backupcontroller

Documentation

Index

• func NewBackupController(livenessChecker *health.MultiAlivenessChecker, ...) factory.Controller
• func NewBackupRemovalController(livenessChecker *health.MultiAlivenessChecker, ...) factory.Controller
• type BackupController
• type BackupRemovalController

Functions

func NewBackupController

func NewBackupController(
	livenessChecker *health.MultiAlivenessChecker,
	backupsClient operatorv1alpha1client.EtcdBackupsGetter,
	kubeClient kubernetes.Interface,
	eventRecorder events.Recorder,
	operatorImagePullSpec string,
	accessor featuregates.FeatureGateAccess,
	backupInformer factory.Informer,
	jobInformer factory.Informer) factory.Controller

func NewBackupRemovalController

func NewBackupRemovalController(
	livenessChecker *health.MultiAlivenessChecker,
	backupsClient operatorv1alpha1client.EtcdBackupsGetter,
	kubeClient kubernetes.Interface,
	eventRecorder events.Recorder,
	accessor featuregates.FeatureGateAccess,
	backupInformer factory.Informer,
	jobInformer factory.Informer) factory.Controller

Types

type BackupController

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

type BackupRemovalController

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

BackupRemovalController is necessary to clean up dangling etcdbackups that have no existing owner refs anymore. This scenario arises from periodically triggered backups, which are not automatically garbage collected by kube-controller-manager. The reason is that etcdbackups are global, whereas the Job/Pods that create them are namespaced. This is a known and documented limitation, which requires us to build this garbage collection below.