controller

Documentation

Index

• type Configuration
  • func NewConfiguration() *Configuration
  • func (c *Configuration) AddFlags(fs *pflag.FlagSet)
• type MachineController
  • func NewMachineController(config *Configuration) *MachineController
  • func (c *MachineController) Run() error
• type NodeWatcher
  • func NewNodeWatcher(kubeconfig string) (*NodeWatcher, error)
  • func (c *NodeWatcher) Run() error

Types

type Configuration

type Configuration struct {
	Kubeconfig   string
	Cloud        string
	KubeadmToken string
}

func NewConfiguration

func NewConfiguration() *Configuration

func (*Configuration) AddFlags

func (c *Configuration) AddFlags(fs *pflag.FlagSet)

type MachineController

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

func NewMachineController

func NewMachineController(config *Configuration) *MachineController

func (*MachineController) Run

func (c *MachineController) Run() error

type NodeWatcher

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

NodeWatcher watches updates to core "Node" objects and takes action on their events. Currently, its only responsibility is to find nodes that have a "machine" annotation and link the corresponding "Machine" object to them.

The "machine" annotation is an implementation detail of how the two objects can get linked together, but it is not required behavior. However, in the event that a Machine.Spec update requires replacing the Node, this can allow for faster turn-around time by allowing a new Node to be created with a new name while the old node is being deleted.

Currently, these annotations are added by the node itself as part of its bootup script after "kubeadm join" succeeds.

func NewNodeWatcher

func NewNodeWatcher(kubeconfig string) (*NodeWatcher, error)

func (*NodeWatcher) Run

func (c *NodeWatcher) Run() error