scheduler

package
Version: v1.24.0 Latest Latest
Warning

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

Go to latest
Published: May 3, 2022 License: Apache-2.0 Imports: 53 Imported by: 24

Documentation

Index

Constants

View Source
const (
	// DefaultExtenderTimeout defines the default extender timeout in second.
	DefaultExtenderTimeout = 5 * time.Second
)
View Source
const (
	// SchedulerError is the reason recorded for events when an error occurs during scheduling a pod.
	SchedulerError = "SchedulerError"
)

Variables

View Source
var ErrNoNodesAvailable = fmt.Errorf("no nodes available to schedule pods")

ErrNoNodesAvailable is used to describe the error that no nodes available to schedule pods.

Functions

func Equal added in v1.22.0

func Equal(e1, e2 *HTTPExtender) bool

Equal is used to check if two extenders are equal ignoring the client field, exported for testing

func MakeDefaultErrorFunc added in v1.17.0

func MakeDefaultErrorFunc(client clientset.Interface, podLister corelisters.PodLister, podQueue internalqueue.SchedulingQueue, schedulerCache internalcache.Cache) func(*framework.QueuedPodInfo, error)

MakeDefaultErrorFunc construct a function to handle pod scheduler error

func NewHTTPExtender added in v1.22.0

func NewHTTPExtender(config *schedulerapi.Extender) (framework.Extender, error)

NewHTTPExtender creates an HTTPExtender object.

func NewInformerFactory added in v1.19.15

func NewInformerFactory(cs clientset.Interface, resyncPeriod time.Duration) informers.SharedInformerFactory

NewInformerFactory creates a SharedInformerFactory and initializes a scheduler specific in-place podInformer.

Types

type AdmissionResult added in v1.24.0

type AdmissionResult struct {
	Name                 string
	Reason               string
	InsufficientResource *noderesources.InsufficientResource
}

AdmissionResult describes the reason why Scheduler can't admit the pod. If the reason is a resource fit one, then AdmissionResult.InsufficientResource includes the details.

func AdmissionCheck added in v1.24.0

func AdmissionCheck(pod *v1.Pod, nodeInfo *framework.NodeInfo, includeAllFailures bool) []AdmissionResult

AdmissionCheck calls the filtering logic of noderesources/nodeport/nodeAffinity/nodename and returns the failure reasons. It's used in kubelet(pkg/kubelet/lifecycle/predicate.go) and scheduler. It returns the first failure if `includeAllFailures` is set to false; otherwise returns all failures.

type FrameworkCapturer added in v1.19.0

type FrameworkCapturer func(schedulerapi.KubeSchedulerProfile)

FrameworkCapturer is used for registering a notify function in building framework.

type HTTPExtender added in v1.22.0

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

HTTPExtender implements the Extender interface.

func (*HTTPExtender) Bind added in v1.22.0

func (h *HTTPExtender) Bind(binding *v1.Binding) error

Bind delegates the action of binding a pod to a node to the extender.

func (*HTTPExtender) Filter added in v1.22.0

func (h *HTTPExtender) Filter(
	pod *v1.Pod,
	nodes []*v1.Node,
) (filteredList []*v1.Node, failedNodes, failedAndUnresolvableNodes extenderv1.FailedNodesMap, err error)

Filter based on extender implemented predicate functions. The filtered list is expected to be a subset of the supplied list; otherwise the function returns an error. The failedNodes and failedAndUnresolvableNodes optionally contains the list of failed nodes and failure reasons, except nodes in the latter are unresolvable.

func (*HTTPExtender) IsBinder added in v1.22.0

func (h *HTTPExtender) IsBinder() bool

IsBinder returns whether this extender is configured for the Bind method.

func (*HTTPExtender) IsIgnorable added in v1.22.0

func (h *HTTPExtender) IsIgnorable() bool

IsIgnorable returns true indicates scheduling should not fail when this extender is unavailable

func (*HTTPExtender) IsInterested added in v1.22.0

func (h *HTTPExtender) IsInterested(pod *v1.Pod) bool

IsInterested returns true if at least one extended resource requested by this pod is managed by this extender.

func (*HTTPExtender) Name added in v1.22.0

func (h *HTTPExtender) Name() string

Name returns extenderURL to identify the extender.

func (*HTTPExtender) Prioritize added in v1.22.0

func (h *HTTPExtender) Prioritize(pod *v1.Pod, nodes []*v1.Node) (*extenderv1.HostPriorityList, int64, error)

Prioritize based on extender implemented priority functions. Weight*priority is added up for each such priority function. The returned score is added to the score computed by Kubernetes scheduler. The total score is used to do the host selection.

func (*HTTPExtender) ProcessPreemption added in v1.22.0

func (h *HTTPExtender) ProcessPreemption(
	pod *v1.Pod,
	nodeNameToVictims map[string]*extenderv1.Victims,
	nodeInfos framework.NodeInfoLister,
) (map[string]*extenderv1.Victims, error)

ProcessPreemption returns filtered candidate nodes and victims after running preemption logic in extender.

func (*HTTPExtender) SupportsPreemption added in v1.22.0

func (h *HTTPExtender) SupportsPreemption() bool

SupportsPreemption returns true if an extender supports preemption. An extender should have preempt verb defined and enabled its own node cache.

type Option added in v1.13.0

type Option func(*schedulerOptions)

Option configures a Scheduler

func WithBuildFrameworkCapturer added in v1.19.0

func WithBuildFrameworkCapturer(fc FrameworkCapturer) Option

WithBuildFrameworkCapturer sets a notify function for getting buildFramework details.

func WithComponentConfigVersion added in v1.22.0

func WithComponentConfigVersion(apiVersion string) Option

WithComponentConfigVersion sets the component config version to the KubeSchedulerConfiguration version used. The string should be the full scheme group/version of the external type we converted from (for example "kubescheduler.config.k8s.io/v1beta2")

func WithExtenders added in v1.18.0

func WithExtenders(e ...schedulerapi.Extender) Option

WithExtenders sets extenders for the Scheduler

func WithFrameworkOutOfTreeRegistry added in v1.17.0

func WithFrameworkOutOfTreeRegistry(registry frameworkruntime.Registry) Option

WithFrameworkOutOfTreeRegistry sets the registry for out-of-tree plugins. Those plugins will be appended to the default registry.

func WithKubeConfig added in v1.22.0

func WithKubeConfig(cfg *restclient.Config) Option

WithKubeConfig sets the kube config for Scheduler.

func WithParallelism added in v1.20.0

func WithParallelism(threads int32) Option

WithParallelism sets the parallelism for all scheduler algorithms. Default is 16.

func WithPercentageOfNodesToScore added in v1.13.0

func WithPercentageOfNodesToScore(percentageOfNodesToScore int32) Option

WithPercentageOfNodesToScore sets percentageOfNodesToScore for Scheduler, the default value is 50

func WithPodInitialBackoffSeconds added in v1.17.0

func WithPodInitialBackoffSeconds(podInitialBackoffSeconds int64) Option

WithPodInitialBackoffSeconds sets podInitialBackoffSeconds for Scheduler, the default value is 1

func WithPodMaxBackoffSeconds added in v1.17.0

func WithPodMaxBackoffSeconds(podMaxBackoffSeconds int64) Option

WithPodMaxBackoffSeconds sets podMaxBackoffSeconds for Scheduler, the default value is 10

func WithPodMaxInUnschedulablePodsDuration added in v1.24.0

func WithPodMaxInUnschedulablePodsDuration(duration time.Duration) Option

WithPodMaxInUnschedulablePodsDuration sets podMaxInUnschedulablePodsDuration for PriorityQueue.

func WithProfiles added in v1.18.0

func WithProfiles(p ...schedulerapi.KubeSchedulerProfile) Option

WithProfiles sets profiles for Scheduler. By default, there is one profile with the name "default-scheduler".

type ScheduleResult added in v1.22.0

type ScheduleResult struct {
	// Name of the selected node.
	SuggestedHost string
	// The number of nodes the scheduler evaluated the pod against in the filtering
	// phase and beyond.
	EvaluatedNodes int
	// The number of nodes out of the evaluated ones that fit the pod.
	FeasibleNodes int
}

ScheduleResult represents the result of scheduling a pod.

type Scheduler

type Scheduler struct {
	// It is expected that changes made via Cache will be observed
	// by NodeLister and Algorithm.
	Cache internalcache.Cache

	Extenders []framework.Extender

	// NextPod should be a function that blocks until the next pod
	// is available. We don't use a channel for this, because scheduling
	// a pod may take some amount of time and we don't want pods to get
	// stale while they sit in a channel.
	NextPod func() *framework.QueuedPodInfo

	// Error is called if there is an error. It is passed the pod in
	// question, and the error
	Error func(*framework.QueuedPodInfo, error)

	// SchedulePod tries to schedule the given pod to one of the nodes in the node list.
	// Return a struct of ScheduleResult with the name of suggested host on success,
	// otherwise will return a FitError with reasons.
	SchedulePod func(ctx context.Context, fwk framework.Framework, state *framework.CycleState, pod *v1.Pod) (ScheduleResult, error)

	// Close this to shut down the scheduler.
	StopEverything <-chan struct{}

	// SchedulingQueue holds pods to be scheduled
	SchedulingQueue internalqueue.SchedulingQueue

	// Profiles are the scheduling profiles.
	Profiles profile.Map
	// contains filtered or unexported fields
}

Scheduler watches for new unscheduled pods. It attempts to find nodes that they fit on and writes bindings back to the api server.

func New added in v1.13.0

func New(client clientset.Interface,
	informerFactory informers.SharedInformerFactory,
	dynInformerFactory dynamicinformer.DynamicSharedInformerFactory,
	recorderFactory profile.RecorderFactory,
	stopCh <-chan struct{},
	opts ...Option) (*Scheduler, error)

New returns a Scheduler

func (*Scheduler) Run added in v1.10.0

func (sched *Scheduler) Run(ctx context.Context)

Run begins watching and scheduling. It starts scheduling and blocked until the context is done.

Directories

Path Synopsis
apis
resources
Package resources provides a metrics collector that reports the resource consumption (requests and limits) of the pods in the cluster as the scheduler and kubelet would interpret it.
Package resources provides a metrics collector that reports the resource consumption (requests and limits) of the pods in the cluster as the scheduler and kubelet would interpret it.
Package profile holds the definition of a scheduling Profile.
Package profile holds the definition of a scheduling Profile.
internal

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL