job

Documentation

Overview

Package job contains logic for watching and synchronizing jobs.

Index

• Constants
• Variables
• type Controller
  • func NewController(ctx context.Context, podInformer coreinformers.PodInformer, ...) (*Controller, error)
  • func (jm *Controller) Run(ctx context.Context, workers int)

Constants

const (
	// "key"
	OrphanPodKeyKindName orphanPodKeyKind = iota
	// "selector"
	OrphanPodKeyKindSelector
)

Variables

var (

	// DefaultJobApiBackOff is the default API backoff period. Exported for tests.
	DefaultJobApiBackOff = time.Second
	// MaxJobApiBackOff is the max API backoff period. Exported for tests.
	MaxJobApiBackOff = time.Minute
	// DefaultJobPodFailureBackOff is the default pod failure backoff period. Exported for tests.
	DefaultJobPodFailureBackOff = 10 * time.Second
	// MaxJobPodFailureBackOff is the max  pod failure backoff period. Exported for tests.
	MaxJobPodFailureBackOff = 10 * time.Minute
	// MaxUncountedPods is the maximum size the slices in
	// .status.uncountedTerminatedPods should have to keep their representation
	// roughly below 20 KB. Exported for tests
	MaxUncountedPods = 500
	// MaxPodCreateDeletePerSync is the maximum number of pods that can be
	// created or deleted in a single sync call. Exported for tests.
	MaxPodCreateDeletePerSync = 500
)

Types

type Controller

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

Controller ensures that all Job objects have corresponding pods to run their configured workload.

func NewController

func NewController(ctx context.Context, podInformer coreinformers.PodInformer, jobInformer batchinformers.JobInformer, kubeClient clientset.Interface) (*Controller, error)

NewController creates a new Job controller that keeps the relevant pods in sync with their corresponding Job objects.

func (*Controller) Run

func (jm *Controller) Run(ctx context.Context, workers int)

Run the main goroutine responsible for watching and syncing jobs.