replication

Documentation

Overview

Package replication contains logic for watching and synchronizing replication controllers.

Index

• Constants
• func GetCondition(status api.ReplicationControllerStatus, ...) *api.ReplicationControllerCondition
• func NewReplicationControllerCondition(condType api.ReplicationControllerConditionType, status api.ConditionStatus, ...) api.ReplicationControllerCondition
• func RemoveCondition(status *api.ReplicationControllerStatus, ...)
• func SetCondition(status *api.ReplicationControllerStatus, ...)
• type OverlappingControllers
  • func (o OverlappingControllers) Len() int
  • func (o OverlappingControllers) Less(i, j int) bool
  • func (o OverlappingControllers) Swap(i, j int)
• type ReplicationManager
  • func NewReplicationManager(podInformer cache.SharedIndexInformer, kubeClient clientset.Interface, ...) *ReplicationManager
  • func NewReplicationManagerFromClient(kubeClient clientset.Interface, resyncPeriod controller.ResyncPeriodFunc, ...) *ReplicationManager
  • func NewReplicationManagerFromClientForIntegration(kubeClient clientset.Interface, resyncPeriod controller.ResyncPeriodFunc, ...) *ReplicationManager
  • func (rm *ReplicationManager) Run(workers int, stopCh <-chan struct{})
  • func (rm *ReplicationManager) SetEventRecorder(recorder record.EventRecorder)

Constants

const (
	// We'll attempt to recompute the required replicas of all replication controllers
	// that have fulfilled their expectations at least this often. This recomputation
	// happens based on contents in local pod storage.
	// Full Resync shouldn't be needed at all in a healthy system. This is a protection
	// against disappearing objects and watch notification, that we believe should not
	// happen at all.
	// TODO: We should get rid of it completely in the fullness of time.
	FullControllerResyncPeriod = 10 * time.Minute

	// Realistic value of the burstReplica field for the replication manager based off
	// performance requirements for kubernetes 1.0.
	BurstReplicas = 500

	// We must avoid counting pods until the pod store has synced. If it hasn't synced, to
	// avoid a hot loop, we'll wait this long between checks.
	PodStoreSyncedPollPeriod = 100 * time.Millisecond
)

Functions

func GetCondition

func GetCondition(status api.ReplicationControllerStatus, condType api.ReplicationControllerConditionType) *api.ReplicationControllerCondition

GetCondition returns a replication controller condition with the provided type if it exists.

func NewReplicationControllerCondition

func NewReplicationControllerCondition(condType api.ReplicationControllerConditionType, status api.ConditionStatus, reason, msg string) api.ReplicationControllerCondition

NewReplicationControllerCondition creates a new replication controller condition.

func RemoveCondition

func RemoveCondition(status *api.ReplicationControllerStatus, condType api.ReplicationControllerConditionType)

RemoveCondition removes the condition with the provided type from the replication controller status.

func SetCondition

func SetCondition(status *api.ReplicationControllerStatus, condition api.ReplicationControllerCondition)

SetCondition adds/replaces the given condition in the replication controller status.

Types

type OverlappingControllers

type OverlappingControllers []*api.ReplicationController

OverlappingControllers sorts a list of controllers by creation timestamp, using their names as a tie breaker.

func (OverlappingControllers) Len

func (o OverlappingControllers) Len() int

func (OverlappingControllers) Less

func (o OverlappingControllers) Less(i, j int) bool

func (OverlappingControllers) Swap

func (o OverlappingControllers) Swap(i, j int)

type ReplicationManager

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

ReplicationManager is responsible for synchronizing ReplicationController objects stored in the system with actual running pods. TODO: this really should be called ReplicationController. The only reason why it's a Manager is to distinguish this type from API object "ReplicationController". We should fix this.

func NewReplicationManager

func NewReplicationManager(podInformer cache.SharedIndexInformer, kubeClient clientset.Interface, resyncPeriod controller.ResyncPeriodFunc, burstReplicas int, lookupCacheSize int, garbageCollectorEnabled bool) *ReplicationManager

NewReplicationManager creates a replication manager

func NewReplicationManagerFromClient

func NewReplicationManagerFromClient(kubeClient clientset.Interface, resyncPeriod controller.ResyncPeriodFunc, burstReplicas int, lookupCacheSize int) *ReplicationManager

NewReplicationManagerFromClient creates a new ReplicationManager that runs its own informer.

func NewReplicationManagerFromClientForIntegration

func NewReplicationManagerFromClientForIntegration(kubeClient clientset.Interface, resyncPeriod controller.ResyncPeriodFunc, burstReplicas int, lookupCacheSize int) *ReplicationManager

NewReplicationManagerFromClientForIntegration creates a new ReplicationManager that runs its own informer. It disables event recording for use in integration tests.

func (*ReplicationManager) Run

func (rm *ReplicationManager) Run(workers int, stopCh <-chan struct{})

Run begins watching and syncing.

func (*ReplicationManager) SetEventRecorder

func (rm *ReplicationManager) SetEventRecorder(recorder record.EventRecorder)

SetEventRecorder replaces the event recorder used by the replication manager with the given recorder. Only used for testing.