queue

package
v1.32.0 Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Dec 11, 2024 License: Apache-2.0 Imports: 29 Imported by: 0

Documentation

Index

Constants

View Source 
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
)
View Source 
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
)

Variables

This section is empty.

Functions

This section is empty.

Types

type Option 

type Option func(*priorityQueueOptions)

Option configures a PriorityQueue

func WithClock 

func WithClock(clock clock.Clock) Option

WithClock sets clock for PriorityQueue, the default clock is clock.RealClock.

func WithMetricsRecorder 

func WithMetricsRecorder(recorder metrics.MetricAsyncRecorder) Option

WithMetricsRecorder sets metrics recorder.

func WithPluginMetricsSamplePercent 

func WithPluginMetricsSamplePercent(percent int) Option

WithPluginMetricsSamplePercent sets the percentage of plugin metrics to be sampled.

func WithPodInitialBackoffDuration 

func WithPodInitialBackoffDuration(duration time.Duration) Option

WithPodInitialBackoffDuration sets pod initial backoff duration for PriorityQueue.

func WithPodLister 

func WithPodLister(pl listersv1.PodLister) Option

WithPodLister sets pod lister for PriorityQueue.

func WithPodMaxBackoffDuration 

func WithPodMaxBackoffDuration(duration time.Duration) Option

WithPodMaxBackoffDuration sets pod max backoff duration for PriorityQueue.

func WithPodMaxInUnschedulablePodsDuration 

func WithPodMaxInUnschedulablePodsDuration(duration time.Duration) Option

WithPodMaxInUnschedulablePodsDuration sets podMaxInUnschedulablePodsDuration for PriorityQueue.

func WithPreEnqueuePluginMap 

func WithPreEnqueuePluginMap(m map[string][]framework.PreEnqueuePlugin) Option

WithPreEnqueuePluginMap sets preEnqueuePluginMap for PriorityQueue.

func WithQueueingHintMapPerProfile 

func WithQueueingHintMapPerProfile(m QueueingHintMapPerProfile) Option

WithQueueingHintMapPerProfile sets queueingHintMap for PriorityQueue.

type PreEnqueueCheck 

type PreEnqueueCheck func(pod *v1.Pod) bool

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 PriorityQueue 

type PriorityQueue struct {
	// contains filtered or unexported fields
}

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.

func NewPriorityQueue 

func NewPriorityQueue(
	lessFn framework.LessFunc,
	informerFactory informers.SharedInformerFactory,
	opts ...Option,
) *PriorityQueue

NewPriorityQueue creates a PriorityQueue object.

func NewTestQueue 

func NewTestQueue(ctx context.Context, lessFn framework.LessFunc, opts ...Option) *PriorityQueue

NewTestQueue creates a priority queue with an empty informer factory.

func NewTestQueueWithInformerFactory 

func NewTestQueueWithInformerFactory(
	ctx context.Context,
	lessFn framework.LessFunc,
	informerFactory informers.SharedInformerFactory,
	opts ...Option,
) *PriorityQueue

func NewTestQueueWithObjects 

func NewTestQueueWithObjects(
	ctx context.Context,
	lessFn framework.LessFunc,
	objs []runtime.Object,
	opts ...Option,
) *PriorityQueue

NewTestQueueWithObjects creates a priority queue with an informer factory populated with the provided objects.

func (*PriorityQueue) Activate 

func (p *PriorityQueue) Activate(logger klog.Logger, pods map[string]*v1.Pod)

Activate moves the given pods to activeQ. If a pod isn't found in unschedulablePods or backoffQ and it's in-flight, the wildcard event is registered so that the pod will be requeued when it comes back. But, if a pod isn't found in unschedulablePods or backoffQ and it's not in-flight (i.e., completely unknown pod), Activate would ignore the pod.

func (*PriorityQueue) Add 

func (p *PriorityQueue) Add(logger klog.Logger, pod *v1.Pod)

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 (PriorityQueue) AddNominatedPod 

func (npm PriorityQueue) AddNominatedPod(logger klog.Logger, pi *framework.PodInfo, nominatingInfo *framework.NominatingInfo)

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 (*PriorityQueue) AddUnschedulableIfNotPresent 

func (p *PriorityQueue) AddUnschedulableIfNotPresent(logger klog.Logger, pInfo *framework.QueuedPodInfo, podSchedulingCycle int64) error

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 (*PriorityQueue) AssignedPodAdded 

func (p *PriorityQueue) AssignedPodAdded(logger klog.Logger, pod *v1.Pod)

AssignedPodAdded is called when a bound pod is added. Creation of this pod may make pending pods with matching affinity terms schedulable.

func (*PriorityQueue) AssignedPodUpdated 

func (p *PriorityQueue) AssignedPodUpdated(logger klog.Logger, oldPod, newPod *v1.Pod, event framework.ClusterEvent)

AssignedPodUpdated is called when a bound pod is updated. Change of labels may make pending pods with matching affinity terms schedulable.

func (*PriorityQueue) Close 

func (p *PriorityQueue) Close()

Close closes the priority queue.

func (*PriorityQueue) Delete 

func (p *PriorityQueue) Delete(pod *v1.Pod)

Delete deletes the item from either of the two queues. It assumes the pod is only in one queue.

func (PriorityQueue) DeleteNominatedPodIfExists 

func (npm PriorityQueue) DeleteNominatedPodIfExists(pod *v1.Pod)

DeleteNominatedPodIfExists deletes <pod> from nominatedPods.

func (*PriorityQueue) Done 

func (p *PriorityQueue) Done(pod types.UID)

Done must be called for pod returned by Pop. This allows the queue to keep track of which pods are currently being processed.

func (*PriorityQueue) GetPod 

func (p *PriorityQueue) GetPod(name, namespace string) (pInfo *framework.QueuedPodInfo, ok bool)

GetPod searches for a pod in the activeQ, backoffQ, and unschedulablePods.

func (*PriorityQueue) InFlightPods 

func (p *PriorityQueue) InFlightPods() []*v1.Pod

func (*PriorityQueue) MoveAllToActiveOrBackoffQueue 

func (p *PriorityQueue) MoveAllToActiveOrBackoffQueue(logger klog.Logger, event framework.ClusterEvent, oldObj, newObj interface{}, preCheck PreEnqueueCheck)

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 (*PriorityQueue) NominatedPodsForNode 

func (p *PriorityQueue) NominatedPodsForNode(nodeName string) []*framework.PodInfo

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. CAUTION: Make sure you don't call this function while taking any queue's lock in any scenario.

func (*PriorityQueue) PendingPods 

func (p *PriorityQueue) PendingPods() ([]*v1.Pod, string)

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 (*PriorityQueue) PodsInActiveQ 

func (p *PriorityQueue) PodsInActiveQ() []*v1.Pod

PodsInActiveQ returns all the Pods in the activeQ.

func (*PriorityQueue) Pop 

func (p *PriorityQueue) Pop(logger klog.Logger) (*framework.QueuedPodInfo, error)

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. Note: This method should NOT be locked by the p.lock at any moment, as it would lead to scheduling throughput degradation.

func (*PriorityQueue) Run 

func (p *PriorityQueue) Run(logger klog.Logger)

Run starts the goroutine to pump from podBackoffQ to activeQ

func (*PriorityQueue) SchedulingCycle 

func (p *PriorityQueue) SchedulingCycle() int64

SchedulingCycle returns current scheduling cycle.

func (*PriorityQueue) Update 

func (p *PriorityQueue) Update(logger klog.Logger, oldPod, newPod *v1.Pod)

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 (PriorityQueue) UpdateNominatedPod 

func (npm PriorityQueue) UpdateNominatedPod(logger klog.Logger, oldPod *v1.Pod, newPodInfo *framework.PodInfo)

UpdateNominatedPod updates the <oldPod> with <newPod>.

type QueueingHintFunction 

type QueueingHintFunction struct {
	PluginName     string
	QueueingHintFn framework.QueueingHintFn
}

QueueingHintFunction is the wrapper of QueueingHintFn that has PluginName.

type QueueingHintMap 

type QueueingHintMap map[framework.ClusterEvent][]*QueueingHintFunction

QueueingHintMap is keyed with ClusterEvent, valued with queueing hint functions registered for the event.

type QueueingHintMapPerProfile 

type QueueingHintMapPerProfile map[string]QueueingHintMap

QueueingHintMapPerProfile is keyed with profile name, valued with queueing hint map registered for the profile.

type SchedulingQueue 

type SchedulingQueue interface {
	framework.PodNominator
	Add(logger klog.Logger, pod *v1.Pod)
	// Activate moves the given pods to activeQ.
	// If a pod isn't found in unschedulablePods or backoffQ and it's in-flight,
	// the wildcard event is registered so that the pod will be requeued when it comes back.
	// But, if a pod isn't found in unschedulablePods or backoffQ and it's not in-flight (i.e., completely unknown pod),
	// Activate would ignore the pod.
	Activate(logger klog.Logger, 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(logger klog.Logger, 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(logger klog.Logger) (*framework.QueuedPodInfo, error)
	// Done must be called for pod returned by Pop. This allows the queue to
	// keep track of which pods are currently being processed.
	Done(types.UID)
	Update(logger klog.Logger, oldPod, newPod *v1.Pod)
	Delete(pod *v1.Pod)
	// Important Note: preCheck shouldn't include anything that depends on the in-tree plugins' logic.
	// (e.g., filter Pods based on added/updated Node's capacity, etc.)
	// We know currently some do, but we'll eventually remove them in favor of the scheduling queue hint.
	MoveAllToActiveOrBackoffQueue(logger klog.Logger, event framework.ClusterEvent, oldObj, newObj interface{}, preCheck PreEnqueueCheck)
	AssignedPodAdded(logger klog.Logger, pod *v1.Pod)
	AssignedPodUpdated(logger klog.Logger, oldPod, newPod *v1.Pod, event framework.ClusterEvent)

	// 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(logger klog.Logger)

	// The following functions are supposed to be used only for testing or debugging.
	GetPod(name, namespace string) (*framework.QueuedPodInfo, bool)
	PendingPods() ([]*v1.Pod, string)
	InFlightPods() []*v1.Pod
	PodsInActiveQ() []*v1.Pod
}

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.

func NewSchedulingQueue 

func NewSchedulingQueue(
	lessFn framework.LessFunc,
	informerFactory informers.SharedInformerFactory,
	opts ...Option) SchedulingQueue

NewSchedulingQueue initializes a priority queue as a new scheduling queue.

type UnschedulablePods 

type UnschedulablePods struct {
	// contains filtered or unexported fields
}

UnschedulablePods holds pods that cannot be scheduled. This data structure is used to implement unschedulablePods.

Source Files

View all Source files

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.