMaxInUnschedulablePodsDuration: options.podMaxInUnschedulablePodsDuration,
activeQ: heap.NewWithRecorder(podInfoKeyFunc, comp, metrics.NewActivePodsRecorder()),
unschedulablePods: newUnschedulablePods(metrics.NewUnschedulablePodsRecorder(), metrics.NewGatedPodsRecorder()),
moveRequestCycle: -1,
clusterEventMap: options.clusterEventMap,
preEnqueuePluginMap: options.preEnqueuePluginMap,
metricsRecorder: options.metricsRecorder,
pluginMetricsSamplePercent: options.pluginMetricsSamplePercent,
}
pq.cond.L = &pq.lock
pq.podBackoffQ = heap.NewWithRecorder(podInfoKeyFunc, pq.podsCompareBackoffCompleted, metrics.NewBackoffPodsRecorder())
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.flushUnschedulablePodsLeftover, 30*time.Second, p.stop)
}
// runPreEnqueuePlugins iterates PreEnqueue function in each registered PreEnqueuePlugin.
// It returns true if all PreEnqueue function run successfully; otherwise returns false
// upon the first failure.
// Note: we need to associate the failed plugin to `pInfo`, so that the pod can be moved back
// to activeQ by related cluster event.
func (p *PriorityQueue) runPreEnqueuePlugins(ctx context.Context, pInfo *framework.QueuedPodInfo) bool {
var s *framework.Status
pod := pInfo.Pod
startTime := time.Now()
defer func() {
metrics.FrameworkExtensionPointDuration.WithLabelValues(preEnqueue, s.Code().String(), pod.Spec.SchedulerName).Observe(metrics.SinceInSeconds(startTime))
}()
shouldRecordMetric := rand.Intn(100) < p.pluginMetricsSamplePercent
for _, pl := range p.preEnqueuePluginMap[pod.Spec.SchedulerName] {
s = p.runPreEnqueuePlugin(ctx, pl, pod, shouldRecordMetric)
if s.IsSuccess() {
continue
}
pInfo.UnschedulablePlugins.Insert(pl.Name())
metrics.UnschedulableReason(pl.Name(), pod.Spec.SchedulerName).Inc()
if s.Code() == framework.Error {
klog.ErrorS(s.AsError(), "Unexpected error running PreEnqueue plugin", "pod", klog.KObj(pod), "plugin", pl.Name())
} else {
klog.V(5).InfoS("Status after running PreEnqueue plugin", "pod", klog.KObj(pod), "plugin", pl.Name(), "status", s)
}
return false
}
return true
}
func (p *PriorityQueue) runPreEnqueuePlugin(ctx context.Context, pl framework.PreEnqueuePlugin, pod *v1.Pod, shouldRecordMetric bool) *framework.Status {
if !shouldRecordMetric {
return pl.PreEnqueue(ctx, pod)
}
startTime := p.clock.Now()
s := pl.PreEnqueue(ctx, pod)
p.metricsRecorder.ObservePluginDurationAsync(preEnqueue, pl.Name(), s.Code().String(), p.clock.Since(startTime).Seconds())
return s
}
// addToActiveQ tries to add pod to active queue. It returns 2 parameters:
// 1. a boolean flag to indicate whether the pod is added successfully.
// 2. an error for the caller to act on.
func (p *PriorityQueue) addToActiveQ(pInfo *framework.QueuedPodInfo) (bool, error) {
pInfo.Gated = !p.runPreEnqueuePlugins(context.Background(), pInfo)
if pInfo.Gated {
// Add the Pod to unschedulablePods if it's not passing PreEnqueuePlugins.
p.unschedulablePods.addOrUpdate(pInfo)
return false, nil
}
if err := p.activeQ.Add(pInfo); err != nil {
klog.ErrorS(err, "Error adding pod to the active queue", "pod", klog.KObj(pInfo.Pod))
return false, err
}
return true, nil
}
// 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)
gated := pInfo.Gated
if added, err := p.addToActiveQ(pInfo); !added {
return err
}
if p.unschedulablePods.get(pod) != nil {
klog.ErrorS(nil, "Error: pod is already in the unschedulable queue", "pod", klog.KObj(pod))
p.unschedulablePods.delete(pod, gated)
}
// 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))
}
klog.V(5).InfoS("Pod moved to an internal scheduling queue", "pod", klog.KObj(pod), "event", PodAdd, "queue", activeQName)
metrics.SchedulerQueueIncomingPods.WithLabelValues("active", PodAdd).Inc()
p.addNominatedPodUnlocked(pInfo.PodInfo, nil)
p.cond.Broadcast()
return nil
}
// Activate moves the given pods to activeQ iff they're in unschedulablePods 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 unschedulablePods or backoffQ.
if pInfo = p.unschedulablePods.get(pod); pInfo == nil {
// If the pod doesn't belong to unschedulablePods 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 unschedulablePods 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
}
gated := pInfo.Gated
if added, _ := p.addToActiveQ(pInfo); !added {
return false
}
p.unschedulablePods.delete(pInfo.Pod, gated)
p.podBackoffQ.Delete(pInfo)
metrics.SchedulerQueueIncomingPods.WithLabelValues("active", ForceActivate).Inc()
p.addNominatedPodUnlocked(pInfo.PodInfo, nil)
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 {
if podInfo.Gated {
return false
}
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 `unschedulablePods`. 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.unschedulablePods.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 unschedulablePods.
for plugin := range pInfo.UnschedulablePlugins {
metrics.UnschedulableReason(plugin, pInfo.Pod.Spec.SchedulerName).Inc()
}
if p.moveRequestCycle >= podSchedulingCycle {
if err := p.podBackoffQ.Add(pInfo); err != nil {
return fmt.Errorf("error adding pod %v to the backoff queue: %v", klog.KObj(pod), err)
}
klog.V(5).InfoS("Pod moved to an internal scheduling queue", "pod", klog.KObj(pod), "event", ScheduleAttemptFailure, "queue", backoffQName)
metrics.SchedulerQueueIncomingPods.WithLabelValues("backoff", ScheduleAttemptFailure).Inc()
} else {
p.unschedulablePods.addOrUpdate(pInfo)
klog.V(5).InfoS("Pod moved to an internal scheduling queue", "pod", klog.KObj(pod), "event", ScheduleAttemptFailure, "queue", unschedulablePods)
metrics.SchedulerQueueIncomingPods.WithLabelValues("unschedulable", ScheduleAttemptFailure).Inc()
}
p.addNominatedPodUnlocked(pInfo.PodInfo, nil)
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()
activated := false
for {
rawPodInfo := p.podBackoffQ.Peek()
if rawPodInfo == nil {
break
}
pInfo := rawPodInfo.(*framework.QueuedPodInfo)
pod := pInfo.Pod
if p.isPodBackingoff(pInfo) {
break
}
_, err := p.podBackoffQ.Pop()
if err != nil {
klog.ErrorS(err, "Unable to pop pod from backoff queue despite backoff completion", "pod", klog.KObj(pod))
break
}
if err := p.activeQ.Add(pInfo); err != nil {
klog.ErrorS(err, "Error adding pod to the active queue", "pod", klog.KObj(pInfo.Pod))
} else {
klog.V(5).InfoS("Pod moved to an internal scheduling queue", "pod", klog.KObj(pod), "event", BackoffComplete, "queue", activeQName)
metrics.SchedulerQueueIncomingPods.WithLabelValues("active", BackoffComplete).Inc()
activated = true
}
}
if activated {
p.cond.Broadcast()
}
}
// flushUnschedulablePodsLeftover moves pods which stay in unschedulablePods
// longer than podMaxInUnschedulablePodsDuration to backoffQ or activeQ.
func (p *PriorityQueue) flushUnschedulablePodsLeftover() {
p.lock.Lock()
defer p.lock.Unlock()
var podsToMove []*framework.QueuedPodInfo
currentTime := p.clock.Now()
for _, pInfo := range p.unschedulablePods.podInfoMap {
lastScheduleTime := pInfo.Timestamp
if currentTime.Sub(lastScheduleTime) > p.podMaxInUnschedulablePodsDuration {
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, nil
}
// 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.updateNominatedPodUnlocked(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.updateNominatedPodUnlocked(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.unschedulablePods.get(newPod); usPodInfo != nil {
pInfo := updatePod(usPodInfo, newPod)
p.updateNominatedPodUnlocked(oldPod, pInfo.PodInfo)
if isPodUpdated(oldPod, newPod) {
gated := usPodInfo.Gated
if p.isPodBackingoff(usPodInfo) {
if err := p.podBackoffQ.Add(pInfo); err != nil {
return err
}
p.unschedulablePods.delete(usPodInfo.Pod, gated)
klog.V(5).InfoS("Pod moved to an internal scheduling queue", "pod", klog.KObj(pInfo.Pod), "event", PodUpdate, "queue", backoffQName)
} else {
if added, err := p.addToActiveQ(pInfo); !added {
return err
}
p.unschedulablePods.delete(usPodInfo.Pod, gated)
klog.V(5).InfoS("Pod moved to an internal scheduling queue", "pod", klog.KObj(pInfo.Pod), "event", BackoffComplete, "queue", activeQName)
p.cond.Broadcast()
}
} else {
// Pod update didn't make it schedulable, keep it in the unschedulable queue.
p.unschedulablePods.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 added, err := p.addToActiveQ(pInfo); !added {
return err
}
p.addNominatedPodUnlocked(pInfo.PodInfo, nil)
klog.V(5).InfoS("Pod moved to an internal scheduling queue", "pod", klog.KObj(pInfo.Pod), "event", PodUpdate, "queue", activeQName)
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.deleteNominatedPodIfExistsUnlocked(pod)
pInfo := newQueuedPodInfoForLookup(pod)
if err := p.activeQ.Delete(pInfo); err != nil {
// The item was probably not found in the activeQ.
p.podBackoffQ.Delete(pInfo)
if pInfo = p.unschedulablePods.get(pod); pInfo != nil {
p.unschedulablePods.delete(pod, pInfo.Gated)
}
}
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()
}
// isPodResourcesResizedDown returns true if a pod CPU and/or memory resize request has been
// admitted by kubelet, is 'InProgress', and results in a net sizing down of updated resources.
// It returns false if either CPU or memory resource is net resized up, or if no resize is in progress.
func isPodResourcesResizedDown(pod *v1.Pod) bool {
if utilfeature.DefaultFeatureGate.Enabled(features.InPlacePodVerticalScaling) {
// TODO(vinaykul,wangchen615,InPlacePodVerticalScaling): Fix this to determine when a
// pod is truly resized down (might need oldPod if we cannot determine from Status alone)
if pod.Status.Resize == v1.PodResizeStatusInProgress {
return true
}
}
return false
}
// 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()
if isPodResourcesResizedDown(pod) {
p.moveAllToActiveOrBackoffQueue(AssignedPodUpdate, nil)
} else {
p.movePodsToActiveOrBackoffQueue(p.getUnschedulablePodsWithMatchingAffinityTerm(pod), AssignedPodUpdate)
}
p.lock.Unlock()
}
// NOTE: this function assumes a lock has been acquired in the caller.
// moveAllToActiveOrBackoffQueue moves all pods from unschedulablePods 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 the signal 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) {
unschedulablePods := make([]*framework.QueuedPodInfo, 0, len(p.unschedulablePods.podInfoMap))
for _, pInfo := range p.unschedulablePods.podInfoMap {
if preCheck == nil || preCheck(pInfo.Pod) {
unschedulablePods = append(unschedulablePods, pInfo)
}
}
p.movePodsToActiveOrBackoffQueue(unschedulablePods, event)
}
// MoveAllToActiveOrBackoffQueue moves all pods from unschedulablePods 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 the signal 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()
p.moveAllToActiveOrBackoffQueue(event, preCheck)
}
// NOTE: this function assumes lock has been acquired in caller
func (p *PriorityQueue) movePodsToActiveOrBackoffQueue(podInfoList []*framework.QueuedPodInfo, event framework.ClusterEvent) {
activated := 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
}
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 {
klog.V(5).InfoS("Pod moved to an internal scheduling queue", "pod", klog.KObj(pInfo.Pod), "event", event, "queue", backoffQName)
metrics.SchedulerQueueIncomingPods.WithLabelValues("backoff", event.Label).Inc()
p.unschedulablePods.delete(pod, pInfo.Gated)
}
} else {
gated := pInfo.Gated
if added, _ := p.addToActiveQ(pInfo); added {
klog.V(5).InfoS("Pod moved to an internal scheduling queue", "pod", klog.KObj(pInfo.Pod), "event", event, "queue", activeQName)
activated = true
metrics.SchedulerQueueIncomingPods.WithLabelValues("active", event.Label).Inc()
p.unschedulablePods.delete(pod, gated)
}
}
}
p.moveRequestCycle = p.schedulingCycle
if activated {
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 {
nsLabels := interpodaffinity.GetNamespaceLabelsSnapshot(pod.Namespace, p.nsLister)
var podsToMove []*framework.QueuedPodInfo
for _, pInfo := range p.unschedulablePods.podInfoMap {
for _, term := range pInfo.RequiredAffinityTerms {
if term.Matches(pod, nsLabels) {
podsToMove = append(podsToMove, pInfo)
break
}
}
}
return podsToMove
}
var pendingPodsSummary = "activeQ:%v; backoffQ:%v; unschedulablePods:%v"
// PendingPods returns all the pending pods in the queue; accompanied by a debugging string
// recording showing the number of pods in each queue respectively.
// This function is used for debugging purposes in the scheduler cache dumper and comparer.
func (p *PriorityQueue) PendingPods() ([]*v1.Pod, string) {
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.unschedulablePods.podInfoMap {
result = append(result, pInfo.Pod)
}
return result, fmt.Sprintf(pendingPodsSummary, p.activeQ.Len(), p.podBackoffQ.Len(), len(p.unschedulablePods.podInfoMap))
}
// 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.Lock()
npm.deleteNominatedPodIfExistsUnlocked(pod)
npm.lock.Unlock()
}
func (npm *nominator) deleteNominatedPodIfExistsUnlocked(pod *v1.Pod) {
npm.delete(pod)
}
// 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, nominatingInfo *framework.NominatingInfo) {
npm.lock.Lock()
npm.addNominatedPodUnlocked(pi, nominatingInfo)
npm.lock.Unlock()
}
// NominatedPodsForNode returns a copy of 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.lock.RLock()
defer npm.lock.RUnlock()
// Make a copy of the nominated Pods so the caller can mutate safely.
pods := make([]*framework.PodInfo, len(npm.nominatedPods[nodeName]))
for i := 0; i < len(pods); i++ {
pods[i] = npm.nominatedPods[nodeName][i].DeepCopy()
}
return pods
}
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)
}
// newQueuedPodInfo builds a QueuedPodInfo object.
func (p *PriorityQueue) newQueuedPodInfo(pod *v1.Pod, plugins ...string) *framework.QueuedPodInfo {
now := p.clock.Now()
// ignore this err since apiserver doesn't properly validate affinity terms
// and we can't fix the validation for backwards compatibility.
podInfo, _ := framework.NewPodInfo(pod)
return &framework.QueuedPodInfo{
PodInfo: podInfo,
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++ {
// Use subtraction instead of addition or multiplication to avoid overflow.
if duration > p.podMaxBackoffDuration-duration {
return p.podMaxBackoffDuration
}
duration += duration
}
return duration
}
func updatePod(oldPodInfo interface{}, newPod *v1.Pod) *framework.QueuedPodInfo {
pInfo := oldPodInfo.(*framework.QueuedPodInfo)
pInfo.Update(newPod)
return pInfo
}
// UnschedulablePods holds pods that cannot be scheduled. This data structure
// is used to implement unschedulablePods.
type UnschedulablePods 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
// unschedulableRecorder/gatedRecorder updates the counter when elements of an unschedulablePodsMap
// get added or removed, and it does nothing if it's nil.
unschedulableRecorder, gatedRecorder metrics.MetricRecorder
}
// addOrUpdate adds a pod to the unschedulable podInfoMap.
func (u *UnschedulablePods) addOrUpdate(pInfo *framework.QueuedPodInfo) {
podID := u.keyFunc(pInfo.Pod)
if _, exists := u.podInfoMap[podID]; !exists {
if pInfo.Gated && u.gatedRecorder != nil {
u.gatedRecorder.Inc()
} else if !pInfo.Gated && u.unschedulableRecorder != nil {
u.unschedulableRecorder.Inc()
}
}
u.podInfoMap[podID] = pInfo
}
// delete deletes a pod from the unschedulable podInfoMap.
// The `gated` parameter is used to figure out which metric should be decreased.
func (u *UnschedulablePods) delete(pod *v1.Pod, gated bool) {
podID := u.keyFunc(pod)
if _, exists := u.podInfoMap[podID]; exists {
if gated && u.gatedRecorder != nil {
u.gatedRecorder.Dec()
} else if !gated && u.unschedulableRecorder != nil {
u.unschedulableRecorder.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 *UnschedulablePods) 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 *UnschedulablePods) clear() {
u.podInfoMap = make(map[string]*framework.QueuedPodInfo)
if u.unschedulableRecorder != nil {
u.unschedulableRecorder.Clear()
}
if u.gatedRecorder != nil {
u.gatedRecorder.Clear()
}
}
// newUnschedulablePods initializes a new object of UnschedulablePods.
func newUnschedulablePods(unschedulableRecorder, gatedRecorder metrics.MetricRecorder) *UnschedulablePods {
return &UnschedulablePods{
podInfoMap: make(map[string]*framework.QueuedPodInfo),
keyFunc: util.GetPodFullName,