Documentation ¶
Overview ¶
Package kubelet is the package that contains the libraries that drive the Kubelet binary. The kubelet is responsible for node level pod management. It runs on each worker in the cluster.
Index ¶
- Constants
- Variables
- func CapacityFromMachineInfo(info *cadvisorApi.MachineInfo) api.ResourceList
- func GetValidatedSources(sources []string) ([]string, error)
- func ListenAndServeKubeletReadOnlyServer(host HostInterface, address net.IP, port uint)
- func ListenAndServeKubeletServer(host HostInterface, address net.IP, port uint, tlsOptions *TLSOptions, ...)
- func ResolvePort(portReference util.IntOrString, container *api.Container) (int, error)
- func SetupCapabilities(allowPrivileged bool, hostNetworkSources []string)
- type ContainerGCPolicy
- type HostInterface
- type ImageGCPolicy
- type Kubelet
- func (kl *Kubelet) BirthCry()
- func (kl *Kubelet) ExecInContainer(podFullName string, podUID types.UID, containerName string, cmd []string, ...) error
- func (kl *Kubelet) GetCachedMachineInfo() (*cadvisorApi.MachineInfo, error)
- func (kl *Kubelet) GetContainerInfo(podFullName string, uid types.UID, containerName string, ...) (*cadvisorApi.ContainerInfo, error)
- func (kl *Kubelet) GetDockerVersion() ([]uint, error)
- func (kl *Kubelet) GetHostIP() (net.IP, error)
- func (kl *Kubelet) GetHostname() string
- func (kl *Kubelet) GetKubeletContainerLogs(podFullName, containerName, tail string, follow bool, stdout, stderr io.Writer) error
- func (kl *Kubelet) GetNode() (*api.Node, error)
- func (kl *Kubelet) GetPodByFullName(podFullName string) (*api.Pod, bool)
- func (kl *Kubelet) GetPodByName(namespace, name string) (*api.Pod, bool)
- func (kl *Kubelet) GetPodStatus(podFullName string) (api.PodStatus, error)
- func (kl *Kubelet) GetPods() []api.Pod
- func (kl *Kubelet) GetRootInfo(req *cadvisorApi.ContainerInfoRequest) (*cadvisorApi.ContainerInfo, error)
- func (kl *Kubelet) ListenAndServe(address net.IP, port uint, tlsOptions *TLSOptions, ...)
- func (kl *Kubelet) ListenAndServeReadOnly(address net.IP, port uint)
- func (kl *Kubelet) PortForward(podFullName string, podUID types.UID, port uint16, stream io.ReadWriteCloser) error
- func (kl *Kubelet) Run(updates <-chan PodUpdate)
- func (kl *Kubelet) RunInContainer(podFullName string, podUID types.UID, containerName string, cmd []string) ([]byte, error)
- func (kl *Kubelet) RunOnce(updates <-chan PodUpdate) ([]RunPodResult, error)
- func (kl *Kubelet) ServeLogs(w http.ResponseWriter, req *http.Request)
- func (kl *Kubelet) StartGarbageCollection()
- func (kl *Kubelet) StreamingConnectionIdleTimeout() time.Duration
- func (kl *Kubelet) SyncPods(allPods []api.Pod, podSyncTypes map[types.UID]metrics.SyncPodType, ...) error
- type PodOperation
- type PodUpdate
- type RunPodResult
- type Server
- type SourcesReadyFn
- type SyncHandler
- type TLSOptions
Constants ¶
const ( RunOnceManifestDelay = 1 * time.Second RunOnceMaxRetries = 10 RunOnceRetryDelay = 1 * time.Second RunOnceRetryDelayBackoff = 2 )
const ( // This is the current pod configuration SET PodOperation = iota // Pods with the given ids are new to this source ADD // Pods with the given ids have been removed from this source REMOVE // Pods with the given ids have been updated in this source UPDATE // These constants identify the sources of pods // Updates from a file FileSource = "file" // Updates from querying a web page HTTPSource = "http" // Updates from Kubernetes API Server ApiserverSource = "api" // Updates from all sources AllSource = "*" // Used for ConfigMirrorAnnotationKey. MirrorType = "mirror" NamespaceDefault = api.NamespaceDefault )
const ConfigMirrorAnnotationKey = "kubernetes.io/config.mirror"
const ConfigSourceAnnotationKey = "kubernetes.io/config.source"
const (
PodInfraContainerImage = "gcr.io/google_containers/pause:0.8.0"
)
Variables ¶
var ( // ErrNoKubeletContainers returned when there are not containers managed by // the kubelet (ie: either no containers on the node, or none that the kubelet cares about). ErrNoKubeletContainers = errors.New("no containers managed by kubelet") // ErrContainerNotFound returned when a container in the given pod with the // given container name was not found, amongst those managed by the kubelet. ErrContainerNotFound = errors.New("no matching container") )
Functions ¶
func CapacityFromMachineInfo ¶ added in v0.4.2
func CapacityFromMachineInfo(info *cadvisorApi.MachineInfo) api.ResourceList
func GetValidatedSources ¶ added in v0.4.3
Gets all validated sources from the specified sources.
func ListenAndServeKubeletReadOnlyServer ¶ added in v0.4.4
func ListenAndServeKubeletReadOnlyServer(host HostInterface, address net.IP, port uint)
ListenAndServeKubeletReadOnlyServer initializes a server to respond to HTTP network requests on the Kubelet.
func ListenAndServeKubeletServer ¶
func ListenAndServeKubeletServer(host HostInterface, address net.IP, port uint, tlsOptions *TLSOptions, enableDebuggingHandlers bool)
ListenAndServeKubeletServer initializes a server to respond to HTTP network requests on the Kubelet.
func ResolvePort ¶
ResolvePort attempts to turn a IntOrString port reference into a concrete port number. If portReference has an int value, it is treated as a literal, and simply returns that value. If portReference is a string, an attempt is first made to parse it as an integer. If that fails, an attempt is made to find a port with the same name in the container spec. If a port with the same name is found, it's ContainerPort value is returned. If no matching port is found, an error is returned.
func SetupCapabilities ¶ added in v0.2.1
TODO: move this into pkg/capabilities
Types ¶
type ContainerGCPolicy ¶ added in v0.4.2
type ContainerGCPolicy struct { // Minimum age at which a container can be garbage collected, zero for no limit. MinAge time.Duration // Max number of dead containers any single pod (UID, container name) pair is // allowed to have, less than zero for no limit. MaxPerPodContainer int // Max number of total dead containers, less than zero for no limit. MaxContainers int }
Specified a policy for garbage collecting containers.
type HostInterface ¶
type HostInterface interface { GetContainerInfo(podFullName string, uid types.UID, containerName string, req *cadvisorApi.ContainerInfoRequest) (*cadvisorApi.ContainerInfo, error) GetRootInfo(req *cadvisorApi.ContainerInfoRequest) (*cadvisorApi.ContainerInfo, error) GetDockerVersion() ([]uint, error) GetCachedMachineInfo() (*cadvisorApi.MachineInfo, error) GetPods() []api.Pod GetPodByName(namespace, name string) (*api.Pod, bool) GetPodStatus(name string) (api.PodStatus, error) RunInContainer(name string, uid types.UID, container string, cmd []string) ([]byte, error) ExecInContainer(name string, uid types.UID, container string, cmd []string, in io.Reader, out, err io.WriteCloser, tty bool) error GetKubeletContainerLogs(podFullName, containerName, tail string, follow bool, stdout, stderr io.Writer) error ServeLogs(w http.ResponseWriter, req *http.Request) PortForward(name string, uid types.UID, port uint16, stream io.ReadWriteCloser) error StreamingConnectionIdleTimeout() time.Duration GetHostname() string }
HostInterface contains all the kubelet methods required by the server. For testablitiy.
type ImageGCPolicy ¶ added in v0.4.2
type ImageGCPolicy struct { // Any usage above this threshold will always trigger garbage collection. // This is the highest usage we will allow. HighThresholdPercent int // Any usage below this threshold will never trigger garbage collection. // This is the lowest threshold we will try to garbage collect to. LowThresholdPercent int }
A policy for garbage collecting images. Policy defines an allowed band in which garbage collection will be run.
type Kubelet ¶
type Kubelet struct {
// contains filtered or unexported fields
}
Kubelet is the main kubelet implementation.
func NewMainKubelet ¶
func NewMainKubelet( hostname string, dockerClient dockertools.DockerInterface, kubeClient client.Interface, rootDirectory string, podInfraContainerImage string, resyncInterval time.Duration, pullQPS float32, pullBurst int, containerGCPolicy ContainerGCPolicy, sourcesReady SourcesReadyFn, clusterDomain string, clusterDNS net.IP, masterServiceNamespace string, volumePlugins []volume.VolumePlugin, networkPlugins []network.NetworkPlugin, networkPluginName string, streamingConnectionIdleTimeout time.Duration, recorder record.EventRecorder, cadvisorInterface cadvisor.Interface, imageGCPolicy ImageGCPolicy, cloud cloudprovider.Interface, nodeStatusUpdateFrequency time.Duration) (*Kubelet, error)
New creates a new Kubelet for use in main
func (*Kubelet) BirthCry ¶ added in v0.2.1
func (kl *Kubelet) BirthCry()
BirthCry sends an event that the kubelet has started up.
func (*Kubelet) ExecInContainer ¶ added in v0.4.1
func (kl *Kubelet) ExecInContainer(podFullName string, podUID types.UID, containerName string, cmd []string, stdin io.Reader, stdout, stderr io.WriteCloser, tty bool) error
ExecInContainer executes a command in a container, connecting the supplied stdin/stdout/stderr to the command's IO streams.
func (*Kubelet) GetCachedMachineInfo ¶ added in v0.4.2
func (kl *Kubelet) GetCachedMachineInfo() (*cadvisorApi.MachineInfo, error)
GetCachedMachineInfo assumes that the machine info can't change without a reboot
func (*Kubelet) GetContainerInfo ¶
func (kl *Kubelet) GetContainerInfo(podFullName string, uid types.UID, containerName string, req *cadvisorApi.ContainerInfoRequest) (*cadvisorApi.ContainerInfo, error)
GetContainerInfo returns stats (from Cadvisor) for a container.
func (*Kubelet) GetDockerVersion ¶ added in v0.3.2
Returns Docker version for this Kubelet.
func (*Kubelet) GetHostname ¶ added in v0.4.2
GetHostname Returns the hostname as the kubelet sees it.
func (*Kubelet) GetKubeletContainerLogs ¶
func (kl *Kubelet) GetKubeletContainerLogs(podFullName, containerName, tail string, follow bool, stdout, stderr io.Writer) error
GetKubeletContainerLogs returns logs from the container TODO: this method is returning logs of random container attempts, when it should be returning the most recent attempt or all of them.
func (*Kubelet) GetPodByFullName ¶ added in v0.4.2
func (*Kubelet) GetPodByName ¶ added in v0.2.1
GetPodByName provides the first pod that matches namespace and name, as well as whether the pod was found.
func (*Kubelet) GetPodStatus ¶ added in v0.2.2
GetPodStatus returns information from Docker about the containers in a pod
func (*Kubelet) GetPods ¶ added in v0.4.2
GetPods returns all pods bound to the kubelet and their spec, and the mirror pods.
func (*Kubelet) GetRootInfo ¶
func (kl *Kubelet) GetRootInfo(req *cadvisorApi.ContainerInfoRequest) (*cadvisorApi.ContainerInfo, error)
GetRootInfo returns stats (from Cadvisor) of current machine (root container).
func (*Kubelet) ListenAndServe ¶ added in v0.4.4
func (*Kubelet) ListenAndServeReadOnly ¶ added in v0.4.4
func (*Kubelet) PortForward ¶ added in v0.4.1
func (kl *Kubelet) PortForward(podFullName string, podUID types.UID, port uint16, stream io.ReadWriteCloser) error
PortForward connects to the pod's port and copies data between the port and the stream.
func (*Kubelet) RunInContainer ¶
func (kl *Kubelet) RunInContainer(podFullName string, podUID types.UID, containerName string, cmd []string) ([]byte, error)
Run a command in a container, returns the combined stdout, stderr as an array of bytes
func (*Kubelet) RunOnce ¶
func (kl *Kubelet) RunOnce(updates <-chan PodUpdate) ([]RunPodResult, error)
RunOnce polls from one configuration update and run the associated pods.
func (*Kubelet) ServeLogs ¶
func (kl *Kubelet) ServeLogs(w http.ResponseWriter, req *http.Request)
Returns logs of current machine.
func (*Kubelet) StartGarbageCollection ¶ added in v0.4.2
func (kl *Kubelet) StartGarbageCollection()
Starts garbage collection theads.
func (*Kubelet) StreamingConnectionIdleTimeout ¶ added in v0.4.1
type PodOperation ¶
type PodOperation int
PodOperation defines what changes will be made on a pod configuration.
type PodUpdate ¶
type PodUpdate struct { Pods []api.Pod Op PodOperation Source string }
PodUpdate defines an operation sent on the channel. You can add or remove single services by sending an array of size one and Op == ADD|REMOVE (with REMOVE, only the ID is required). For setting the state of the system to a given state for this source configuration, set Pods as desired and Op to SET, which will reset the system state to that specified in this operation for this source channel. To remove all pods, set Pods to empty object and Op to SET.
Additionally, Pods should never be nil - it should always point to an empty slice. While functionally similar, this helps our unit tests properly check that the correct PodUpdates are generated.
type RunPodResult ¶
type Server ¶
type Server struct {
// contains filtered or unexported fields
}
Server is a http.Handler which exposes kubelet functionality over HTTP.
func NewServer ¶
func NewServer(host HostInterface, enableDebuggingHandlers bool) Server
NewServer initializes and configures a kubelet.Server object to handle HTTP requests.
func (*Server) InstallDebuggingHandlers ¶
func (s *Server) InstallDebuggingHandlers()
InstallDeguggingHandlers registers the HTTP request patterns that serve logs or run commands/containers
func (*Server) InstallDefaultHandlers ¶
func (s *Server) InstallDefaultHandlers()
InstallDefaultHandlers registers the default set of supported HTTP request patterns with the mux.
type SourcesReadyFn ¶ added in v0.4.2
type SourcesReadyFn func() bool
type SyncHandler ¶
type SyncHandler interface { // Syncs current state to match the specified pods. SyncPodType specified what // type of sync is occuring per pod. StartTime specifies the time at which // syncing began (for use in monitoring). SyncPods(pods []api.Pod, podSyncTypes map[types.UID]metrics.SyncPodType, mirrorPods map[string]api.Pod, startTime time.Time) error }
SyncHandler is an interface implemented by Kubelet, for testability
Source Files ¶
Directories ¶
Path | Synopsis |
---|---|
Kubelet interactions with cAdvisor.
|
Kubelet interactions with cAdvisor. |
Reads the pod configuration from the Kubernetes apiserver.
|
Reads the pod configuration from the Kubernetes apiserver. |
Package envvars is the package that build the environment variables that kubernetes provides to the containers run by it.
|
Package envvars is the package that build the environment variables that kubernetes provides to the containers run by it. |
Package leaky holds bits of kubelet that should be internal but have leaked out through bad abstractions.
|
Package leaky holds bits of kubelet that should be internal but have leaked out through bad abstractions. |
exec
Package exec scans and loads networking plugins that are installed under /usr/libexec/kubernetes/kubelet-plugins/net/exec/ The layout convention for a plugin is: plugin-name/ (plugins have to be directories first) plugin-name/plugin-name (executable that will be called out, see Vendoring Note for more nuances) plugin-name/<other-files> where, 'executable' has the following requirements: - should have exec permissions - should give non-zero exit code on failure, and zero on sucess - the arguments will be <action> <pod_namespace> <pod_name> <docker_id_of_infra_container> whereupon, <action> will be one of: - init, called when the kubelet loads the plugin - setup, called after the infra container of a pod is created, but before other containers of the pod are created - teardown, called before the pod infra container is killed As the executables are called, the file-descriptors stdin, stdout, stderr remain open.
|
Package exec scans and loads networking plugins that are installed under /usr/libexec/kubernetes/kubelet-plugins/net/exec/ The layout convention for a plugin is: plugin-name/ (plugins have to be directories first) plugin-name/plugin-name (executable that will be called out, see Vendoring Note for more nuances) plugin-name/<other-files> where, 'executable' has the following requirements: - should have exec permissions - should give non-zero exit code on failure, and zero on sucess - the arguments will be <action> <pod_namespace> <pod_name> <docker_id_of_infra_container> whereupon, <action> will be one of: - init, called when the kubelet loads the plugin - setup, called after the infra container of a pod is created, but before other containers of the pod are created - teardown, called before the pod infra container is killed As the executables are called, the file-descriptors stdin, stdout, stderr remain open. |