libpod

package
v3.4.7 Latest Latest
Warning

This package is not in the latest version of its module.

Go to latest
Published: Apr 20, 2022 License: Apache-2.0 Imports: 133 Imported by: 16

Documentation

Overview

+linux

Index

Constants

View Source
const (
	// MaxHealthCheckNumberLogs is the maximum number of attempts we keep
	// in the healthcheck history file
	MaxHealthCheckNumberLogs int = 5
	// MaxHealthCheckLogLength in characters
	MaxHealthCheckLogLength = 500
)
View Source
const (
	// MountPrivate represents the private mount option.
	MountPrivate = "private"
	// MountRPrivate represents the rprivate mount option.
	MountRPrivate = "rprivate"
	// MountShared represents the shared mount option.
	MountShared = "shared"
	// MountRShared represents the rshared mount option.
	MountRShared = "rshared"
	// MountSlave represents the slave mount option.
	MountSlave = "slave"
	// MountRSlave represents the rslave mount option.
	MountRSlave = "rslave"
)
View Source
const (
	AttachPipeStdin  = 1
	AttachPipeStdout = 2
	AttachPipeStderr = 3
)

Sync with stdpipe_t in conmon.c

View Source
const CgroupfsDefaultCgroupParent = "/libpod_parent"

CgroupfsDefaultCgroupParent is the cgroup parent for CGroupFS in libpod

View Source
const DefaultWaitInterval = 250 * time.Millisecond

DefaultWaitInterval is the default interval between container status checks while waiting.

View Source
const SystemdDefaultCgroupParent = "machine.slice"

SystemdDefaultCgroupParent is the cgroup parent for the systemd cgroup manager in libpod

View Source
const SystemdDefaultRootlessCgroupParent = "user.slice"

SystemdDefaultRootlessCgroupParent is the cgroup parent for the systemd cgroup manager in libpod when running as rootless

Variables

View Source
var (
	// SdNotifyModeValues describes the only values that SdNotifyMode can be
	SdNotifyModeValues = []string{define.SdNotifyModeContainer, define.SdNotifyModeConmon, define.SdNotifyModeIgnore}
)

Runtime Creation Options

Functions

func DefaultSeccompPath

func DefaultSeccompPath() (string, error)

DefaultSeccompPath returns the path to the default seccomp.json file if it exists, first it checks OverrideSeccomp and then default. If neither exist function returns ""

func DownloadFromFile

func DownloadFromFile(reader *os.File) (string, error)

DownloadFromFile reads all of the content from the reader and temporarily saves in it $TMPDIR/importxyz, which is deleted after the image is imported

func FuncTimer

func FuncTimer(funcName string)

FuncTimer helps measure the execution time of a function For debug purposes, do not leave in code used like defer FuncTimer("foo")

func GenerateForKube

func GenerateForKube(ctx context.Context, ctrs []*Container) (*v1.Pod, error)

GenerateForKube takes a slice of libpod containers and generates one v1.Pod description that includes just a single container.

func GetSlirp4netnsDNS added in v3.2.0

func GetSlirp4netnsDNS(subnet *net.IPNet) (*net.IP, error)

Get expected slirp DNS ipv4 address based on subnet Reference: https://github.com/rootless-containers/slirp4netns/blob/master/slirp4netns.1.md#description

func GetSlirp4netnsGateway added in v3.2.0

func GetSlirp4netnsGateway(subnet *net.IPNet) (*net.IP, error)

Get expected slirp Gateway ipv4 address based on subnet Reference: https://github.com/rootless-containers/slirp4netns/blob/master/slirp4netns.1.md#description

func GetSlirp4netnsIP added in v3.2.0

func GetSlirp4netnsIP(subnet *net.IPNet) (*net.IP, error)

Get expected slirp ipv4 address based on subnet. If subnet is null use default subnet Reference: https://github.com/rootless-containers/slirp4netns/blob/master/slirp4netns.1.md#description

func JSONDeepCopy

func JSONDeepCopy(from, to interface{}) error

JSONDeepCopy performs a deep copy by performing a JSON encode/decode of the given structures. From and To should be identically typed structs.

func LabelVolumePath

func LabelVolumePath(path string) error

LabelVolumePath takes a mount path for a volume and gives it an selinux label of either shared or not

func MountExists

func MountExists(specMounts []spec.Mount, dest string) bool

MountExists returns true if dest exists in the list of mounts

func SetXdgDirs

func SetXdgDirs() error

SetXdgDirs ensures the XDG_RUNTIME_DIR env and XDG_CONFIG_HOME variables are set. containers/image uses XDG_RUNTIME_DIR to locate the auth file, XDG_CONFIG_HOME is use for the containers.conf configuration file.

func Unmount

func Unmount(mount string)

Unmount umounts a target directory

func WaitForFile

func WaitForFile(path string, chWait chan error, timeout time.Duration) (bool, error)

WaitForFile waits until a file has been created or the given timeout has occurred

func WithInfraContainerPorts

func WithInfraContainerPorts(bindings []ocicni.PortMapping, infraSpec *specgen.SpecGenerator) []netTypes.PortMapping

WithInfraContainerPorts tells the pod to add port bindings to the pause container

Types

type BoltState

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

BoltState is a state implementation backed by a Bolt DB

func (*BoltState) AddContainer

func (s *BoltState) AddContainer(ctr *Container) error

AddContainer adds a container to the state The container being added cannot belong to a pod

func (*BoltState) AddContainerToPod

func (s *BoltState) AddContainerToPod(pod *Pod, ctr *Container) error

AddContainerToPod adds the given container to an existing pod The container will be added to the state and the pod

func (*BoltState) AddExecSession

func (s *BoltState) AddExecSession(ctr *Container, session *ExecSession) error

AddExecSession adds an exec session to the state.

func (*BoltState) AddPod

func (s *BoltState) AddPod(pod *Pod) error

AddPod adds the given pod to the state.

func (*BoltState) AddVolume

func (s *BoltState) AddVolume(volume *Volume) error

AddVolume adds the given volume to the state. It also adds ctrDepID to the sub bucket holding the container dependencies that this volume has

func (*BoltState) AllContainers

func (s *BoltState) AllContainers() ([]*Container, error)

AllContainers retrieves all the containers in the database

func (*BoltState) AllPods

func (s *BoltState) AllPods() ([]*Pod, error)

AllPods returns all pods present in the state

func (*BoltState) AllVolumes

func (s *BoltState) AllVolumes() ([]*Volume, error)

AllVolumes returns all volumes present in the state

func (*BoltState) Close

func (s *BoltState) Close() error

Close closes the state and prevents further use

func (*BoltState) Container

func (s *BoltState) Container(id string) (*Container, error)

Container retrieves a single container from the state by its full ID

func (*BoltState) ContainerInUse

func (s *BoltState) ContainerInUse(ctr *Container) ([]string, error)

ContainerInUse checks if other containers depend on the given container It returns a slice of the IDs of the containers depending on the given container. If the slice is empty, no containers depend on the given container

func (*BoltState) GetAllNetworkAliases

func (s *BoltState) GetAllNetworkAliases(ctr *Container) (map[string][]string, error)

GetAllNetworkAliases retrieves the network aliases for the given container in all CNI networks.

func (*BoltState) GetContainerConfig

func (s *BoltState) GetContainerConfig(id string) (*ContainerConfig, error)

GetContainerConfig returns a container config from the database by full ID

func (*BoltState) GetContainerExecSessions

func (s *BoltState) GetContainerExecSessions(ctr *Container) ([]string, error)

GetContainerExecSessions retrieves the IDs of all exec sessions running in a container that the database is aware of (IE, were added via AddExecSession).

func (*BoltState) GetDBConfig

func (s *BoltState) GetDBConfig() (*DBConfig, error)

GetDBConfig retrieves runtime configuration fields that were created when the database was first initialized

func (*BoltState) GetExecSession

func (s *BoltState) GetExecSession(id string) (string, error)

GetExecSession returns the ID of the container an exec session is associated with.

func (*BoltState) GetName

func (s *BoltState) GetName(id string) (string, error)

GetName returns the name associated with a given ID. Since IDs are globally unique, it works for both containers and pods. Returns ErrNoSuchCtr if the ID does not exist.

func (*BoltState) GetNetworkAliases

func (s *BoltState) GetNetworkAliases(ctr *Container, network string) ([]string, error)

GetNetworkAliases retrieves the network aliases for the given container in the given CNI network.

func (*BoltState) GetNetworks

func (s *BoltState) GetNetworks(ctr *Container) ([]string, error)

GetNetworks returns the CNI networks this container is a part of.

func (*BoltState) HasContainer

func (s *BoltState) HasContainer(id string) (bool, error)

HasContainer checks if a container is present in the state

func (*BoltState) HasPod

func (s *BoltState) HasPod(id string) (bool, error)

HasPod checks if a pod with the given ID exists in the state

func (*BoltState) HasVolume

func (s *BoltState) HasVolume(name string) (bool, error)

HasVolume returns true if the given volume exists in the state, otherwise it returns false

func (*BoltState) LookupContainer

func (s *BoltState) LookupContainer(idOrName string) (*Container, error)

LookupContainer retrieves a container from the state by full or unique partial ID or name

func (*BoltState) LookupContainerID

func (s *BoltState) LookupContainerID(idOrName string) (string, error)

LookupContainerID retrieves a container ID from the state by full or unique partial ID or name

func (*BoltState) LookupPod

func (s *BoltState) LookupPod(idOrName string) (*Pod, error)

LookupPod retrieves a pod from full or unique partial ID or name

func (*BoltState) LookupVolume

func (s *BoltState) LookupVolume(name string) (*Volume, error)

LookupVolume locates a volume from a partial name.

func (*BoltState) NetworkConnect

func (s *BoltState) NetworkConnect(ctr *Container, network string, aliases []string) error

NetworkConnect adds the given container to the given network. If aliases are specified, those will be added to the given network.

func (*BoltState) NetworkDisconnect

func (s *BoltState) NetworkDisconnect(ctr *Container, network string) error

NetworkDisconnect disconnects the container from the given network, also removing any aliases in the network.

func (*BoltState) Pod

func (s *BoltState) Pod(id string) (*Pod, error)

Pod retrieves a pod given its full ID

func (*BoltState) PodContainers

func (s *BoltState) PodContainers(pod *Pod) ([]*Container, error)

PodContainers returns all the containers present in the given pod

func (*BoltState) PodContainersByID

func (s *BoltState) PodContainersByID(pod *Pod) ([]string, error)

PodContainersByID returns the IDs of all containers present in the given pod

func (*BoltState) PodHasContainer

func (s *BoltState) PodHasContainer(pod *Pod, id string) (bool, error)

PodHasContainer checks if the given pod has a container with the given ID

func (*BoltState) Refresh

func (s *BoltState) Refresh() error

Refresh clears container and pod states after a reboot

func (*BoltState) RemoveContainer

func (s *BoltState) RemoveContainer(ctr *Container) error

RemoveContainer removes a container from the state Only removes containers not in pods - for containers that are a member of a pod, use RemoveContainerFromPod

func (*BoltState) RemoveContainerExecSessions

func (s *BoltState) RemoveContainerExecSessions(ctr *Container) error

RemoveContainerExecSessions removes all exec sessions attached to a given container.

func (*BoltState) RemoveContainerFromPod

func (s *BoltState) RemoveContainerFromPod(pod *Pod, ctr *Container) error

RemoveContainerFromPod removes a container from an existing pod The container will also be removed from the state

func (*BoltState) RemoveExecSession

func (s *BoltState) RemoveExecSession(session *ExecSession) error

RemoveExecSession removes references to the given exec session in the database.

func (*BoltState) RemovePod

func (s *BoltState) RemovePod(pod *Pod) error

RemovePod removes the given pod from the state Only empty pods can be removed

func (*BoltState) RemovePodContainers

func (s *BoltState) RemovePodContainers(pod *Pod) error

RemovePodContainers removes all containers in a pod

func (*BoltState) RemoveVolume

func (s *BoltState) RemoveVolume(volume *Volume) error

RemoveVolume removes the given volume from the state

func (*BoltState) RewriteContainerConfig

func (s *BoltState) RewriteContainerConfig(ctr *Container, newCfg *ContainerConfig) error

RewriteContainerConfig rewrites a container's configuration. WARNING: This function is DANGEROUS. Do not use without reading the full comment on this function in state.go.

func (*BoltState) RewritePodConfig

func (s *BoltState) RewritePodConfig(pod *Pod, newCfg *PodConfig) error

RewritePodConfig rewrites a pod's configuration. WARNING: This function is DANGEROUS. Do not use without reading the full comment on this function in state.go.

func (*BoltState) RewriteVolumeConfig

func (s *BoltState) RewriteVolumeConfig(volume *Volume, newCfg *VolumeConfig) error

RewriteVolumeConfig rewrites a volume's configuration. WARNING: This function is DANGEROUS. Do not use without reading the full comment on this function in state.go.

func (*BoltState) SafeRewriteContainerConfig added in v3.1.0

func (s *BoltState) SafeRewriteContainerConfig(ctr *Container, oldName, newName string, newCfg *ContainerConfig) error

SafeRewriteContainerConfig rewrites a container's configuration in a more limited fashion than RewriteContainerConfig. It is marked as safe to use under most circumstances, unlike RewriteContainerConfig. DO NOT USE TO: Change container dependencies, change pod membership, change locks, change container ID.

func (*BoltState) SaveContainer

func (s *BoltState) SaveContainer(ctr *Container) error

SaveContainer saves a container's current state in the database

func (*BoltState) SavePod

func (s *BoltState) SavePod(pod *Pod) error

SavePod saves a pod's state to the database

func (*BoltState) SaveVolume

func (s *BoltState) SaveVolume(volume *Volume) error

SaveVolume saves the volume's state to the database.

func (*BoltState) SetNamespace

func (s *BoltState) SetNamespace(ns string) error

SetNamespace sets the namespace that will be used for container and pod retrieval

func (*BoltState) UpdateContainer

func (s *BoltState) UpdateContainer(ctr *Container) error

UpdateContainer updates a container's state from the database

func (*BoltState) UpdatePod

func (s *BoltState) UpdatePod(pod *Pod) error

UpdatePod updates a pod's state from the database

func (*BoltState) UpdateVolume

func (s *BoltState) UpdateVolume(volume *Volume) error

UpdateVolume updates the volume's state from the database.

func (*BoltState) ValidateDBConfig

func (s *BoltState) ValidateDBConfig(runtime *Runtime) error

ValidateDBConfig validates paths in the given runtime against the database

func (*BoltState) Volume

func (s *BoltState) Volume(name string) (*Volume, error)

Volume retrieves a volume from full name

func (*BoltState) VolumeInUse

func (s *BoltState) VolumeInUse(volume *Volume) ([]string, error)

VolumeInUse checks if any container is using the volume It returns a slice of the IDs of the containers using the given volume. If the slice is empty, no containers use the given volume

type ConmonOCIRuntime

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

ConmonOCIRuntime is an OCI runtime managed by Conmon. TODO: Make all calls to OCI runtime have a timeout.

func (*ConmonOCIRuntime) AttachResize

func (r *ConmonOCIRuntime) AttachResize(ctr *Container, newSize define.TerminalSize) error

AttachResize resizes the terminal used by the given container.

func (*ConmonOCIRuntime) AttachSocketPath

func (r *ConmonOCIRuntime) AttachSocketPath(ctr *Container) (string, error)

AttachSocketPath is the path to a single container's attach socket.

func (*ConmonOCIRuntime) CheckConmonRunning

func (r *ConmonOCIRuntime) CheckConmonRunning(ctr *Container) (bool, error)

func (*ConmonOCIRuntime) CheckpointContainer

func (r *ConmonOCIRuntime) CheckpointContainer(ctr *Container, options ContainerCheckpointOptions) error

CheckpointContainer checkpoints the given container.

func (*ConmonOCIRuntime) CreateContainer

func (r *ConmonOCIRuntime) CreateContainer(ctr *Container, restoreOptions *ContainerCheckpointOptions) error

CreateContainer creates a container.

func (*ConmonOCIRuntime) DeleteContainer

func (r *ConmonOCIRuntime) DeleteContainer(ctr *Container) error

DeleteContainer deletes a container from the OCI runtime.

func (*ConmonOCIRuntime) ExecAttachResize

func (r *ConmonOCIRuntime) ExecAttachResize(ctr *Container, sessionID string, newSize define.TerminalSize) error

ExecAttachResize resizes the TTY of the given exec session.

func (*ConmonOCIRuntime) ExecAttachSocketPath

func (r *ConmonOCIRuntime) ExecAttachSocketPath(ctr *Container, sessionID string) (string, error)

ExecAttachSocketPath is the path to a container's exec session attach socket.

func (*ConmonOCIRuntime) ExecContainer

func (r *ConmonOCIRuntime) ExecContainer(c *Container, sessionID string, options *ExecOptions, streams *define.AttachStreams, newSize *define.TerminalSize) (int, chan error, error)

ExecContainer executes a command in a running container

func (*ConmonOCIRuntime) ExecContainerDetached

func (r *ConmonOCIRuntime) ExecContainerDetached(ctr *Container, sessionID string, options *ExecOptions, stdin bool) (int, error)

ExecContainerDetached executes a command in a running container, but does not attach to it.

func (*ConmonOCIRuntime) ExecContainerHTTP

func (r *ConmonOCIRuntime) ExecContainerHTTP(ctr *Container, sessionID string, options *ExecOptions, req *http.Request, w http.ResponseWriter,
	streams *HTTPAttachStreams, cancel <-chan bool, hijackDone chan<- bool, holdConnOpen <-chan bool, newSize *define.TerminalSize) (int, chan error, error)

ExecContainerHTTP executes a new command in an existing container and forwards its standard streams over an attach

func (*ConmonOCIRuntime) ExecStopContainer

func (r *ConmonOCIRuntime) ExecStopContainer(ctr *Container, sessionID string, timeout uint) error

ExecStopContainer stops a given exec session in a running container.

func (*ConmonOCIRuntime) ExecUpdateStatus

func (r *ConmonOCIRuntime) ExecUpdateStatus(ctr *Container, sessionID string) (bool, error)

ExecUpdateStatus checks if the given exec session is still running.

func (*ConmonOCIRuntime) ExitFilePath

func (r *ConmonOCIRuntime) ExitFilePath(ctr *Container) (string, error)

ExitFilePath is the path to a container's exit file.

func (*ConmonOCIRuntime) HTTPAttach

func (r *ConmonOCIRuntime) HTTPAttach(ctr *Container, req *http.Request, w http.ResponseWriter, streams *HTTPAttachStreams, detachKeys *string, cancel <-chan bool, hijackDone chan<- bool, streamAttach, streamLogs bool) (deferredErr error)

HTTPAttach performs an attach for the HTTP API. The caller must handle closing the HTTP connection after this returns. The cancel channel is not closed; it is up to the caller to do so after this function returns. If this is a container with a terminal, we will stream raw. If it is not, we will stream with an 8-byte header to multiplex STDOUT and STDERR. Returns any errors that occurred, and whether the connection was successfully hijacked before that error occurred.

func (*ConmonOCIRuntime) KillContainer

func (r *ConmonOCIRuntime) KillContainer(ctr *Container, signal uint, all bool) error

KillContainer sends the given signal to the given container. If all is set, send to all PIDs in the container. All is only supported if the container created cgroups.

func (*ConmonOCIRuntime) Name

func (r *ConmonOCIRuntime) Name() string

Name returns the name of the runtime being wrapped by Conmon.

func (*ConmonOCIRuntime) Path

func (r *ConmonOCIRuntime) Path() string

Path returns the path of the OCI runtime being wrapped by Conmon.

func (*ConmonOCIRuntime) PauseContainer

func (r *ConmonOCIRuntime) PauseContainer(ctr *Container) error

PauseContainer pauses the given container.

func (*ConmonOCIRuntime) RuntimeInfo

RuntimeInfo provides information on the runtime.

func (*ConmonOCIRuntime) StartContainer

func (r *ConmonOCIRuntime) StartContainer(ctr *Container) error

StartContainer starts the given container. Sets time the container was started, but does not save it.

func (*ConmonOCIRuntime) StopContainer

func (r *ConmonOCIRuntime) StopContainer(ctr *Container, timeout uint, all bool) error

StopContainer stops a container, first using its given stop signal (or SIGTERM if no signal was specified), then using SIGKILL. Timeout is given in seconds. If timeout is 0, the container will be immediately kill with SIGKILL. Does not set finished time for container, assumes you will run updateStatus after to pull the exit code.

func (*ConmonOCIRuntime) SupportsCheckpoint

func (r *ConmonOCIRuntime) SupportsCheckpoint() bool

SupportsCheckpoint checks if the OCI runtime supports checkpointing containers.

func (*ConmonOCIRuntime) SupportsJSONErrors

func (r *ConmonOCIRuntime) SupportsJSONErrors() bool

SupportsJSONErrors checks if the OCI runtime supports JSON-formatted error messages.

func (*ConmonOCIRuntime) SupportsKVM

func (r *ConmonOCIRuntime) SupportsKVM() bool

SupportsKVM checks if the OCI runtime supports running containers without KVM separation

func (*ConmonOCIRuntime) SupportsNoCgroups

func (r *ConmonOCIRuntime) SupportsNoCgroups() bool

SupportsNoCgroups checks if the OCI runtime supports running containers without cgroups (the --cgroup-manager=disabled flag).

func (*ConmonOCIRuntime) UnpauseContainer

func (r *ConmonOCIRuntime) UnpauseContainer(ctr *Container) error

UnpauseContainer unpauses the given container.

func (*ConmonOCIRuntime) UpdateContainerStatus

func (r *ConmonOCIRuntime) UpdateContainerStatus(ctr *Container) error

UpdateContainerStatus retrieves the current status of the container from the runtime. It updates the container's state but does not save it. If useRuntime is false, we will not directly hit runc to see the container's status, but will instead only check for the existence of the conmon exit file and update state to stopped if it exists.

type Container

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

Container is a single OCI container. All operations on a Container that access state must begin with a call to syncContainer(). There is no guarantee that state exists in a readable state before syncContainer() is run, and even if it does, its contents will be out of date and must be refreshed from the database. Generally, this requirement applies only to top-level functions; helpers can assume that their callers handled this requirement. Generally speaking, if a function takes the container lock and accesses any part of state, it should syncContainer() immediately after locking.

func (*Container) AddArtifact

func (c *Container) AddArtifact(name string, data []byte) error

AddArtifact creates and writes to an artifact file for the container

func (*Container) Attach

func (c *Container) Attach(streams *define.AttachStreams, keys string, resize <-chan define.TerminalSize) error

Attach attaches to a container. This function returns when the attach finishes. It does not hold the lock for the duration of its runtime, only using it at the beginning to verify state.

func (*Container) AttachResize

func (c *Container) AttachResize(newSize define.TerminalSize) error

AttachResize resizes the container's terminal, which is displayed by Attach and HTTPAttach.

func (*Container) AttachSocketPath

func (c *Container) AttachSocketPath() (string, error)

AttachSocketPath retrieves the path of the container's attach socket

func (*Container) AutoRemove

func (c *Container) AutoRemove() bool

AutoRemove indicates whether the container will be removed after it is executed

func (*Container) Batch

func (c *Container) Batch(batchFunc func(*Container) error) error

Batch starts a batch operation on the given container All commands in the passed function will execute under the same lock and without synchronizing state after each operation This will result in substantial performance benefits when running numerous commands on the same container Note that the container passed into the Batch function cannot be removed during batched operations. runtime.RemoveContainer can only be called outside of Batch Any error returned by the given batch function will be returned unmodified by Batch As Batch normally disables updating the current state of the container, the Sync() function is provided to enable container state to be updated and checked within Batch.

func (*Container) BindMounts

func (c *Container) BindMounts() (map[string]string, error)

BindMounts retrieves bind mounts that were created by libpod and will be added to the container All these mounts except /dev/shm are ignored if a mount in the given spec has the same destination These mounts include /etc/resolv.conf, /etc/hosts, and /etc/hostname The return is formatted as a map from destination (mountpoint in the container) to source (path of the file that will be mounted into the container) If the container has not been started yet, an empty map will be returned, as the files in question are only created when the container is started.

func (*Container) CGroupPath

func (c *Container) CGroupPath() (string, error)

CGroupPath returns a cgroups "path" for the given container. Note that the container must be running. Otherwise, an error is returned.

func (*Container) CgroupManager

func (c *Container) CgroupManager() string

CgroupManager returns the cgroup manager used by the given container.

func (*Container) CgroupParent

func (c *Container) CgroupParent() string

CgroupParent gets the container's CGroup parent

func (*Container) Checkpoint

func (c *Container) Checkpoint(ctx context.Context, options ContainerCheckpointOptions) error

Checkpoint checkpoints a container

func (*Container) CheckpointPath

func (c *Container) CheckpointPath() string

CheckpointPath returns the path to the directory containing the checkpoint

func (*Container) Cleanup

func (c *Container) Cleanup(ctx context.Context) error

Cleanup unmounts all mount points in container and cleans up container storage It also cleans up the network stack

func (*Container) Command

func (c *Container) Command() []string

Command is the container's command This is not added to the spec, but is instead used during image commit

func (*Container) Commit

func (c *Container) Commit(ctx context.Context, destImage string, options ContainerCommitOptions) (*libimage.Image, error)

Commit commits the changes between a container and its image, creating a new image

func (*Container) Config

func (c *Container) Config() *ContainerConfig

Config returns the configuration used to create the container

func (*Container) ConmonPID

func (c *Container) ConmonPID() (int, error)

ConmonPID Returns the PID of the container's conmon process. If the container is not running, a PID of 0 will be returned. No error will occur.

func (*Container) ContainerState

func (c *Container) ContainerState() (*ContainerState, error)

ContainerState returns containerstate struct

func (*Container) ControlSocketPath

func (c *Container) ControlSocketPath() string

ControlSocketPath returns the path to the containers control socket for things like tty resizing

func (*Container) CopyFromArchive added in v3.1.0

func (c *Container) CopyFromArchive(ctx context.Context, containerPath string, chown bool, rename map[string]string, tarStream io.Reader) (func() error, error)

CopyFromArchive copies the contents from the specified tarStream to path *inside* the container.

func (*Container) CopyToArchive added in v3.1.0

func (c *Container) CopyToArchive(ctx context.Context, containerPath string, tarStream io.Writer) (func() error, error)

CopyToArchive copies the contents from the specified path *inside* the container to the tarStream.

func (*Container) CreatedTime

func (c *Container) CreatedTime() time.Time

CreatedTime gets the time when the container was created

func (*Container) DNSOption

func (c *Container) DNSOption() []string

DNSOption returns the DNS options that will be used in the container's resolv.conf If empty, options from the host's resolv.conf will be used instead

func (*Container) DNSSearch

func (c *Container) DNSSearch() []string

DNSSearch returns the DNS search domains that will be used in the container's resolv.conf If empty, DNS Search domains from the host's resolv.conf will be used instead

func (*Container) DNSServers

func (c *Container) DNSServers() []net.IP

DNSServers returns DNS servers that will be used in the container's resolv.conf If empty, DNS server from the host's resolv.conf will be used instead

func (*Container) Dependencies

func (c *Container) Dependencies() []string

Dependencies gets the containers this container depends upon

func (*Container) Entrypoint

func (c *Container) Entrypoint() []string

Entrypoint is the container's entrypoint. This is not added to the spec, but is instead used during image commit.

func (*Container) Exec

func (c *Container) Exec(config *ExecConfig, streams *define.AttachStreams, resize <-chan define.TerminalSize) (int, error)

Exec emulates the old Libpod exec API, providing a single call to create, run, and remove an exec session. Returns exit code and error. Exit code is not guaranteed to be set sanely if error is not nil.

func (*Container) ExecCleanup

func (c *Container) ExecCleanup(sessionID string) error

ExecCleanup cleans up an exec session in the container, removing temporary files associated with it.

func (*Container) ExecCreate

func (c *Container) ExecCreate(config *ExecConfig) (string, error)

ExecCreate creates a new exec session for the container. The session is not started. The ID of the new exec session will be returned.

func (*Container) ExecHTTPStartAndAttach

func (c *Container) ExecHTTPStartAndAttach(sessionID string, r *http.Request, w http.ResponseWriter,
	streams *HTTPAttachStreams, detachKeys *string, cancel <-chan bool, hijackDone chan<- bool, newSize *define.TerminalSize) error

ExecHTTPStartAndAttach starts and performs an HTTP attach to an exec session. newSize resizes the tty to this size before the process is started, must be nil if the exec session has no tty

func (*Container) ExecRemove

func (c *Container) ExecRemove(sessionID string, force bool) error

ExecRemove removes an exec session in the container. If force is given, the session will be stopped first if it is running.

func (*Container) ExecResize

func (c *Container) ExecResize(sessionID string, newSize define.TerminalSize) error

ExecResize resizes the TTY of the given exec session. Only available if the exec session created a TTY.

func (*Container) ExecSession

func (c *Container) ExecSession(id string) (*ExecSession, error)

ExecSession retrieves detailed information on a single active exec session in a container

func (*Container) ExecSessions

func (c *Container) ExecSessions() ([]string, error)

ExecSessions retrieves active exec sessions running in the container

func (*Container) ExecStart

func (c *Container) ExecStart(sessionID string) error

ExecStart starts an exec session in the container, but does not attach to it. Returns immediately upon starting the exec session, unlike other ExecStart functions, which will only return when the exec session exits.

func (*Container) ExecStartAndAttach

func (c *Container) ExecStartAndAttach(sessionID string, streams *define.AttachStreams, newSize *define.TerminalSize) error

ExecStartAndAttach starts and attaches to an exec session in a container. newSize resizes the tty to this size before the process is started, must be nil if the exec session has no tty TODO: Should we include detach keys in the signature to allow override? TODO: How do we handle AttachStdin/AttachStdout/AttachStderr?

func (*Container) ExecStop

func (c *Container) ExecStop(sessionID string, timeout *uint) error

ExecStop stops an exec session in the container. If a timeout is provided, it will be used; otherwise, the timeout will default to the stop timeout of the container. Cleanup will be invoked automatically once the session is stopped.

func (*Container) ExitCode

func (c *Container) ExitCode() (int32, bool, error)

ExitCode returns the exit code of the container as an int32, and whether the container has exited. If the container has not exited, exit code will always be 0. If the container restarts, the exit code is reset to 0.

func (*Container) Export

func (c *Container) Export(path string) error

Export exports a container's root filesystem as a tar archive The archive will be saved as a file at the given path

func (*Container) FinishedTime

func (c *Container) FinishedTime() (time.Time, error)

FinishedTime is the time the container was stopped

func (*Container) GetArtifact

func (c *Container) GetArtifact(name string) ([]byte, error)

GetArtifact reads the specified artifact file from the container

func (*Container) GetContainerPidInformation

func (c *Container) GetContainerPidInformation(descriptors []string) ([]string, error)

GetContainerPidInformation returns process-related data of all processes in the container. The output data can be controlled via the `descriptors` argument which expects format descriptors and supports all AIXformat descriptors of ps (1) plus some additional ones to for instance inspect the set of effective capabilities. Each element in the returned string slice is a tab-separated string.

For more details, please refer to github.com/containers/psgo.

func (*Container) GetContainerStats

func (c *Container) GetContainerStats(previousStats *define.ContainerStats) (*define.ContainerStats, error)

GetContainerStats gets the running stats for a given container

func (*Container) HTTPAttach

func (c *Container) HTTPAttach(r *http.Request, w http.ResponseWriter, streams *HTTPAttachStreams, detachKeys *string, cancel <-chan bool, streamAttach, streamLogs bool, hijackDone chan<- bool) error

HTTPAttach forwards an attach session over a hijacked HTTP session. HTTPAttach will consume and close the included httpCon, which is expected to be sourced from a hijacked HTTP connection. The cancel channel is optional, and can be used to asynchronously cancel the attach session. The streams variable is only supported if the container was not a terminal, and allows specifying which of the container's standard streams will be forwarded to the client. This function returns when the attach finishes. It does not hold the lock for the duration of its runtime, only using it at the beginning to verify state. The streamLogs parameter indicates that all the container's logs until present will be streamed at the beginning of the attach. The streamAttach parameter indicates that the attach itself will be streamed over the socket; if this is not set, but streamLogs is, only the logs will be sent. At least one of streamAttach and streamLogs must be set.

func (*Container) HasHealthCheck

func (c *Container) HasHealthCheck() bool

HasHealthCheck returns bool as to whether there is a health check defined for the container

func (*Container) HealthCheckConfig

func (c *Container) HealthCheckConfig() *manifest.Schema2HealthConfig

HealthCheckConfig returns the command and timing attributes of the health check

func (*Container) HealthCheckStatus

func (c *Container) HealthCheckStatus() (string, error)

HealthCheckStatus returns the current state of a container with a healthcheck

func (*Container) HostNetwork

func (c *Container) HostNetwork() bool

func (*Container) Hostname

func (c *Container) Hostname() string

Hostname gets the container's hostname

func (*Container) HostsAdd

func (c *Container) HostsAdd() []string

HostsAdd returns hosts that will be added to the container's hosts file The host system's hosts file is used as a base, and these are appended to it

func (*Container) ID

func (c *Container) ID() string

ID returns the container's ID

func (*Container) IDMappings

func (c *Container) IDMappings() storage.IDMappingOptions

IDMappings returns the UID/GID mapping used for the container

func (*Container) IPs

func (c *Container) IPs() ([]net.IPNet, error)

IPs retrieves a container's IP address(es) This will only be populated if the container is configured to created a new network namespace, and that namespace is presently active

func (*Container) Image

func (c *Container) Image() (string, string)

Image returns the ID and name of the image used as the container's rootfs.

func (*Container) Init

func (c *Container) Init(ctx context.Context, recursive bool) error

Init creates a container in the OCI runtime, moving a container from ContainerStateConfigured, ContainerStateStopped, or ContainerStateExited to ContainerStateCreated. Once in Created state, Conmon will be running, which allows the container to be attached to. The container can subsequently transition to ContainerStateRunning via Start(), or be transitioned back to ContainerStateConfigured by Cleanup() (which will stop conmon and unmount the container). Init requires that all dependency containers be started (e.g. pod infra containers). The `recursive` parameter will, if set to true, start these dependency containers before initializing this container.

func (*Container) Inspect

func (c *Container) Inspect(size bool) (*define.InspectContainerData, error)

Inspect a container for low-level information

func (*Container) IsInfra

func (c *Container) IsInfra() bool

IsInfra returns whether the container is an infra container

func (*Container) IsInitCtr added in v3.4.0

func (c *Container) IsInitCtr() bool

IsInitCtr returns whether the container is an init container

func (*Container) IsReadOnly

func (c *Container) IsReadOnly() bool

IsReadOnly returns whether the container is running in read only mode

func (*Container) Kill

func (c *Container) Kill(signal uint) error

Kill sends a signal to a container

func (*Container) Labels

func (c *Container) Labels() map[string]string

Labels returns the container's labels

func (*Container) LogDriver

func (c *Container) LogDriver() string

LogDriver returns the log driver for this container

func (*Container) LogPath

func (c *Container) LogPath() string

LogPath returns the path to the container's log file This file will only be present after Init() is called to create the container in the runtime

func (*Container) LogTag

func (c *Container) LogTag() string

LogTag returns the tag to the container's log file

func (*Container) Mount

func (c *Container) Mount() (string, error)

Mount mounts a container's filesystem on the host The path where the container has been mounted is returned

func (*Container) MountLabel

func (c *Container) MountLabel() string

MountLabel returns the SELinux mount label of the container

func (*Container) Mounted

func (c *Container) Mounted() (bool, string, error)

Mounted returns whether the container is mounted and the path it is mounted at (if it is mounted). If the container is not mounted, no error is returned, and the mountpoint will be set to "".

func (*Container) Name

func (c *Container) Name() string

Name returns the container's name

func (*Container) NamedVolumes

func (c *Container) NamedVolumes() []*ContainerNamedVolume

NamedVolumes returns the container's named volumes. The name of each is guaranteed to point to a valid libpod Volume present in the state.

func (*Container) Namespace

func (c *Container) Namespace() string

Namespace returns the libpod namespace the container is in. Namespaces are used to logically separate containers and pods in the state.

func (*Container) NamespacePath

func (c *Container) NamespacePath(linuxNS LinuxNS) (string, error)

NamespacePath returns the path of one of the container's namespaces If the container is not running, an error will be returned

func (*Container) NetworkConnect

func (c *Container) NetworkConnect(nameOrID, netName string, aliases []string) error

ConnectNetwork connects a container to a given network

func (*Container) NetworkDisabled

func (c *Container) NetworkDisabled() (bool, error)

NetworkDisabled returns whether the container is running with a disabled network

func (*Container) NetworkDisconnect

func (c *Container) NetworkDisconnect(nameOrID, netName string, force bool) error

NetworkDisconnect removes a container from the network

func (*Container) NetworkMode added in v3.3.0

func (c *Container) NetworkMode() string

NetworkMode gets the configured network mode for the container. Get actual value from the database

func (*Container) Networks

func (c *Container) Networks() ([]string, bool, error)

Networks gets all the networks this container is connected to. Please do NOT use ctr.config.Networks, as this can be changed from those values at runtime via network connect and disconnect. If the container is configured to use CNI and this function returns an empty array, the container will still be connected to the default network. The second return parameter, a bool, indicates that the container container is joining the default CNI network - the network name will be included in the returned array of network names, but the container did not explicitly join this network.

func (*Container) NewNetNS

func (c *Container) NewNetNS() bool

NewNetNS returns whether the container will create a new network namespace

func (*Container) OOMKilled

func (c *Container) OOMKilled() (bool, error)

OOMKilled returns whether the container was killed by an OOM condition

func (*Container) PID

func (c *Container) PID() (int, error)

PID returns the PID of the container. If the container is not running, a pid of 0 will be returned. No error will occur.

func (*Container) Pause

func (c *Container) Pause() error

Pause pauses a container

func (*Container) PodID

func (c *Container) PodID() string

PodID returns the full ID of the pod the container belongs to, or "" if it does not belong to a pod

func (*Container) PortMappings

func (c *Container) PortMappings() ([]ocicni.PortMapping, error)

PortMappings returns the ports that will be mapped into a container if a new network namespace is created If NewNetNS() is false, this value is unused

func (*Container) PreCheckPointPath

func (c *Container) PreCheckPointPath() string

PreCheckpointPath returns the path to the directory containing the pre-checkpoint-images

func (*Container) Privileged

func (c *Container) Privileged() bool

Privileged returns whether the container is privileged

func (*Container) ProcessLabel

func (c *Container) ProcessLabel() string

ProcessLabel returns the selinux ProcessLabel of the container

func (*Container) RWSize

func (c *Container) RWSize() (int64, error)

RWSize returns the rw size of the container

func (*Container) RawImageName

func (c *Container) RawImageName() string

RawImageName returns the unprocessed and not-normalized user-specified image name.

func (*Container) ReadLog

func (c *Container) ReadLog(ctx context.Context, options *logs.LogOptions, logChannel chan *logs.LogLine) error

ReadLog reads a containers log based on the input options and returns log lines over a channel.

func (*Container) Refresh

func (c *Container) Refresh(ctx context.Context) error

Refresh is DEPRECATED and REMOVED.

func (*Container) ReloadNetwork

func (c *Container) ReloadNetwork() error

ReloadNetwork reconfigures the container's network. Technically speaking, it will tear down and then reconfigure the container's network namespace, which will result in all firewall rules being recreated. It is mostly intended to be used in cases where the system firewall has been reloaded, and existing rules have been wiped out. It is expected that some downtime will result, as the rules are destroyed as part of this process. At present, this only works on root containers; it may be expanded to restart slirp4netns in the future to work with rootless containers as well. Requires that the container must be running or created.

func (*Container) RemoveArtifact

func (c *Container) RemoveArtifact(name string) error

RemoveArtifact deletes the specified artifacts file

func (*Container) RestartPolicy

func (c *Container) RestartPolicy() string

RestartPolicy returns the container's restart policy.

func (*Container) RestartRetries

func (c *Container) RestartRetries() uint

RestartRetries returns the number of retries that will be attempted when using the "on-failure" restart policy

func (*Container) RestartWithTimeout

func (c *Container) RestartWithTimeout(ctx context.Context, timeout uint) error

RestartWithTimeout restarts a running container and takes a given timeout in uint

func (*Container) Restore

func (c *Container) Restore(ctx context.Context, options ContainerCheckpointOptions) error

Restore restores a container

func (*Container) RootFsSize

func (c *Container) RootFsSize() (int64, error)

RootFsSize returns the root FS size of the container

func (*Container) RootGID

func (c *Container) RootGID() int

RootGID returns the root user mapping from container

func (*Container) RootUID

func (c *Container) RootUID() int

RootUID returns the root user mapping from container

func (*Container) Routes

func (c *Container) Routes() ([]types.Route, error)

Routes retrieves a container's routes This will only be populated if the container is configured to created a new network namespace, and that namespace is presently active

func (*Container) Runtime

func (c *Container) Runtime() *Runtime

Runtime returns the container's Runtime.

func (*Container) RuntimeName

func (c *Container) RuntimeName() string

RuntimeName returns the name of the runtime

func (*Container) Secrets added in v3.1.0

func (c *Container) Secrets() []*ContainerSecret

Secrets return the secrets in the container

func (*Container) ShmDir

func (c *Container) ShmDir() string

ShmDir returns the sources path to be mounted on /dev/shm in container

func (*Container) ShmSize

func (c *Container) ShmSize() int64

ShmSize returns the size of SHM device to be mounted into the container

func (*Container) ShouldRestart

func (c *Container) ShouldRestart(ctx context.Context) bool

Indicate whether or not the container should restart

func (*Container) Spec

func (c *Container) Spec() *spec.Spec

Spec returns the container's OCI runtime spec The spec returned is the one used to create the container. The running spec may differ slightly as mounts are added based on the image

func (*Container) Start

func (c *Container) Start(ctx context.Context, recursive bool) error

Start starts the given container. Start will accept container in ContainerStateConfigured, ContainerStateCreated, ContainerStateStopped, and ContainerStateExited, and transition them to ContainerStateRunning (all containers not in ContainerStateCreated will make an intermediate stop there via the Init API). Once in ContainerStateRunning, the container can be transitioned to ContainerStatePaused via Pause(), or to ContainerStateStopped by the process stopping (either due to exit, or being forced to stop by the Kill or Stop API calls). Start requites that all dependency containers (e.g. pod infra containers) be running before being run. The recursive parameter, if set, will start all dependencies before starting this container.

func (*Container) StartAndAttach

func (c *Container) StartAndAttach(ctx context.Context, streams *define.AttachStreams, keys string, resize <-chan define.TerminalSize, recursive bool) (<-chan error, error)

StartAndAttach starts a container and attaches to it. This acts as a combination of the Start and Attach APIs, ensuring proper ordering of the two such that no output from the container is lost (e.g. the Attach call occurs before Start). In overall functionality, it is identical to the Start call, with the added side effect that an attach session will also be started.

func (*Container) StartedTime

func (c *Container) StartedTime() (time.Time, error)

StartedTime is the time the container was started

func (*Container) Stat added in v3.1.0

func (c *Container) Stat(ctx context.Context, containerPath string) (*define.FileInfo, error)

Stat the specified path *inside* the container and return a file info.

func (*Container) State

func (c *Container) State() (define.ContainerStatus, error)

State returns the current state of the container

func (*Container) StaticDir

func (c *Container) StaticDir() string

StaticDir returns the directory used to store persistent container files

func (*Container) Stdin

func (c *Container) Stdin() bool

Stdin returns whether STDIN on the container will be kept open

func (*Container) Stop

func (c *Container) Stop() error

Stop uses the container's stop signal (or SIGTERM if no signal was specified) to stop the container, and if it has not stopped after container's stop timeout, SIGKILL is used to attempt to forcibly stop the container Default stop timeout is 10 seconds, but can be overridden when the container is created

func (*Container) StopSignal

func (c *Container) StopSignal() uint

StopSignal is the signal that will be used to stop the container If it fails to stop the container, SIGKILL will be used after a timeout If StopSignal is 0, the default signal of SIGTERM will be used

func (*Container) StopTimeout

func (c *Container) StopTimeout() uint

StopTimeout returns the container's stop timeout If the container's default stop signal fails to kill the container, SIGKILL will be used after this timeout

func (*Container) StopWithTimeout

func (c *Container) StopWithTimeout(timeout uint) error

StopWithTimeout is a version of Stop that allows a timeout to be specified manually. If timeout is 0, SIGKILL will be used immediately to kill the container.

func (*Container) StoppedByUser

func (c *Container) StoppedByUser() (bool, error)

StoppedByUser returns whether the container was last stopped by an explicit call to the Stop() API, or whether it exited naturally.

func (*Container) Sync

func (c *Container) Sync() error

Sync updates the status of a container by querying the OCI runtime. If the container has not been created inside the OCI runtime, nothing will be done. Most of the time, Podman does not explicitly query the OCI runtime for container status, and instead relies upon exit files created by conmon. This can cause a disconnect between running state and what Podman sees in cases where Conmon was killed unexpected, or runc was upgraded. Running a manual Sync() ensures that container state will be correct in such situations.

func (*Container) Systemd

func (c *Container) Systemd() bool

Systemd returns whether the container will be running in systemd mode

func (*Container) Timezone

func (c *Container) Timezone() string

Timezone returns the timezone configured inside the container. Local means it has the same timezone as the host machine

func (*Container) Top

func (c *Container) Top(descriptors []string) ([]string, error)

Top gathers statistics about the running processes in a container. It returns a []string for output

func (*Container) Umask

func (c *Container) Umask() string

Umask returns the Umask bits configured inside the container.

func (*Container) Unmount

func (c *Container) Unmount(force bool) error

Unmount unmounts a container's filesystem on the host

func (*Container) Unpause

func (c *Container) Unpause() error

Unpause unpauses a container

func (*Container) User

func (c *Container) User() string

User returns the user who the container is run as

func (*Container) UserVolumes

func (c *Container) UserVolumes() []string

UserVolumes returns user-added volume mounts in the container. These are not added to the spec, but are used during image commit and to trigger some OCI hooks.

func (*Container) Wait

func (c *Container) Wait(ctx context.Context) (int32, error)

Wait blocks until the container exits and returns its exit code.

func (*Container) WaitForConditionWithInterval

func (c *Container) WaitForConditionWithInterval(ctx context.Context, waitTimeout time.Duration, conditions ...define.ContainerStatus) (int32, error)

func (*Container) WaitWithInterval

func (c *Container) WaitWithInterval(ctx context.Context, waitTimeout time.Duration) (int32, error)

WaitWithInterval blocks until the container to exit and returns its exit code. The argument is the interval at which checks the container's status.

func (*Container) WorkingDir

func (c *Container) WorkingDir() string

WorkingDir returns the containers working dir

type ContainerCheckpointOptions

type ContainerCheckpointOptions struct {
	// Keep tells the API to not delete checkpoint artifacts
	Keep bool
	// KeepRunning tells the API to keep the container running
	// after writing the checkpoint to disk
	KeepRunning bool
	// TCPEstablished tells the API to checkpoint a container
	// even if it contains established TCP connections
	TCPEstablished bool
	// TargetFile tells the API to read (or write) the checkpoint image
	// from (or to) the filename set in TargetFile
	TargetFile string
	// Name tells the API that during restore from an exported
	// checkpoint archive a new name should be used for the
	// restored container
	Name string
	// IgnoreRootfs tells the API to not export changes to
	// the container's root file-system (or to not import)
	IgnoreRootfs bool
	// IgnoreStaticIP tells the API to ignore the IP set
	// during 'podman run' with '--ip'. This is especially
	// important to be able to restore a container multiple
	// times with '--import --name'.
	IgnoreStaticIP bool
	// IgnoreStaticMAC tells the API to ignore the MAC set
	// during 'podman run' with '--mac-address'. This is especially
	// important to be able to restore a container multiple
	// times with '--import --name'.
	IgnoreStaticMAC bool
	// IgnoreVolumes tells the API to not export or not to import
	// the content of volumes associated with the container
	IgnoreVolumes bool
	// Pre Checkpoint container and leave container running
	PreCheckPoint bool
	// Dump container with Pre Checkpoint images
	WithPrevious bool
	// ImportPrevious tells the API to restore container with two
	// images. One is TargetFile, the other is ImportPrevious.
	ImportPrevious string
	// Compression tells the API which compression to use for
	// the exported checkpoint archive.
	Compression archive.Compression
	// If Pod is set the container should be restored into the
	// given Pod. If Pod is empty it is a restore without a Pod.
	// Restoring a non Pod container into a Pod or a Pod container
	// without a Pod is theoretically possible, but will
	// probably not work if a PID namespace is shared.
	// A shared PID namespace means that a Pod container has PID 1
	// in the infrastructure container, but without the infrastructure
	// container no PID 1 will be in the namespace and that is not
	// possible.
	Pod string
}

ContainerCheckpointOptions is a struct used to pass the parameters for checkpointing (and restoring) to the corresponding functions

type ContainerCommitOptions

type ContainerCommitOptions struct {
	buildah.CommitOptions
	Pause          bool
	IncludeVolumes bool
	Author         string
	Message        string
	Changes        []string
}

ContainerCommitOptions is a struct used to commit a container to an image It uses buildah's CommitOptions as a base. Long-term we might wish to add these to the buildah struct once buildah is more integrated with libpod

type ContainerConfig

type ContainerConfig struct {
	// Spec is OCI runtime spec used to create the container. This is passed
	// in when the container is created, but it is not the final spec used
	// to run the container - it will be modified by Libpod to add things we
	// manage (e.g. bind mounts for /etc/resolv.conf, named volumes, a
	// network namespace prepared by CNI or slirp4netns) in the
	// generateSpec() function.
	Spec *spec.Spec `json:"spec"`

	// ID is a hex-encoded 256-bit pseudorandom integer used as a unique
	// identifier for the container. IDs are globally unique in Libpod -
	// once an ID is in use, no other container or pod will be created with
	// the same one until the holder of the ID has been removed.
	// ID is generated by Libpod, and cannot be chosen or influenced by the
	// user (except when restoring a checkpointed container).
	// ID is guaranteed to be 64 characters long.
	ID string `json:"id"`

	// Name is a human-readable name for the container. All containers must
	// have a non-empty name. Name may be provided when the container is
	// created; if no name is chosen, a name will be auto-generated.
	Name string `json:"name"`

	// Pod is the full ID of the pod the container belongs to. If the
	// container does not belong to a pod, this will be empty.
	// If this is not empty, a pod with this ID is guaranteed to exist in
	// the state for the duration of this container's existence.
	Pod string `json:"pod,omitempty"`

	// Namespace is the libpod Namespace the container is in.
	// Namespaces are used to divide containers in the state.
	Namespace string `json:"namespace,omitempty"`

	// LockID is the ID of this container's lock. Each container, pod, and
	// volume is assigned a unique Lock (from one of several backends) by
	// the libpod Runtime. This lock will belong only to this container for
	// the duration of the container's lifetime.
	LockID uint32 `json:"lockID"`

	// CreateCommand is the full command plus arguments that were used to
	// create the container. It is shown in the output of Inspect, and may
	// be used to recreate an identical container for automatic updates or
	// portable systemd unit files.
	CreateCommand []string `json:"CreateCommand,omitempty"`

	// RawImageName is the raw and unprocessed name of the image when creating
	// the container (as specified by the user).  May or may not be set.  One
	// use case to store this data are auto-updates where we need the _exact_
	// name and not some normalized instance of it.
	RawImageName string `json:"RawImageName,omitempty"`

	// IDMappings are UID/GID mappings used by the container's user
	// namespace. They are used by the OCI runtime when creating the
	// container, and by c/storage to ensure that the container's files have
	// the appropriate owner.
	IDMappings storage.IDMappingOptions `json:"idMappingsOptions,omitempty"`

	// Dependencies are the IDs of dependency containers.
	// These containers must be started before this container is started.
	Dependencies []string

	// embedded sub-configs
	ContainerRootFSConfig
	ContainerSecurityConfig
	ContainerNameSpaceConfig
	ContainerNetworkConfig
	ContainerImageConfig
	ContainerMiscConfig
}

ContainerConfig contains all information that was used to create the container. It may not be changed once created. It is stored, read-only, on disk in Libpod's State. Any changes will not be written back to the database, and will cause inconsistencies with other Libpod instances.

type ContainerFilter

type ContainerFilter func(*Container) bool

ContainerFilter is a function to determine whether a container is included in command output. Containers to be outputted are tested using the function. A true return will include the container, a false return will exclude it.

type ContainerGraph

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

ContainerGraph is a dependency graph based on a set of containers.

func BuildContainerGraph

func BuildContainerGraph(ctrs []*Container) (*ContainerGraph, error)

BuildContainerGraph builds a dependency graph based on the container slice.

func (*ContainerGraph) DependencyMap

func (cg *ContainerGraph) DependencyMap() (dependencies map[*Container][]*Container)

DependencyMap returns the dependency graph as map with the key being a container and the value being the containers the key depends on.

type ContainerImageConfig

type ContainerImageConfig struct {
	// UserVolumes contains user-added volume mounts in the container.
	// These will not be added to the container's spec, as it is assumed
	// they are already present in the spec given to Libpod. Instead, it is
	// used when committing containers to generate the VOLUMES field of the
	// image that is created, and for triggering some OCI hooks which do not
	// fire unless user-added volume mounts are present.
	UserVolumes []string `json:"userVolumes,omitempty"`
	// Entrypoint is the container's entrypoint.
	// It is not used in spec generation, but will be used when the
	// container is committed to populate the entrypoint of the new image.
	Entrypoint []string `json:"entrypoint,omitempty"`
	// Command is the container's command.
	// It is not used in spec generation, but will be used when the
	// container is committed to populate the command of the new image.
	Command []string `json:"command,omitempty"`
}

ContainerImageConfig is an embedded sub-config providing image configuration to the container.

type ContainerImageVolume

type ContainerImageVolume struct {
	// Source is the source of the image volume.  The image can be referred
	// to by name and by ID.
	Source string `json:"source"`
	// Dest is the absolute path of the mount in the container.
	Dest string `json:"dest"`
	// ReadWrite sets the volume writable.
	ReadWrite bool `json:"rw"`
}

ContainerImageVolume is a volume based on a container image. The container image is first mounted on the host and is then bind-mounted into the container.

type ContainerInfo

type ContainerInfo struct {
	Dir          string
	RunDir       string
	Config       *v1.Image
	ProcessLabel string
	MountLabel   string
	UIDMap       []idtools.IDMap
	GIDMap       []idtools.IDMap
}

ContainerInfo wraps a subset of information about a container: the locations of its nonvolatile and volatile per-container directories, along with a copy of the configuration blob from the image that was used to create the container, if the image had a configuration. It also returns the ProcessLabel and MountLabel selected for the container

type ContainerMiscConfig

type ContainerMiscConfig struct {
	// Whether to keep container STDIN open
	Stdin bool `json:"stdin,omitempty"`
	// Labels is a set of key-value pairs providing additional information
	// about a container
	Labels map[string]string `json:"labels,omitempty"`
	// StopSignal is the signal that will be used to stop the container
	StopSignal uint `json:"stopSignal,omitempty"`
	// StopTimeout is the signal that will be used to stop the container
	StopTimeout uint `json:"stopTimeout,omitempty"`
	// Timeout is maximum time a container will run before getting the kill signal
	Timeout uint `json:"timeout,omitempty"`
	// Time container was created
	CreatedTime time.Time `json:"createdTime"`
	// CgroupManager is the cgroup manager used to create this container.
	// If empty, the runtime default will be used.
	CgroupManager string `json:"cgroupManager,omitempty"`
	// NoCgroups indicates that the container will not create CGroups. It is
	// incompatible with CgroupParent.  Deprecated in favor of CgroupsMode.
	NoCgroups bool `json:"noCgroups,omitempty"`
	// CgroupsMode indicates how the container will create cgroups
	// (disabled, no-conmon, enabled).  It supersedes NoCgroups.
	CgroupsMode string `json:"cgroupsMode,omitempty"`
	// Cgroup parent of the container.
	CgroupParent string `json:"cgroupParent"`
	// LogPath log location
	LogPath string `json:"logPath"`
	// LogTag is the tag used for logging
	LogTag string `json:"logTag"`
	// LogSize is the tag used for logging
	LogSize int64 `json:"logSize"`
	// LogDriver driver for logs
	LogDriver string `json:"logDriver"`
	// File containing the conmon PID
	ConmonPidFile string `json:"conmonPidFile,omitempty"`
	// RestartPolicy indicates what action the container will take upon
	// exiting naturally.
	// Allowed options are "no" (take no action), "on-failure" (restart on
	// non-zero exit code, up an a maximum of RestartRetries times),
	// and "always" (always restart the container on any exit code).
	// The empty string is treated as the default ("no")
	RestartPolicy string `json:"restart_policy,omitempty"`
	// RestartRetries indicates the number of attempts that will be made to
	// restart the container. Used only if RestartPolicy is set to
	// "on-failure".
	RestartRetries uint `json:"restart_retries,omitempty"`
	// TODO log options for log drivers
	// PostConfigureNetNS needed when a user namespace is created by an OCI runtime
	// if the network namespace is created before the user namespace it will be
	// owned by the wrong user namespace.
	PostConfigureNetNS bool `json:"postConfigureNetNS"`
	// OCIRuntime used to create the container
	OCIRuntime string `json:"runtime,omitempty"`
	// ExitCommand is the container's exit command.
	// This Command will be executed when the container exits by Conmon.
	// It is usually used to invoke post-run cleanup - for example, in
	// Podman, it invokes `podman container cleanup`, which in turn calls
	// Libpod's Cleanup() API to unmount the container and clean up its
	// network.
	ExitCommand []string `json:"exitCommand,omitempty"`
	// IsInfra is a bool indicating whether this container is an infra container used for
	// sharing kernel namespaces in a pod
	IsInfra bool `json:"pause"`
	// SdNotifyMode tells libpod what to do with a NOTIFY_SOCKET if passed
	SdNotifyMode string `json:"sdnotifyMode,omitempty"`
	// Systemd tells libpod to setup the container in systemd mode
	Systemd bool `json:"systemd"`
	// HealthCheckConfig has the health check command and related timings
	HealthCheckConfig *manifest.Schema2HealthConfig `json:"healthcheck"`
	// PreserveFDs is a number of additional file descriptors (in addition
	// to 0, 1, 2) that will be passed to the executed process. The total FDs
	// passed will be 3 + PreserveFDs.
	PreserveFDs uint `json:"preserveFds,omitempty"`
	// Timezone is the timezone inside the container.
	// Local means it has the same timezone as the host machine
	Timezone string `json:"timezone,omitempty"`
	// Umask is the umask inside the container.
	Umask string `json:"umask,omitempty"`
	// PidFile is the file that saves the pid of the container process
	PidFile string `json:"pid_file,omitempty"`
	// CDIDevices contains devices that use the CDI
	CDIDevices []string `json:"cdiDevices,omitempty"`
	// EnvSecrets are secrets that are set as environment variables
	EnvSecrets map[string]*secrets.Secret `json:"secret_env,omitempty"`
	// InitContainerType specifies if the container is an initcontainer
	// and if so, what type: always or once are possible non-nil entries
	InitContainerType string `json:"init_container_type,omitempty"`
}

ContainerMiscConfig is an embedded sub-config providing misc configuration to the container.

type ContainerNameSpaceConfig

type ContainerNameSpaceConfig struct {
	// IDs of container to share namespaces with
	// NetNsCtr conflicts with the CreateNetNS bool
	// These containers are considered dependencies of the given container
	// They must be started before the given container is started
	IPCNsCtr    string `json:"ipcNsCtr,omitempty"`
	MountNsCtr  string `json:"mountNsCtr,omitempty"`
	NetNsCtr    string `json:"netNsCtr,omitempty"`
	PIDNsCtr    string `json:"pidNsCtr,omitempty"`
	UserNsCtr   string `json:"userNsCtr,omitempty"`
	UTSNsCtr    string `json:"utsNsCtr,omitempty"`
	CgroupNsCtr string `json:"cgroupNsCtr,omitempty"`
}

ContainerNameSpaceConfig is an embedded sub-config providing namespace configuration to the container.

type ContainerNamedVolume

type ContainerNamedVolume struct {
	// Name is the name of the volume to mount in.
	// Must resolve to a valid volume present in this Podman.
	Name string `json:"volumeName"`
	// Dest is the mount's destination
	Dest string `json:"dest"`
	// Options are fstab style mount options
	Options []string `json:"options,omitempty"`
}

ContainerNamedVolume is a named volume that will be mounted into the container. Each named volume is a libpod Volume present in the state.

type ContainerNetworkConfig

type ContainerNetworkConfig struct {
	// CreateNetNS indicates that libpod should create and configure a new
	// network namespace for the container.
	// This cannot be set if NetNsCtr is also set.
	CreateNetNS bool `json:"createNetNS"`
	// StaticIP is a static IP to request for the container.
	// This cannot be set unless CreateNetNS is set.
	// If not set, the container will be dynamically assigned an IP by CNI.
	StaticIP net.IP `json:"staticIP"`
	// StaticMAC is a static MAC to request for the container.
	// This cannot be set unless CreateNetNS is set.
	// If not set, the container will be dynamically assigned a MAC by CNI.
	StaticMAC net.HardwareAddr `json:"staticMAC"`
	// PortMappings are the ports forwarded to the container's network
	// namespace
	// These are not used unless CreateNetNS is true
	PortMappings []ocicni.PortMapping `json:"portMappings,omitempty"`
	// ExposedPorts are the ports which are exposed but not forwarded
	// into the container.
	// The map key is the port and the string slice contains the protocols,
	// e.g. tcp and udp
	// These are only set when exposed ports are given but not published.
	ExposedPorts map[uint16][]string `json:"exposedPorts,omitempty"`
	// UseImageResolvConf indicates that resolv.conf should not be
	// bind-mounted inside the container.
	// Conflicts with DNSServer, DNSSearch, DNSOption.
	UseImageResolvConf bool
	// DNS servers to use in container resolv.conf
	// Will override servers in host resolv if set
	DNSServer []net.IP `json:"dnsServer,omitempty"`
	// DNS Search domains to use in container resolv.conf
	// Will override search domains in host resolv if set
	DNSSearch []string `json:"dnsSearch,omitempty"`
	// DNS options to be set in container resolv.conf
	// With override options in host resolv if set
	DNSOption []string `json:"dnsOption,omitempty"`
	// UseImageHosts indicates that /etc/hosts should not be
	// bind-mounted inside the container.
	// Conflicts with HostAdd.
	UseImageHosts bool
	// Hosts to add in container
	// Will be appended to host's host file
	HostAdd []string `json:"hostsAdd,omitempty"`
	// Network names (CNI) to add container to. Empty to use default network.
	// Please note that these can be altered at runtime. The actual list is
	// stored in the DB and should be retrieved from there; this is only the
	// set of networks the container was *created* with.
	Networks []string `json:"networks,omitempty"`
	// Network mode specified for the default network.
	NetMode namespaces.NetworkMode `json:"networkMode,omitempty"`
	// NetworkOptions are additional options for each network
	NetworkOptions map[string][]string `json:"network_options,omitempty"`
	// NetworkAliases are aliases that will be added to each network.
	// These are additional names that this container can be accessed as via
	// DNS when the CNI dnsname plugin is in use.
	// Please note that these can be altered at runtime. As such, the actual
	// list is stored in the database and should be retrieved from there;
	// this is only the set of aliases the container was *created with*.
	// Formatted as map of network name to aliases. All network names must
	// be present in the Networks list above.
	NetworkAliases map[string][]string `json:"network_alises,omitempty"`
}

ContainerNetworkConfig is an embedded sub-config providing network configuration to the container.

type ContainerNetworkDescriptions

type ContainerNetworkDescriptions map[string]int

ContainerNetworkDescriptions describes the relationship between the CNI network and the ethN where N is an integer

type ContainerOverlayVolume

type ContainerOverlayVolume struct {
	// Destination is the absolute path where the mount will be placed in the container.
	Dest string `json:"dest"`
	// Source specifies the source path of the mount.
	Source string `json:"source,omitempty"`
	// Options holds overlay volume options.
	Options []string `json:"options,omitempty"`
}

ContainerOverlayVolume is a overlay volume that will be mounted into the container. Each volume is a libpod Volume present in the state.

type ContainerRootFSConfig

type ContainerRootFSConfig struct {
	// RootfsImageID is the ID of the image used to create the container.
	// If the container was created from a Rootfs, this will be empty.
	// If non-empty, Podman will create a root filesystem for the container
	// based on an image with this ID.
	// This conflicts with Rootfs.
	RootfsImageID string `json:"rootfsImageID,omitempty"`
	// RootfsImageName is the (normalized) name of the image used to create
	// the container. If the container was created from a Rootfs, this will
	// be empty.
	RootfsImageName string `json:"rootfsImageName,omitempty"`
	// Rootfs is a directory to use as the container's root filesystem.
	// If RootfsImageID is set, this will be empty.
	// If this is set, Podman will not create a root filesystem for the
	// container based on an image, and will instead use the given directory
	// as the container's root.
	// Conflicts with RootfsImageID.
	Rootfs string `json:"rootfs,omitempty"`
	// ShmDir is the path to be mounted on /dev/shm in container.
	// If not set manually at creation time, Libpod will create a tmpfs
	// with the size specified in ShmSize and populate this with the path of
	// said tmpfs.
	ShmDir string `json:"ShmDir,omitempty"`
	// ShmSize is the size of the container's SHM. Only used if ShmDir was
	// not set manually at time of creation.
	ShmSize int64 `json:"shmSize"`
	// Static directory for container content that will persist across
	// reboot.
	// StaticDir is a persistent directory for Libpod files that will
	// survive system reboot. It is not part of the container's rootfs and
	// is not mounted into the container. It will be removed when the
	// container is removed.
	// Usually used to store container log files, files that will be bind
	// mounted into the container (e.g. the resolv.conf we made for the
	// container), and other per-container content.
	StaticDir string `json:"staticDir"`
	// Mounts contains all additional mounts into the container rootfs.
	// It is presently only used for the container's SHM directory.
	// These must be unmounted before the container's rootfs is unmounted.
	Mounts []string `json:"mounts,omitempty"`
	// NamedVolumes lists the Libpod named volumes to mount into the
	// container. Each named volume is guaranteed to exist so long as this
	// container exists.
	NamedVolumes []*ContainerNamedVolume `json:"namedVolumes,omitempty"`
	// OverlayVolumes lists the overlay volumes to mount into the container.
	OverlayVolumes []*ContainerOverlayVolume `json:"overlayVolumes,omitempty"`
	// ImageVolumes lists the image volumes to mount into the container.
	// Please note that this is named ctrImageVolumes in JSON to
	// distinguish between these and the old `imageVolumes` field in Podman
	// pre-1.8, which was used in very old Podman versions to determine how
	// image volumes were handled in Libpod (support for these eventually
	// moved out of Libpod into pkg/specgen).
	// Please DO NOT re-use the `imageVolumes` name in container JSON again.
	ImageVolumes []*ContainerImageVolume `json:"ctrImageVolumes,omitempty"`
	// CreateWorkingDir indicates that Libpod should create the container's
	// working directory if it does not exist. Some OCI runtimes do this by
	// default, but others do not.
	CreateWorkingDir bool `json:"createWorkingDir,omitempty"`
	// Secrets lists secrets to mount into the container
	Secrets []*ContainerSecret `json:"secrets,omitempty"`
	// SecretPath is the secrets location in storage
	SecretsPath string `json:"secretsPath"`
	// Volatile specifies whether the container storage can be optimized
	// at the cost of not syncing all the dirty files in memory.
	Volatile bool `json:"volatile,omitempty"`
}

ContainerRootFSConfig is an embedded sub-config providing config info about the container's root fs.

type ContainerSecret added in v3.2.3

type ContainerSecret struct {
	// Secret is the secret
	*secrets.Secret
	// UID is the UID of the secret file
	UID uint32
	// GID is the GID of the secret file
	GID uint32
	// Mode is the mode of the secret file
	Mode uint32
	// Secret target inside container
	Target string
}

ContainerSecret is a secret that is mounted in a container

type ContainerSecurityConfig

type ContainerSecurityConfig struct {
	// Privileged is whether the container is privileged. Privileged
	// containers have lessened security and increased access to the system.
	// Note that this does NOT directly correspond to Podman's --privileged
	// flag - most of the work of that flag is done in creating the OCI spec
	// given to Libpod. This only enables a small subset of the overall
	// operation, mostly around mounting the container image with reduced
	// security.
	Privileged bool `json:"privileged"`
	// ProcessLabel is the SELinux process label for the container.
	ProcessLabel string `json:"ProcessLabel,omitempty"`
	// MountLabel is the SELinux mount label for the container's root
	// filesystem. Only used if the container was created from an image.
	// If not explicitly set, an unused random MLS label will be assigned by
	// containers/storage (but only if SELinux is enabled).
	MountLabel string `json:"MountLabel,omitempty"`
	// LabelOpts are options passed in by the user to setup SELinux labels.
	// These are used by the containers/storage library.
	LabelOpts []string `json:"labelopts,omitempty"`
	// User and group to use in the container. Can be specified as only user
	// (in which case we will attempt to look up the user in the container
	// to determine the appropriate group) or user and group separated by a
	// colon.
	// Can be specified by name or UID/GID.
	// If unset, this will default to UID and GID 0 (root).
	User string `json:"user,omitempty"`
	// Groups are additional groups to add the container's user to. These
	// are resolved within the container using the container's /etc/passwd.
	Groups []string `json:"groups,omitempty"`
	// AddCurrentUserPasswdEntry indicates that Libpod should ensure that
	// the container's /etc/passwd contains an entry for the user running
	// Libpod - mostly used in rootless containers where the user running
	// Libpod wants to retain their UID inside the container.
	AddCurrentUserPasswdEntry bool `json:"addCurrentUserPasswdEntry,omitempty"`
}

ContainerSecurityConfig is an embedded sub-config providing security configuration to the container.

type ContainerState

type ContainerState struct {
	// The current state of the running container
	State define.ContainerStatus `json:"state"`
	// The path to the JSON OCI runtime spec for this container
	ConfigPath string `json:"configPath,omitempty"`
	// RunDir is a per-boot directory for container content
	RunDir string `json:"runDir,omitempty"`
	// Mounted indicates whether the container's storage has been mounted
	// for use
	Mounted bool `json:"mounted,omitempty"`
	// Mountpoint contains the path to the container's mounted storage as given
	// by containers/storage.
	Mountpoint string `json:"mountPoint,omitempty"`
	// StartedTime is the time the container was started
	StartedTime time.Time `json:"startedTime,omitempty"`
	// FinishedTime is the time the container finished executing
	FinishedTime time.Time `json:"finishedTime,omitempty"`
	// ExitCode is the exit code returned when the container stopped
	ExitCode int32 `json:"exitCode,omitempty"`
	// Exited is whether the container has exited
	Exited bool `json:"exited,omitempty"`
	// OOMKilled indicates that the container was killed as it ran out of
	// memory
	OOMKilled bool `json:"oomKilled,omitempty"`
	// Checkpointed indicates that the container was stopped by a checkpoint
	// operation.
	Checkpointed bool `json:"checkpointed,omitempty"`
	// PID is the PID of a running container
	PID int `json:"pid,omitempty"`
	// ConmonPID is the PID of the container's conmon
	ConmonPID int `json:"conmonPid,omitempty"`
	// ExecSessions contains all exec sessions that are associated with this
	// container.
	ExecSessions map[string]*ExecSession `json:"newExecSessions,omitempty"`
	// LegacyExecSessions are legacy exec sessions from older versions of
	// Podman.
	// These are DEPRECATED and will be removed in a future release.
	LegacyExecSessions map[string]*legacyExecSession `json:"execSessions,omitempty"`
	// NetworkStatus contains the configuration results for all networks
	// the pod is attached to. Only populated if we created a network
	// namespace for the container, and the network namespace is currently
	// active
	NetworkStatus []*cnitypes.Result `json:"networkResults,omitempty"`
	// BindMounts contains files that will be bind-mounted into the
	// container when it is mounted.
	// These include /etc/hosts and /etc/resolv.conf
	// This maps the path the file will be mounted to in the container to
	// the path of the file on disk outside the container
	BindMounts map[string]string `json:"bindMounts,omitempty"`
	// StoppedByUser indicates whether the container was stopped by an
	// explicit call to the Stop() API.
	StoppedByUser bool `json:"stoppedByUser,omitempty"`
	// RestartPolicyMatch indicates whether the conditions for restart
	// policy have been met.
	RestartPolicyMatch bool `json:"restartPolicyMatch,omitempty"`
	// RestartCount is how many times the container was restarted by its
	// restart policy. This is NOT incremented by normal container restarts
	// (only by restart policy).
	RestartCount uint `json:"restartCount,omitempty"`

	// ExtensionStageHooks holds hooks which will be executed by libpod
	// and not delegated to the OCI runtime.
	ExtensionStageHooks map[string][]spec.Hook `json:"extensionStageHooks,omitempty"`

	// NetInterfaceDescriptions describe the relationship between a CNI
	// network and an interface names
	NetInterfaceDescriptions ContainerNetworkDescriptions `json:"networkDescriptions,omitempty"`
	// contains filtered or unexported fields
}

ContainerState contains the current state of the container It is stored on disk in a tmpfs and recreated on reboot

type CtrCreateOption

type CtrCreateOption func(*Container) error

A CtrCreateOption is a functional option which alters the Container created by NewContainer

func WithAddCurrentUserPasswdEntry

func WithAddCurrentUserPasswdEntry() CtrCreateOption

WithAddCurrentUserPasswdEntry indicates that container should add current user entry to /etc/passwd, since the UID will be mapped into the container, via user namespace

func WithCDI added in v3.2.0

func WithCDI(devices []string) CtrCreateOption

WithCDI sets the devices to check for for CDI configuration.

func WithCgroupNSFrom

func WithCgroupNSFrom(nsCtr *Container) CtrCreateOption

WithCgroupNSFrom indicates the the container should join the CGroup namespace of the given container. If the container has joined a pod, it can only join the namespaces of containers in the same pod.

func WithCgroupParent

func WithCgroupParent(parent string) CtrCreateOption

WithCgroupParent sets the Cgroup Parent of the new container.

func WithCgroupsMode

func WithCgroupsMode(mode string) CtrCreateOption

WithCgroupsMode disables the creation of CGroups for the conmon process.

func WithCommand

func WithCommand(command []string) CtrCreateOption

WithCommand sets the command of the container. This is not used to change the container's spec, but will instead be used during commit to populate the command of the new image. If not explicitly set it will default to the image's command. A nil command is allowed, and will clear command on the created image.

func WithConmonPidFile

func WithConmonPidFile(path string) CtrCreateOption

WithConmonPidFile specifies the path to the file that receives the pid of conmon.

func WithCreateCommand

func WithCreateCommand(cmd []string) CtrCreateOption

WithCreateCommand adds the full command plus arguments of the current process to the container config.

func WithCreateWorkingDir

func WithCreateWorkingDir() CtrCreateOption

WithCreateWorkingDir tells Podman to create the container's working directory if it does not exist.

func WithCtrNamespace

func WithCtrNamespace(ns string) CtrCreateOption

WithCtrNamespace sets the namespace the container will be created in. Namespaces are used to create separate views of Podman's state - runtimes can join a specific namespace and see only containers and pods in that namespace. Empty string namespaces are allowed, and correspond to a lack of namespace.

func WithDNS

func WithDNS(dnsServers []string) CtrCreateOption

WithDNS sets additional name servers for the container.

func WithDNSOption

func WithDNSOption(dnsOptions []string) CtrCreateOption

WithDNSOption sets addition dns options for the container.

func WithDNSSearch

func WithDNSSearch(searchDomains []string) CtrCreateOption

WithDNSSearch sets the additional search domains of a container.

func WithDependencyCtrs

func WithDependencyCtrs(ctrs []*Container) CtrCreateOption

WithDependencyCtrs sets dependency containers of the given container. Dependency containers must be running before this container is started.

func WithEntrypoint

func WithEntrypoint(entrypoint []string) CtrCreateOption

WithEntrypoint sets the entrypoint of the container. This is not used to change the container's spec, but will instead be used during commit to populate the entrypoint of the new image. If not explicitly set it will default to the image's entrypoint. A nil entrypoint is allowed, and will clear entrypoint on the created image.

func WithEnvSecrets added in v3.2.0

func WithEnvSecrets(envSecrets map[string]string) CtrCreateOption

WithSecrets adds environment variable secrets to the container

func WithExitCommand

func WithExitCommand(exitCommand []string) CtrCreateOption

WithExitCommand sets the ExitCommand for the container, appending on the ctr.ID() to the end

func WithGroups

func WithGroups(groups []string) CtrCreateOption

WithGroups sets additional groups for the container, which are defined by the user.

func WithHealthCheck

func WithHealthCheck(healthCheck *manifest.Schema2HealthConfig) CtrCreateOption

WithHealthCheck adds the healthcheck to the container config

func WithHosts

func WithHosts(hosts []string) CtrCreateOption

WithHosts sets additional host:IP for the hosts file.

func WithIDMappings

func WithIDMappings(idmappings storage.IDMappingOptions) CtrCreateOption

WithIDMappings sets the idmappings for the container

func WithIPCNSFrom

func WithIPCNSFrom(nsCtr *Container) CtrCreateOption

WithIPCNSFrom indicates the the container should join the IPC namespace of the given container. If the container has joined a pod, it can only join the namespaces of containers in the same pod.

func WithImageVolumes

func WithImageVolumes(volumes []*ContainerImageVolume) CtrCreateOption

WithImageVolumes adds the given image volumes to the container.

func WithInitCtrType added in v3.4.0

func WithInitCtrType(containerType string) CtrCreateOption

WithInitCtrType indicates the container is a initcontainer

func WithLabels

func WithLabels(labels map[string]string) CtrCreateOption

WithLabels adds labels to the container.

func WithLogDriver

func WithLogDriver(driver string) CtrCreateOption

WithLogDriver sets the log driver for the container

func WithLogPath

func WithLogPath(path string) CtrCreateOption

WithLogPath sets the path to the log file.

func WithLogTag

func WithLogTag(tag string) CtrCreateOption

WithLogTag sets the tag to the log file.

func WithMaxLogSize

func WithMaxLogSize(limit int64) CtrCreateOption

WithMaxLogSize sets the maximum size of container logs. Positive sizes are limits in bytes, -1 is unlimited.

func WithMountNSFrom

func WithMountNSFrom(nsCtr *Container) CtrCreateOption

WithMountNSFrom indicates the the container should join the mount namespace of the given container. If the container has joined a pod, it can only join the namespaces of containers in the same pod.

func WithName

func WithName(name string) CtrCreateOption

WithName sets the container's name.

func WithNamedVolumes

func WithNamedVolumes(volumes []*ContainerNamedVolume) CtrCreateOption

WithNamedVolumes adds the given named volumes to the container.

func WithNetNS

func WithNetNS(portMappings []ocicni.PortMapping, exposedPorts map[uint16][]string, postConfigureNetNS bool, netmode string, networks []string) CtrCreateOption

WithNetNS indicates that the container should be given a new network namespace with a minimal configuration. An optional array of port mappings can be provided. Conflicts with WithNetNSFrom().

func WithNetNSFrom

func WithNetNSFrom(nsCtr *Container) CtrCreateOption

WithNetNSFrom indicates the the container should join the network namespace of the given container. If the container has joined a pod, it can only join the namespaces of containers in the same pod.

func WithNetworkAliases

func WithNetworkAliases(aliases map[string][]string) CtrCreateOption

WithNetworkAliases sets network aliases for the container. Accepts a map of network name to aliases.

func WithNetworkOptions

func WithNetworkOptions(options map[string][]string) CtrCreateOption

WithNetworkOptions sets additional options for the networks.

func WithOverlayVolumes

func WithOverlayVolumes(volumes []*ContainerOverlayVolume) CtrCreateOption

WithOverlayVolumes adds the given overlay volumes to the container.

func WithPIDNSFrom

func WithPIDNSFrom(nsCtr *Container) CtrCreateOption

WithPIDNSFrom indicates the the container should join the PID namespace of the given container. If the container has joined a pod, it can only join the namespaces of containers in the same pod.

func WithPidFile added in v3.2.0

func WithPidFile(pidFile string) CtrCreateOption

WithPidFile adds pidFile to the container

func WithPreserveFDs

func WithPreserveFDs(fd uint) CtrCreateOption

WithPreserveFDs forwards from the process running Libpod into the container the given number of extra FDs (starting after the standard streams) to the created container

func WithPrivileged

func WithPrivileged(privileged bool) CtrCreateOption

WithPrivileged sets the privileged flag in the container runtime.

func WithRestartPolicy

func WithRestartPolicy(policy string) CtrCreateOption

WithRestartPolicy sets the container's restart policy. Valid values are "no", "on-failure", and "always". The empty string is allowed, and will be equivalent to "no".

func WithRestartRetries

func WithRestartRetries(tries uint) CtrCreateOption

WithRestartRetries sets the number of retries to use when restarting a container with the "on-failure" restart policy. 0 is an allowed value, and indicates infinite retries.

func WithRootFS

func WithRootFS(rootfs string) CtrCreateOption

WithRootFS sets the rootfs for the container. This creates a container from a directory on disk and not an image.

func WithRootFSFromImage

func WithRootFSFromImage(imageID, imageName, rawImageName string) CtrCreateOption

WithRootFSFromImage sets up a fresh root filesystem using the given image. If useImageConfig is specified, image volumes, environment variables, and other configuration from the image will be added to the config. TODO: Replace image name and ID with a libpod.Image struct when that is finished.

func WithSdNotifyMode

func WithSdNotifyMode(mode string) CtrCreateOption

WithSdNotifyMode sets the sd-notify method

func WithSecLabels

func WithSecLabels(labelOpts []string) CtrCreateOption

WithSecLabels sets the labels for SELinux.

func WithSecrets added in v3.1.0

func WithSecrets(containerSecrets []*ContainerSecret) CtrCreateOption

WithSecrets adds secrets to the container

func WithShmDir

func WithShmDir(dir string) CtrCreateOption

WithShmDir sets the directory that should be mounted on /dev/shm.

func WithShmSize

func WithShmSize(size int64) CtrCreateOption

WithShmSize sets the size of /dev/shm tmpfs mount.

func WithStaticIP

func WithStaticIP(ip net.IP) CtrCreateOption

WithStaticIP indicates that the container should request a static IP from the CNI plugins. It cannot be set unless WithNetNS has already been passed. Further, it cannot be set if additional CNI networks to join have been specified.

func WithStaticMAC

func WithStaticMAC(mac net.HardwareAddr) CtrCreateOption

WithStaticMAC indicates that the container should request a static MAC from the CNI plugins. It cannot be set unless WithNetNS has already been passed. Further, it cannot be set if additional CNI networks to join have been specified.

func WithStdin

func WithStdin() CtrCreateOption

WithStdin keeps stdin on the container open to allow interaction.

func WithStopSignal

func WithStopSignal(signal syscall.Signal) CtrCreateOption

WithStopSignal sets the signal that will be sent to stop the container.

func WithStopTimeout

func WithStopTimeout(timeout uint) CtrCreateOption

WithStopTimeout sets the time to after initial stop signal is sent to the container, before sending the kill signal.

func WithSystemd

func WithSystemd() CtrCreateOption

WithSystemd turns on systemd mode in the container

func WithTimeout added in v3.2.0

func WithTimeout(timeout uint) CtrCreateOption

WithTimeout sets the maximum time a container is allowed to run"

func WithTimezone

func WithTimezone(path string) CtrCreateOption

WithTimezone sets the timezone in the container

func WithUTSNSFrom

func WithUTSNSFrom(nsCtr *Container) CtrCreateOption

WithUTSNSFrom indicates the the container should join the UTS namespace of the given container. If the container has joined a pod, it can only join the namespaces of containers in the same pod.

func WithUTSNSFromPod

func WithUTSNSFromPod(p *Pod) CtrCreateOption

WithUTSNSFromPod indicates the the container should join the UTS namespace of its pod

func WithUmask

func WithUmask(umask string) CtrCreateOption

WithUmask sets the umask in the container

func WithUseImageHosts

func WithUseImageHosts() CtrCreateOption

WithUseImageHosts tells the container not to bind-mount /etc/hosts in. This conflicts with WithHosts().

func WithUseImageResolvConf

func WithUseImageResolvConf() CtrCreateOption

WithUseImageResolvConf tells the container not to bind-mount resolv.conf in. This conflicts with other DNS-related options.

func WithUser

func WithUser(user string) CtrCreateOption

WithUser sets the user identity field in configuration. Valid uses [user | user:group | uid | uid:gid | user:gid | uid:group ].

func WithUserNSFrom

func WithUserNSFrom(nsCtr *Container) CtrCreateOption

WithUserNSFrom indicates the the container should join the user namespace of the given container. If the container has joined a pod, it can only join the namespaces of containers in the same pod.

func WithUserVolumes

func WithUserVolumes(volumes []string) CtrCreateOption

WithUserVolumes sets the user-added volumes of the container. These are not added to the container's spec, but will instead be used during commit to populate the volumes of the new image, and to trigger some OCI hooks that are only added if volume mounts are present. Furthermore, they are used in the output of inspect, to filter volumes - only volumes included in this list will be included in the output. Unless explicitly set, committed images will have no volumes. The given volumes slice must not be nil.

func WithVolatile added in v3.2.0

func WithVolatile() CtrCreateOption

WithVolatile sets the volatile flag for the container storage. The option can potentially cause data loss when used on a container that must survive a machine reboot.

type DBConfig

type DBConfig struct {
	LibpodRoot  string
	LibpodTmp   string
	StorageRoot string
	StorageTmp  string
	GraphDriver string
	VolumePath  string
}

DBConfig is a set of Libpod runtime configuration settings that are saved in a State when it is first created, and can subsequently be retrieved.

type ExecConfig

type ExecConfig struct {
	// Command the the command that will be invoked in the exec session.
	// Must not be empty.
	Command []string `json:"command"`
	// Terminal is whether the exec session will allocate a pseudoterminal.
	Terminal bool `json:"terminal,omitempty"`
	// AttachStdin is whether the STDIN stream will be forwarded to the exec
	// session's first process when attaching. Only available if Terminal is
	// false.
	AttachStdin bool `json:"attachStdin,omitempty"`
	// AttachStdout is whether the STDOUT stream will be forwarded to the
	// exec session's first process when attaching. Only available if
	// Terminal is false.
	AttachStdout bool `json:"attachStdout,omitempty"`
	// AttachStderr is whether the STDERR stream will be forwarded to the
	// exec session's first process when attaching. Only available if
	// Terminal is false.
	AttachStderr bool `json:"attachStderr,omitempty"`
	// DetachKeys are keys that will be used to detach from the exec
	// session. Here, nil will use the default detach keys, where a pointer
	// to the empty string ("") will disable detaching via detach keys.
	DetachKeys *string `json:"detachKeys,omitempty"`
	// Environment is a set of environment variables that will be set for
	// the first process started by the exec session.
	Environment map[string]string `json:"environment,omitempty"`
	// Privileged is whether the exec session will be privileged - that is,
	// will be granted additional capabilities.
	Privileged bool `json:"privileged,omitempty"`
	// User is the user the exec session will be run as.
	// If set to "" the exec session will be started as the same user the
	// container was started as.
	User string `json:"user,omitempty"`
	// WorkDir is the working directory for the first process that will be
	// launched by the exec session.
	// If set to "" the exec session will be started in / within the
	// container.
	WorkDir string `json:"workDir,omitempty"`
	// PreserveFDs indicates that a number of extra FDs from the process
	// running libpod will be passed into the container. These are assumed
	// to begin at 3 (immediately after the standard streams). The number
	// given is the number that will be passed into the exec session,
	// starting at 3.
	PreserveFDs uint `json:"preserveFds,omitempty"`
	// ExitCommand is the exec session's exit command.
	// This command will be executed when the exec session exits.
	// If unset, no command will be executed.
	// Two arguments will be appended to the exit command by Libpod:
	// The ID of the exec session, and the ID of the container the exec
	// session is a part of (in that order).
	ExitCommand []string `json:"exitCommand,omitempty"`
	// ExitCommandDelay is a delay (in seconds) between the container
	// exiting, and the exit command being executed. If set to 0, there is
	// no delay. If set, ExitCommand must also be set.
	ExitCommandDelay uint `json:"exitCommandDelay,omitempty"`
}

ExecConfig contains the configuration of an exec session

type ExecOptions

type ExecOptions struct {
	// Cmd is the command to execute.
	Cmd []string
	// Env is a set of environment variables to add to the container.
	Env map[string]string
	// Terminal is whether to create a new TTY for the exec session.
	Terminal bool
	// Cwd is the working directory for the executed command. If unset, the
	// working directory of the container will be used.
	Cwd string
	// User is the user the command will be executed as. If unset, the user
	// the container was run as will be used.
	User string
	// Streams are the streams that will be attached to the container.
	Streams *define.AttachStreams
	// PreserveFDs is a number of additional file descriptors (in addition
	// to 0, 1, 2) that will be passed to the executed process. The total FDs
	// passed will be 3 + PreserveFDs.
	PreserveFDs uint
	// DetachKeys is a set of keys that, when pressed in sequence, will
	// detach from the container.
	// If not provided, the default keys will be used.
	// If provided but set to "", detaching from the container will be
	// disabled.
	DetachKeys *string
	// ExitCommand is a command that will be run after the exec session
	// exits.
	ExitCommand []string
	// ExitCommandDelay is a delay (in seconds) between the exec session
	// exiting, and the exit command being invoked.
	ExitCommandDelay uint
	// Privileged indicates the execed process will be launched in Privileged mode
	Privileged bool
}

ExecOptions are options passed into ExecContainer. They control the command that will be executed and how the exec will proceed.

type ExecSession

type ExecSession struct {
	// Id is the ID of the exec session.
	// Named somewhat strangely to not conflict with ID().
	// nolint:stylecheck,golint
	Id string `json:"id"`
	// ContainerId is the ID of the container this exec session belongs to.
	// Named somewhat strangely to not conflict with ContainerID().
	// nolint:stylecheck,golint
	ContainerId string `json:"containerId"`

	// State is the state of the exec session.
	State define.ContainerExecStatus `json:"state"`
	// PID is the PID of the process created by the exec session.
	PID int `json:"pid,omitempty"`
	// ExitCode is the exit code of the exec session, if it has exited.
	ExitCode int `json:"exitCode,omitempty"`

	// Config is the configuration of this exec session.
	// Cannot be empty.
	Config *ExecConfig `json:"config"`
}

ExecSession contains information on a single exec session attached to a given container.

func (*ExecSession) ContainerID

func (e *ExecSession) ContainerID() string

ContainerID returns the ID of the container this exec session was started in.

func (*ExecSession) ID

func (e *ExecSession) ID() string

ID returns the ID of an exec session.

func (*ExecSession) Inspect

func (e *ExecSession) Inspect() (*define.InspectExecSession, error)

Inspect inspects the given exec session and produces detailed output on its configuration and current state.

type HTTPAttachStreams

type HTTPAttachStreams struct {
	Stdin  bool
	Stdout bool
	Stderr bool
}

HTTPAttachStreams informs the HTTPAttach endpoint which of the container's standard streams should be streamed to the client. If this is passed, at least one of the streams must be set to true.

type LinuxNS

type LinuxNS int

LinuxNS represents a Linux namespace

const (
	// InvalidNS is an invalid namespace
	InvalidNS LinuxNS = iota
	// IPCNS is the IPC namespace
	IPCNS LinuxNS = iota
	// MountNS is the mount namespace
	MountNS LinuxNS = iota
	// NetNS is the network namespace
	NetNS LinuxNS = iota
	// PIDNS is the PID namespace
	PIDNS LinuxNS = iota
	// UserNS is the user namespace
	UserNS LinuxNS = iota
	// UTSNS is the UTS namespace
	UTSNS LinuxNS = iota
	// CgroupNS is the CGroup namespace
	CgroupNS LinuxNS = iota
)

func (LinuxNS) String

func (ns LinuxNS) String() string

String returns a string representation of a Linux namespace It is guaranteed to be the name of the namespace in /proc for valid ns types

type MissingRuntime

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

MissingRuntime is used when the OCI runtime requested by the container is missing (not installed or not in the configuration file).

func (*MissingRuntime) AttachResize

func (r *MissingRuntime) AttachResize(ctr *Container, newSize define.TerminalSize) error

AttachResize is not available as the runtime is missing

func (*MissingRuntime) AttachSocketPath

func (r *MissingRuntime) AttachSocketPath(ctr *Container) (string, error)

AttachSocketPath does not work as there is no runtime to attach to. (Theoretically we could follow ExitFilePath but there is no guarantee the container is running and thus has an attach socket...)

func (*MissingRuntime) CheckConmonRunning

func (r *MissingRuntime) CheckConmonRunning(ctr *Container) (bool, error)

CheckConmonRunning is not available as the runtime is missing

func (*MissingRuntime) CheckpointContainer

func (r *MissingRuntime) CheckpointContainer(ctr *Container, options ContainerCheckpointOptions) error

CheckpointContainer is not available as the runtime is missing

func (*MissingRuntime) CreateContainer

func (r *MissingRuntime) CreateContainer(ctr *Container, restoreOptions *ContainerCheckpointOptions) error

CreateContainer is not available as the runtime is missing

func (*MissingRuntime) DeleteContainer

func (r *MissingRuntime) DeleteContainer(ctr *Container) error

DeleteContainer is not available as the runtime is missing

func (*MissingRuntime) ExecAttachResize

func (r *MissingRuntime) ExecAttachResize(ctr *Container, sessionID string, newSize define.TerminalSize) error

ExecAttachResize is not available as the runtime is missing.

func (*MissingRuntime) ExecAttachSocketPath

func (r *MissingRuntime) ExecAttachSocketPath(ctr *Container, sessionID string) (string, error)

ExecAttachSocketPath does not work as there is no runtime to attach to. (Again, we could follow ExitFilePath, but no guarantee there is an existing and running exec session)

func (*MissingRuntime) ExecContainer

func (r *MissingRuntime) ExecContainer(ctr *Container, sessionID string, options *ExecOptions, streams *define.AttachStreams, newSize *define.TerminalSize) (int, chan error, error)

ExecContainer is not available as the runtime is missing

func (*MissingRuntime) ExecContainerDetached

func (r *MissingRuntime) ExecContainerDetached(ctr *Container, sessionID string, options *ExecOptions, stdin bool) (int, error)

ExecContainerDetached is not available as the runtime is missing

func (*MissingRuntime) ExecContainerHTTP

func (r *MissingRuntime) ExecContainerHTTP(ctr *Container, sessionID string, options *ExecOptions, req *http.Request, w http.ResponseWriter,
	streams *HTTPAttachStreams, cancel <-chan bool, hijackDone chan<- bool, holdConnOpen <-chan bool, newSize *define.TerminalSize) (int, chan error, error)

ExecContainerHTTP is not available as the runtime is missing

func (*MissingRuntime) ExecStopContainer

func (r *MissingRuntime) ExecStopContainer(ctr *Container, sessionID string, timeout uint) error

ExecStopContainer is not available as the runtime is missing. TODO: We can also investigate using unix.Kill() on the PID of the exec session here if we want to make stopping containers possible. Won't be perfect, though.

func (*MissingRuntime) ExecUpdateStatus

func (r *MissingRuntime) ExecUpdateStatus(ctr *Container, sessionID string) (bool, error)

ExecUpdateStatus is not available as the runtime is missing.

func (*MissingRuntime) ExitFilePath

func (r *MissingRuntime) ExitFilePath(ctr *Container) (string, error)

ExitFilePath returns the exit file path for containers. Here, we mimic what ConmonOCIRuntime does, because there is a chance that the container in question is still running happily (config file modified to remove a runtime, for example). We can't find the runtime to do anything to the container, but Conmon should still place an exit file for it.

func (*MissingRuntime) HTTPAttach

func (r *MissingRuntime) HTTPAttach(ctr *Container, req *http.Request, w http.ResponseWriter, streams *HTTPAttachStreams, detachKeys *string, cancel <-chan bool, hijackDone chan<- bool, streamAttach, streamLogs bool) error

HTTPAttach is not available as the runtime is missing

func (*MissingRuntime) KillContainer

func (r *MissingRuntime) KillContainer(ctr *Container, signal uint, all bool) error

KillContainer is not available as the runtime is missing TODO: We could attempt to unix.Kill() the PID as recorded in the state if we really want to smooth things out? Won't be perfect, but if the container has a PID namespace it could be enough?

func (*MissingRuntime) Name

func (r *MissingRuntime) Name() string

Name is the name of the missing runtime

func (*MissingRuntime) Path

func (r *MissingRuntime) Path() string

Path is not available as the runtime is missing

func (*MissingRuntime) PauseContainer

func (r *MissingRuntime) PauseContainer(ctr *Container) error

PauseContainer is not available as the runtime is missing

func (*MissingRuntime) RuntimeInfo

func (r *MissingRuntime) RuntimeInfo() (*define.ConmonInfo, *define.OCIRuntimeInfo, error)

RuntimeInfo returns information on the missing runtime

func (*MissingRuntime) StartContainer

func (r *MissingRuntime) StartContainer(ctr *Container) error

StartContainer is not available as the runtime is missing

func (*MissingRuntime) StopContainer

func (r *MissingRuntime) StopContainer(ctr *Container, timeout uint, all bool) error

StopContainer is not available as the runtime is missing

func (*MissingRuntime) SupportsCheckpoint

func (r *MissingRuntime) SupportsCheckpoint() bool

SupportsCheckpoint returns false as checkpointing requires a working runtime

func (*MissingRuntime) SupportsJSONErrors

func (r *MissingRuntime) SupportsJSONErrors() bool

SupportsJSONErrors returns false as there is no runtime to give errors

func (*MissingRuntime) SupportsKVM

func (r *MissingRuntime) SupportsKVM() bool

SupportsKVM checks if the OCI runtime supports running containers without KVM separation

func (*MissingRuntime) SupportsNoCgroups

func (r *MissingRuntime) SupportsNoCgroups() bool

SupportsNoCgroups returns false as there is no runtime to create containers

func (*MissingRuntime) UnpauseContainer

func (r *MissingRuntime) UnpauseContainer(ctr *Container) error

UnpauseContainer is not available as the runtime is missing

func (*MissingRuntime) UpdateContainerStatus

func (r *MissingRuntime) UpdateContainerStatus(ctr *Container) error

UpdateContainerStatus is not available as the runtime is missing

type OCIRuntime

type OCIRuntime interface {
	// Name returns the name of the runtime.
	Name() string
	// Path returns the path to the runtime executable.
	Path() string

	// CreateContainer creates the container in the OCI runtime.
	CreateContainer(ctr *Container, restoreOptions *ContainerCheckpointOptions) error
	// UpdateContainerStatus updates the status of the given container.
	UpdateContainerStatus(ctr *Container) error
	// StartContainer starts the given container.
	StartContainer(ctr *Container) error
	// KillContainer sends the given signal to the given container.
	// If all is set, all processes in the container will be signalled;
	// otherwise, only init will be signalled.
	KillContainer(ctr *Container, signal uint, all bool) error
	// StopContainer stops the given container.
	// The container's stop signal (or SIGTERM if unspecified) will be sent
	// first.
	// After the given timeout, SIGKILL will be sent.
	// If the given timeout is 0, SIGKILL will be sent immediately, and the
	// stop signal will be omitted.
	// If all is set, we will attempt to use the --all flag will `kill` in
	// the OCI runtime to kill all processes in the container, including
	// exec sessions. This is only supported if the container has cgroups.
	StopContainer(ctr *Container, timeout uint, all bool) error
	// DeleteContainer deletes the given container from the OCI runtime.
	DeleteContainer(ctr *Container) error
	// PauseContainer pauses the given container.
	PauseContainer(ctr *Container) error
	// UnpauseContainer unpauses the given container.
	UnpauseContainer(ctr *Container) error

	// HTTPAttach performs an attach intended to be transported over HTTP.
	// For terminal attach, the container's output will be directly streamed
	// to output; otherwise, STDOUT and STDERR will be multiplexed, with
	// a header prepended as follows: 1-byte STREAM (0, 1, 2 for STDIN,
	// STDOUT, STDERR), 3 null (0x00) bytes, 4-byte big endian length.
	// If a cancel channel is provided, it can be used to asynchronously
	// terminate the attach session. Detach keys, if given, will also cause
	// the attach session to be terminated if provided via the STDIN
	// channel. If they are not provided, the default detach keys will be
	// used instead. Detach keys of "" will disable detaching via keyboard.
	// The streams parameter will determine which streams to forward to the
	// client.
	HTTPAttach(ctr *Container, r *http.Request, w http.ResponseWriter, streams *HTTPAttachStreams, detachKeys *string, cancel <-chan bool, hijackDone chan<- bool, streamAttach, streamLogs bool) error
	// AttachResize resizes the terminal in use by the given container.
	AttachResize(ctr *Container, newSize define.TerminalSize) error

	// ExecContainer executes a command in a running container.
	// Returns an int (PID of exec session), error channel (errors from
	// attach), and error (errors that occurred attempting to start the exec
	// session). This returns once the exec session is running - not once it
	// has completed, as one might expect. The attach session will remain
	// running, in a goroutine that will return via the chan error in the
	// return signature.
	// newSize resizes the tty to this size before the process is started, must be nil if the exec session has no tty
	ExecContainer(ctr *Container, sessionID string, options *ExecOptions, streams *define.AttachStreams, newSize *define.TerminalSize) (int, chan error, error)
	// ExecContainerHTTP executes a command in a running container and
	// attaches its standard streams to a provided hijacked HTTP session.
	// Maintains the same invariants as ExecContainer (returns on session
	// start, with a goroutine running in the background to handle attach).
	// The HTTP attach itself maintains the same invariants as HTTPAttach.
	// newSize resizes the tty to this size before the process is started, must be nil if the exec session has no tty
	ExecContainerHTTP(ctr *Container, sessionID string, options *ExecOptions, r *http.Request, w http.ResponseWriter,
		streams *HTTPAttachStreams, cancel <-chan bool, hijackDone chan<- bool, holdConnOpen <-chan bool, newSize *define.TerminalSize) (int, chan error, error)
	// ExecContainerDetached executes a command in a running container, but
	// does not attach to it. Returns the PID of the exec session and an
	// error (if starting the exec session failed)
	ExecContainerDetached(ctr *Container, sessionID string, options *ExecOptions, stdin bool) (int, error)
	// ExecAttachResize resizes the terminal of a running exec session. Only
	// allowed with sessions that were created with a TTY.
	ExecAttachResize(ctr *Container, sessionID string, newSize define.TerminalSize) error
	// ExecStopContainer stops a given exec session in a running container.
	// SIGTERM with be sent initially, then SIGKILL after the given timeout.
	// If timeout is 0, SIGKILL will be sent immediately, and SIGTERM will
	// be omitted.
	ExecStopContainer(ctr *Container, sessionID string, timeout uint) error
	// ExecUpdateStatus checks the status of a given exec session.
	// Returns true if the session is still running, or false if it exited.
	ExecUpdateStatus(ctr *Container, sessionID string) (bool, error)

	// CheckpointContainer checkpoints the given container.
	// Some OCI runtimes may not support this - if SupportsCheckpoint()
	// returns false, this is not implemented, and will always return an
	// error.
	CheckpointContainer(ctr *Container, options ContainerCheckpointOptions) error

	// CheckConmonRunning verifies that the given container's Conmon
	// instance is still running. Runtimes without Conmon, or systems where
	// the PID of conmon is not available, should mock this as True.
	// True indicates that Conmon for the instance is running, False
	// indicates it is not.
	CheckConmonRunning(ctr *Container) (bool, error)

	// SupportsCheckpoint returns whether this OCI runtime
	// implementation supports the CheckpointContainer() operation.
	SupportsCheckpoint() bool
	// SupportsJSONErrors is whether the runtime can return JSON-formatted
	// error messages.
	SupportsJSONErrors() bool
	// SupportsNoCgroups is whether the runtime supports running containers
	// without cgroups.
	SupportsNoCgroups() bool
	// SupportsKVM os whether the OCI runtime supports running containers
	// without KVM separation
	SupportsKVM() bool

	// AttachSocketPath is the path to the socket to attach to a given
	// container.
	// TODO: If we move Attach code in here, this should be made internal.
	// We don't want to force all runtimes to share the same attach
	// implementation.
	AttachSocketPath(ctr *Container) (string, error)
	// ExecAttachSocketPath is the path to the socket to attach to a given
	// exec session in the given container.
	// TODO: Probably should be made internal.
	ExecAttachSocketPath(ctr *Container, sessionID string) (string, error)
	// ExitFilePath is the path to a container's exit file.
	// All runtime implementations must create an exit file when containers
	// exit, containing the exit code of the container (as a string).
	// This is the path to that file for a given container.
	ExitFilePath(ctr *Container) (string, error)

	// RuntimeInfo returns verbose information about the runtime.
	RuntimeInfo() (*define.ConmonInfo, *define.OCIRuntimeInfo, error)
}

OCIRuntime is an implementation of an OCI runtime. The OCI runtime implementation is expected to be a fairly thin wrapper around the actual runtime, and is not expected to include things like state management logic - e.g., we do not expect it to determine on its own that calling 'UnpauseContainer()' on a container that is not paused is an error. The code calling the OCIRuntime will manage this. TODO: May want to move the Attach() code under this umbrella. It's highly OCI runtime dependent. TODO: May want to move the conmon cleanup code here too - it depends on Conmon being in use.

type Pod

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

Pod represents a group of containers that are managed together. Any operations on a Pod that access state must begin with a call to updatePod(). There is no guarantee that state exists in a readable state before this call, and even if it does its contents will be out of date and must be refreshed from the database. Generally, this requirement applies only to top-level functions; helpers can assume their callers handled this requirement. Generally speaking, if a function takes the pod lock and accesses any part of state, it should updatePod() immediately after locking. Pod represents a group of containers that may share namespaces

func (*Pod) AllContainers

func (p *Pod) AllContainers() ([]*Container, error)

AllContainers retrieves the containers in the pod

func (*Pod) AllContainersByID

func (p *Pod) AllContainersByID() ([]string, error)

AllContainersByID returns the container IDs of all the containers in the pod

func (*Pod) CgroupParent

func (p *Pod) CgroupParent() string

CgroupParent returns the pod's CGroup parent

func (*Pod) CgroupPath

func (p *Pod) CgroupPath() (string, error)

CgroupPath returns the path to the pod's CGroup

func (*Pod) Cleanup

func (p *Pod) Cleanup(ctx context.Context) (map[string]error, error)

Cleanup cleans up all containers within a pod that have stopped. All containers are cleaned up independently. An error with one container will not prevent other containers being cleaned up. An error and a map[string]error are returned. If the error is not nil and the map is nil, an error was encountered before any containers were cleaned up. If map is not nil, an error was encountered when working on one or more containers. The container ID is mapped to the error encountered. The error is set to ErrPodPartialFail. If both error and the map are nil, all containers were paused without error

func (*Pod) CreateCommand

func (p *Pod) CreateCommand() []string

CreateCommand returns the os.Args of the process with which the pod has been created.

func (*Pod) CreatedTime

func (p *Pod) CreatedTime() time.Time

CreatedTime gets the time when the pod was created

func (*Pod) GenerateForKube

func (p *Pod) GenerateForKube(ctx context.Context) (*v1.Pod, []v1.ServicePort, error)

GenerateForKube takes a slice of libpod containers and generates one v1.Pod description

func (*Pod) GetPodPidInformation

func (p *Pod) GetPodPidInformation(descriptors []string) ([]string, error)

GetPodPidInformation returns process-related data of all processes in the pod. The output data can be controlled via the `descriptors` argument which expects format descriptors and supports all AIXformat descriptors of ps (1) plus some additional ones to for instance inspect the set of effective capabilities. Each element in the returned string slice is a tab-separated string.

For more details, please refer to github.com/containers/psgo.

func (*Pod) GetPodStats

func (p *Pod) GetPodStats(previousContainerStats map[string]*define.ContainerStats) (map[string]*define.ContainerStats, error)

GetPodStats returns the stats for each of its containers

func (*Pod) GetPodStatus

func (p *Pod) GetPodStatus() (string, error)

GetPodStatus determines the status of the pod based on the statuses of the containers in the pod. Returns a string representation of the pod status

func (*Pod) HasContainer

func (p *Pod) HasContainer(id string) (bool, error)

HasContainer checks if a container is present in the pod

func (*Pod) HasInfraContainer

func (p *Pod) HasInfraContainer() bool

HasInfraContainer returns whether the pod will create an infra container

func (*Pod) Hostname

func (p *Pod) Hostname() string

Hostname returns the hostname of the pod.

func (*Pod) ID

func (p *Pod) ID() string

ID retrieves the pod's ID

func (*Pod) InfraContainer

func (p *Pod) InfraContainer() (*Container, error)

InfraContainer returns the infra container.

func (*Pod) InfraContainerID

func (p *Pod) InfraContainerID() (string, error)

InfraContainerID returns the infra container ID for a pod. If the container returned is "", the pod has no infra container.

func (*Pod) Inspect

func (p *Pod) Inspect() (*define.InspectPodData, error)

Inspect returns a PodInspect struct to describe the pod.

func (*Pod) Kill

func (p *Pod) Kill(ctx context.Context, signal uint) (map[string]error, error)

Kill sends a signal to all running containers within a pod. Signals will only be sent to running containers. Containers that are not running will be ignored. All signals are sent independently, and sending will continue even if some containers encounter errors. An error and a map[string]error are returned. If the error is not nil and the map is nil, an error was encountered before any containers were signalled. If map is not nil, an error was encountered when signalling one or more containers. The container ID is mapped to the error encountered. The error is set to ErrPodPartialFail. If both error and the map are nil, all containers were signalled successfully.

func (*Pod) Labels

func (p *Pod) Labels() map[string]string

Labels returns the pod's labels

func (*Pod) Name

func (p *Pod) Name() string

Name retrieves the pod's name

func (*Pod) Namespace

func (p *Pod) Namespace() string

Namespace returns the pod's libpod namespace. Namespaces are used to logically separate containers and pods in the state.

func (*Pod) Pause

func (p *Pod) Pause(ctx context.Context) (map[string]error, error)

Pause pauses all containers within a pod that are running. Only running containers will be paused. Paused, stopped, or created containers will be ignored. All containers are paused independently. An error pausing one container will not prevent other containers being paused. An error and a map[string]error are returned. If the error is not nil and the map is nil, an error was encountered before any containers were paused. If map is not nil, an error was encountered when pausing one or more containers. The container ID is mapped to the error encountered. The error is set to ErrPodPartialFail. If both error and the map are nil, all containers were paused without error

func (*Pod) PidMode added in v3.3.0

func (p *Pod) PidMode() string

PidMode returns the PID mode given by the user ex: pod, private...

func (*Pod) ProcessLabel

func (p *Pod) ProcessLabel() (string, error)

ProcessLabel returns the SELinux label associated with the pod

func (*Pod) Restart

func (p *Pod) Restart(ctx context.Context) (map[string]error, error)

Restart restarts all containers within a pod that are not paused or in an error state. It combines the effects of Stop() and Start() on a container Each container will use its own stop timeout. All containers are started independently, in order dictated by their dependencies. An error restarting one container will not prevent other containers being restarted. An error and a map[string]error are returned. If the error is not nil and the map is nil, an error was encountered before any containers were restarted. If map is not nil, an error was encountered when restarting one or more containers. The container ID is mapped to the error encountered. The error is set to ErrPodPartialFail. If both error and the map are nil, all containers were restarted without error.

func (*Pod) SharesCgroup

func (p *Pod) SharesCgroup() bool

SharesCgroup returns whether containers in the pod will default to this pod's cgroup instead of the default libpod parent

func (*Pod) SharesIPC

func (p *Pod) SharesIPC() bool

SharesIPC returns whether containers in pod default to use IPC namespace of first container in pod

func (*Pod) SharesMount

func (p *Pod) SharesMount() bool

SharesMount returns whether containers in pod default to use PID namespace of first container in pod

func (*Pod) SharesNamespaces

func (p *Pod) SharesNamespaces() bool

SharesNamespaces checks if the pod has any kernel namespaces set as shared. An infra container will not be created if no kernel namespaces are shared.

func (*Pod) SharesNet

func (p *Pod) SharesNet() bool

SharesNet returns whether containers in pod default to use network namespace of first container in pod

func (*Pod) SharesPID

func (p *Pod) SharesPID() bool

SharesPID returns whether containers in pod default to use PID namespace of first container in pod

func (*Pod) SharesUTS

func (p *Pod) SharesUTS() bool

SharesUTS returns whether containers in pod default to use UTS namespace of first container in pod

func (*Pod) SharesUser

func (p *Pod) SharesUser() bool

SharesUser returns whether containers in pod default to use user namespace of first container in pod

func (*Pod) Start

func (p *Pod) Start(ctx context.Context) (map[string]error, error)

Start starts all containers within a pod. It combines the effects of Init() and Start() on a container. If a container has already been initialized it will be started, otherwise it will be initialized then started. Containers that are already running or have been paused are ignored All containers are started independently, in order dictated by their dependencies. An error and a map[string]error are returned. If the error is not nil and the map is nil, an error was encountered before any containers were started. If map is not nil, an error was encountered when starting one or more containers. The container ID is mapped to the error encountered. The error is set to ErrPodPartialFail. If both error and the map are nil, all containers were started successfully.

func (*Pod) Status

func (p *Pod) Status() (map[string]define.ContainerStatus, error)

Status gets the status of all containers in the pod. Returns a map of Container ID to Container Status.

func (*Pod) Stop

func (p *Pod) Stop(ctx context.Context, cleanup bool) (map[string]error, error)

Stop stops all containers within a pod without a timeout. It assumes -1 for a timeout.

func (*Pod) StopWithTimeout

func (p *Pod) StopWithTimeout(ctx context.Context, cleanup bool, timeout int) (map[string]error, error)

StopWithTimeout stops all containers within a pod that are not already stopped Each container will use its own stop timeout. Only running containers will be stopped. Paused, stopped, or created containers will be ignored. If cleanup is true, mounts and network namespaces will be cleaned up after the container is stopped. All containers are stopped independently. An error stopping one container will not prevent other containers being stopped. An error and a map[string]error are returned. If the error is not nil and the map is nil, an error was encountered before any containers were stopped. If map is not nil, an error was encountered when stopping one or more containers. The container ID is mapped to the error encountered. The error is set to ErrPodPartialFail. If both error and the map are nil, all containers were stopped without error.

func (*Pod) Unpause

func (p *Pod) Unpause(ctx context.Context) (map[string]error, error)

Unpause unpauses all containers within a pod that are running. Only paused containers will be unpaused. Running, stopped, or created containers will be ignored. All containers are unpaused independently. An error unpausing one container will not prevent other containers being unpaused. An error and a map[string]error are returned. If the error is not nil and the map is nil, an error was encountered before any containers were unpaused. If map is not nil, an error was encountered when unpausing one or more containers. The container ID is mapped to the error encountered. The error is set to ErrPodPartialFail. If both error and the map are nil, all containers were unpaused without error.

func (*Pod) UserNSMode added in v3.4.0

func (p *Pod) UserNSMode() string

PidMode returns the PID mode given by the user ex: pod, private...

type PodConfig

type PodConfig struct {
	ID   string `json:"id"`
	Name string `json:"name"`
	// Namespace the pod is in
	Namespace string `json:"namespace,omitempty"`

	Hostname string `json:"hostname,omitempty"`

	// Labels contains labels applied to the pod
	Labels map[string]string `json:"labels"`
	// CgroupParent contains the pod's CGroup parent
	CgroupParent string `json:"cgroupParent"`
	// UsePodCgroup indicates whether the pod will create its own CGroup and
	// join containers to it.
	// If true, all containers joined to the pod will use the pod cgroup as
	// their cgroup parent, and cannot set a different cgroup parent
	UsePodCgroup bool `json:"sharesCgroup,omitempty"`

	// The following UsePod{kernelNamespace} indicate whether the containers
	// in the pod will inherit the namespace from the first container in the pod.
	UsePodPID      bool `json:"sharesPid,omitempty"`
	UsePodIPC      bool `json:"sharesIpc,omitempty"`
	UsePodNet      bool `json:"sharesNet,omitempty"`
	UsePodMount    bool `json:"sharesMnt,omitempty"`
	UsePodUser     bool `json:"sharesUser,omitempty"`
	UsePodUTS      bool `json:"sharesUts,omitempty"`
	UsePodCgroupNS bool `json:"sharesCgroupNS,omitempty"`

	HasInfra bool `json:"hasInfra,omitempty"`

	// Time pod was created
	CreatedTime time.Time `json:"created"`

	// CreateCommand is the full command plus arguments of the process the
	// container has been created with.
	CreateCommand []string `json:"CreateCommand,omitempty"`

	// ID of the pod's lock
	LockID uint32 `json:"lockID"`
}

PodConfig represents a pod's static configuration

type PodContainerStats

type PodContainerStats struct {
	Pod            *Pod
	ContainerStats map[string]*define.ContainerStats
}

PodContainerStats is an organization struct for pods and their containers

type PodCreateOption

type PodCreateOption func(*Pod) error

A PodCreateOption is a functional option which alters the Pod created by NewPod

func WithInfraConmonPidFile

func WithInfraConmonPidFile(path string, infraSpec *specgen.SpecGenerator) PodCreateOption

WithInfraConmonPidFile sets the path to a custom conmon PID file for the infra container.

func WithInfraContainer

func WithInfraContainer() PodCreateOption

WithInfraContainer tells the pod to create a pause container

func WithPodCgroup

func WithPodCgroup() PodCreateOption

WithPodCgroup tells containers in this pod to use the cgroup namespace created for this pod. Containers in a pod will inherit the kernel namespaces from the first container added.

func WithPodCgroupParent

func WithPodCgroupParent(path string) PodCreateOption

WithPodCgroupParent sets the Cgroup Parent of the pod.

func WithPodCgroups

func WithPodCgroups() PodCreateOption

WithPodCgroups tells containers in this pod to use the cgroup created for this pod. This can still be overridden at the container level by explicitly specifying a CGroup parent.

func WithPodCreateCommand

func WithPodCreateCommand(createCmd []string) PodCreateOption

WithPodCreateCommand adds the full command plus arguments of the current process to the pod config.

func WithPodHostname

func WithPodHostname(hostname string) PodCreateOption

WithPodHostname sets the hostname of the pod.

func WithPodIPC

func WithPodIPC() PodCreateOption

WithPodIPC tells containers in this pod to use the ipc namespace created for this pod. Containers in a pod will inherit the kernel namespaces from the first container added.

func WithPodLabels

func WithPodLabels(labels map[string]string) PodCreateOption

WithPodLabels sets the labels of a pod.

func WithPodMount

func WithPodMount() PodCreateOption

WithPodMount tells containers in this pod to use the mount namespace created for this pod. Containers in a pod will inherit the kernel namespaces from the first container added. TODO implement WithMountNSFrom, so WithMountNsFromPod functions properly Then this option can be added on the pod level

func WithPodName

func WithPodName(name string) PodCreateOption

WithPodName sets the name of the pod.

func WithPodNamespace

func WithPodNamespace(ns string) PodCreateOption

WithPodNamespace sets the namespace for the created pod. Namespaces are used to create separate views of Podman's state - runtimes can join a specific namespace and see only containers and pods in that namespace. Empty string namespaces are allowed, and correspond to a lack of namespace. Containers must belong to the same namespace as the pod they join.

func WithPodNet

func WithPodNet() PodCreateOption

WithPodNet tells containers in this pod to use the network namespace created for this pod. Containers in a pod will inherit the kernel namespaces from the first container added.

func WithPodPID

func WithPodPID() PodCreateOption

WithPodPID tells containers in this pod to use the pid namespace created for this pod. Containers in a pod will inherit the kernel namespaces from the first container added.

func WithPodUTS

func WithPodUTS() PodCreateOption

WithPodUTS tells containers in this pod to use the uts namespace created for this pod. Containers in a pod will inherit the kernel namespaces from the first container added.

func WithPodUser

func WithPodUser() PodCreateOption

WithPodUser tells containers in this pod to use the user namespace created for this pod. Containers in a pod will inherit the kernel namespaces from the first container added. TODO implement WithUserNSFrom, so WithUserNsFromPod functions properly Then this option can be added on the pod level

type PodFilter

type PodFilter func(*Pod) bool

PodFilter is a function to determine whether a pod is included in command output. Pods to be outputted are tested using the function. A true return will include the pod, a false return will exclude it.

type RootlessCNI added in v3.2.0

type RootlessCNI struct {
	Lock lockfile.Locker
	// contains filtered or unexported fields
}

func (*RootlessCNI) Cleanup added in v3.2.0

func (r *RootlessCNI) Cleanup(runtime *Runtime) error

Cleanup the rootless cni namespace if needed. It checks if we have running containers with the bridge network mode. Cleanup() will try to lock RootlessCNI, therefore you have to call it with an unlocked

func (*RootlessCNI) Do added in v3.2.0

func (r *RootlessCNI) Do(toRun func() error) error

Do - run the given function in the rootless cni ns. It does not lock the rootlessCNI lock, the caller should only lock when needed, e.g. for cni operations.

type Runtime

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

Runtime is the core libpod runtime

func NewRuntime

func NewRuntime(ctx context.Context, options ...RuntimeOption) (*Runtime, error)

NewRuntime creates a new container runtime Options can be passed to override the default configuration for the runtime

func NewRuntimeFromConfig

func NewRuntimeFromConfig(ctx context.Context, userConfig *config.Config, options ...RuntimeOption) (*Runtime, error)

NewRuntimeFromConfig creates a new container runtime using the given configuration file for its default configuration. Passed RuntimeOption functions can be used to mutate this configuration further. An error will be returned if the configuration file at the given path does not exist or cannot be loaded

func (*Runtime) AddInfra added in v3.4.0

func (r *Runtime) AddInfra(ctx context.Context, pod *Pod, infraCtr *Container) (*Pod, error)

AddInfra adds the created infra container to the pod state

func (*Runtime) Build

func (r *Runtime) Build(ctx context.Context, options buildahDefine.BuildOptions, dockerfiles ...string) (string, reference.Canonical, error)

Build adds the runtime to the imagebuildah call

func (*Runtime) ConnectContainerToNetwork

func (r *Runtime) ConnectContainerToNetwork(nameOrID, netName string, aliases []string) error

ConnectContainerToNetwork connects a container to a CNI network

func (*Runtime) DefaultOCIRuntime added in v3.4.0

func (r *Runtime) DefaultOCIRuntime() OCIRuntime

DefaultOCIRuntime return copy of Default OCI Runtime

func (*Runtime) DeferredShutdown

func (r *Runtime) DeferredShutdown(force bool)

DeferredShutdown shuts down the runtime without exposing any errors. This is only meant to be used when the runtime is being shutdown within a defer statement; else use Shutdown

func (*Runtime) DisconnectContainerFromNetwork

func (r *Runtime) DisconnectContainerFromNetwork(nameOrID, netName string, force bool) error

DisconnectContainerFromNetwork removes a container from its CNI network

func (*Runtime) EnableLabeling

func (r *Runtime) EnableLabeling() bool

func (*Runtime) Events

func (r *Runtime) Events(ctx context.Context, options events.ReadOptions) error

Events is a wrapper function for everyone to begin tailing the events log with options

func (*Runtime) EvictContainer

func (r *Runtime) EvictContainer(ctx context.Context, idOrName string, removeVolume bool) (string, error)

EvictContainer removes the given container partial or full ID or name, and returns the full ID of the evicted container and any error encountered. It should be used to remove a container when obtaining a Container struct pointer has failed. Running container will not be stopped. If removeVolume is specified, named volumes used by the container will be removed also if and only if the container is the sole user.

func (*Runtime) GetAllContainers

func (r *Runtime) GetAllContainers() ([]*Container, error)

GetAllContainers is a helper function for GetContainers

func (*Runtime) GetAllPods

func (r *Runtime) GetAllPods() ([]*Pod, error)

GetAllPods retrieves all pods

func (*Runtime) GetAllVolumes

func (r *Runtime) GetAllVolumes() ([]*Volume, error)

GetAllVolumes retrieves all the volumes

func (*Runtime) GetConfig

func (r *Runtime) GetConfig() (*config.Config, error)

GetConfig returns a copy of the configuration used by the runtime

func (*Runtime) GetContainer

func (r *Runtime) GetContainer(id string) (*Container, error)

GetContainer retrieves a container by its ID

func (*Runtime) GetContainers

func (r *Runtime) GetContainers(filters ...ContainerFilter) ([]*Container, error)

GetContainers retrieves all containers from the state Filters can be provided which will determine what containers are included in the output. Multiple filters are handled by ANDing their output, so only containers matching all filters are returned

func (*Runtime) GetContainersByList

func (r *Runtime) GetContainersByList(containers []string) ([]*Container, error)

GetContainersByList is a helper function for GetContainers which takes a []string of container IDs or names

func (*Runtime) GetContainersWithoutLock

func (r *Runtime) GetContainersWithoutLock(filters ...ContainerFilter) ([]*Container, error)

GetContainersWithoutLock is same as GetContainers but without lock

func (*Runtime) GetDiff

func (r *Runtime) GetDiff(from, to string, diffType define.DiffType) ([]archive.Change, error)

GetDiff returns the differences between the two images, layers, or containers

func (*Runtime) GetEvents

func (r *Runtime) GetEvents(ctx context.Context, filters []string) ([]*events.Event, error)

GetEvents reads the event log and returns events based on input filters

func (*Runtime) GetExecDiedEvent added in v3.3.0

func (r *Runtime) GetExecDiedEvent(ctx context.Context, nameOrID, execSessionID string) (*events.Event, error)

GetExecDiedEvent takes a container name or ID, exec session ID, and returns that exec session's Died event (if it has already occurred).

func (*Runtime) GetExecSessionContainer

func (r *Runtime) GetExecSessionContainer(id string) (*Container, error)

GetExecSessionContainer gets the container that a given exec session ID is attached to.

func (*Runtime) GetHostDistributionInfo

func (r *Runtime) GetHostDistributionInfo() define.DistributionInfo

GetHostDistributionInfo returns a map containing the host's distribution and version

func (*Runtime) GetLastContainerEvent

func (r *Runtime) GetLastContainerEvent(ctx context.Context, nameOrID string, containerEvent events.Status) (*events.Event, error)

GetLastContainerEvent takes a container name or ID and an event status and returns the last occurrence of the container event

func (*Runtime) GetLatestContainer

func (r *Runtime) GetLatestContainer() (*Container, error)

GetLatestContainer returns a container object of the latest created container.

func (*Runtime) GetLatestPod

func (r *Runtime) GetLatestPod() (*Pod, error)

GetLatestPod returns a pod object of the latest created pod.

func (*Runtime) GetName

func (r *Runtime) GetName(id string) (string, error)

GetName retrieves the name associated with a given full ID. This works for both containers and pods, and does not distinguish between the two. If the given ID does not correspond to any existing Pod or Container, ErrNoSuchCtr is returned.

func (*Runtime) GetOCIRuntimePath

func (r *Runtime) GetOCIRuntimePath() string

GetOCIRuntimePath retrieves the path of the default OCI runtime.

func (*Runtime) GetPod

func (r *Runtime) GetPod(id string) (*Pod, error)

GetPod retrieves a pod by its ID

func (*Runtime) GetRootlessCNINetNs added in v3.2.0

func (r *Runtime) GetRootlessCNINetNs(new bool) (*RootlessCNI, error)

GetRootlessCNINetNs returns the rootless cni object. If create is set to true the rootless cni namespace will be created if it does not exists already. If called as root it returns always nil. On success the returned RootlessCNI lock is locked and must be unlocked by the caller.

func (*Runtime) GetRunningContainers

func (r *Runtime) GetRunningContainers() ([]*Container, error)

GetRunningContainers is a helper function for GetContainers

func (*Runtime) GetRunningPods

func (r *Runtime) GetRunningPods() ([]*Pod, error)

GetRunningPods returns an array of running pods

func (*Runtime) GetSecretsStorageDir added in v3.1.0

func (r *Runtime) GetSecretsStorageDir() string

GetSecretsStoreageDir returns the directory that the secrets manager should take

func (*Runtime) GetVolume

func (r *Runtime) GetVolume(name string) (*Volume, error)

GetVolume retrieves a volume given its full name.

func (*Runtime) HasContainer

func (r *Runtime) HasContainer(id string) (bool, error)

HasContainer checks if a container with the given ID is present

func (*Runtime) HasPod

func (r *Runtime) HasPod(id string) (bool, error)

HasPod checks to see if a pod with the given ID exists

func (*Runtime) HasVolume

func (r *Runtime) HasVolume(name string) (bool, error)

HasVolume checks to see if a volume with the given name exists

func (*Runtime) HealthCheck

func (r *Runtime) HealthCheck(name string) (define.HealthCheckStatus, error)

HealthCheck verifies the state and validity of the healthcheck configuration on the container and then executes the healthcheck

func (*Runtime) Info

func (r *Runtime) Info() (*define.Info, error)

Info returns the store and host information

func (*Runtime) IsBuildahContainer

func (r *Runtime) IsBuildahContainer(id string) (bool, error)

func (*Runtime) IsStorageContainerMounted

func (r *Runtime) IsStorageContainerMounted(id string) (bool, string, error)

MountedStorageContainer returns whether a storage container is mounted along with the mount path

func (*Runtime) LibimageRuntime added in v3.2.0

func (r *Runtime) LibimageRuntime() *libimage.Runtime

LibimageRuntime ... to allow for a step-by-step migration to libimage.

func (*Runtime) ListStorageContainers

func (r *Runtime) ListStorageContainers() ([]*StorageContainer, error)

ListStorageContainers lists all containers visible to c/storage.

func (*Runtime) Log

func (r *Runtime) Log(ctx context.Context, containers []*Container, options *logs.LogOptions, logChannel chan *logs.LogLine) error

Log is a runtime function that can read one or more container logs.

func (*Runtime) LookupContainer

func (r *Runtime) LookupContainer(idOrName string) (*Container, error)

LookupContainer looks up a container by its name or a partial ID If a partial ID is not unique, an error will be returned

func (*Runtime) LookupContainerID added in v3.3.0

func (r *Runtime) LookupContainerID(idOrName string) (string, error)

LookupContainerId looks up a container id by its name or a partial ID If a partial ID is not unique, an error will be returned

func (*Runtime) LookupPod

func (r *Runtime) LookupPod(idOrName string) (*Pod, error)

LookupPod retrieves a pod by its name or a partial ID If a partial ID is not unique, an error will be returned

func (*Runtime) LookupVolume

func (r *Runtime) LookupVolume(name string) (*Volume, error)

LookupVolume retrieves a volume by unambiguous partial name.

func (*Runtime) MountStorageContainer

func (r *Runtime) MountStorageContainer(id string) (string, error)

MountStorageContainer mounts the storage container's root filesystem

func (*Runtime) NewContainer

func (r *Runtime) NewContainer(ctx context.Context, rSpec *spec.Spec, spec *specgen.SpecGenerator, infra bool, options ...CtrCreateOption) (*Container, error)

NewContainer creates a new container from a given OCI config.

func (*Runtime) NewPod

func (r *Runtime) NewPod(ctx context.Context, p specgen.PodSpecGenerator, options ...PodCreateOption) (_ *Pod, deferredErr error)

NewPod makes a new, empty pod

func (*Runtime) NewVolume

func (r *Runtime) NewVolume(ctx context.Context, options ...VolumeCreateOption) (*Volume, error)

NewVolume creates a new empty volume

func (*Runtime) Pods

func (r *Runtime) Pods(filters ...PodFilter) ([]*Pod, error)

Pods retrieves all pods Filters can be provided which will determine which pods are included in the output. Multiple filters are handled by ANDing their output, so only pods matching all filters are returned

func (*Runtime) PrepareVolumeOnCreateContainer added in v3.3.0

func (r *Runtime) PrepareVolumeOnCreateContainer(ctx context.Context, ctr *Container) error

func (*Runtime) PruneContainers

func (r *Runtime) PruneContainers(filterFuncs []ContainerFilter) ([]*reports.PruneReport, error)

PruneContainers removes stopped and exited containers from localstorage. A set of optional filters can be provided to be more granular.

func (*Runtime) PrunePods

func (r *Runtime) PrunePods(ctx context.Context) (map[string]error, error)

PrunePods removes unused pods and their containers from local storage.

func (*Runtime) PruneVolumes

func (r *Runtime) PruneVolumes(ctx context.Context, filterFuncs []VolumeFilter) ([]*reports.PruneReport, error)

PruneVolumes removes unused volumes from the system

func (*Runtime) Reload

func (r *Runtime) Reload() error

Reload reloads the configurations files

func (*Runtime) RemoveContainer

func (r *Runtime) RemoveContainer(ctx context.Context, c *Container, force bool, removeVolume bool) error

RemoveContainer removes the given container If force is specified, the container will be stopped first If removeVolume is specified, named volumes used by the container will be removed also if and only if the container is the sole user Otherwise, RemoveContainer will return an error if the container is running

func (*Runtime) RemoveContainersForImageCallback added in v3.2.0

func (r *Runtime) RemoveContainersForImageCallback(ctx context.Context) libimage.RemoveContainerFunc

RemoveContainersForImageCallback returns a callback that can be used in `libimage`. When forcefully removing images, containers using the image should be removed as well. The callback allows for more graceful removal as we can use the libpod-internal removal logic.

func (*Runtime) RemovePod

func (r *Runtime) RemovePod(ctx context.Context, p *Pod, removeCtrs, force bool) error

RemovePod removes a pod If removeCtrs is specified, containers will be removed Otherwise, a pod that is not empty will return an error and not be removed If force is specified with removeCtrs, all containers will be stopped before being removed Otherwise, the pod will not be removed if any containers are running

func (*Runtime) RemoveStorageContainer

func (r *Runtime) RemoveStorageContainer(idOrName string, force bool) error

RemoveStorageContainer removes a container from c/storage. The container WILL NOT be removed if it exists in libpod. Accepts ID or full name of container. If force is set, the container will be unmounted first to ensure removal.

func (*Runtime) RemoveVolume

func (r *Runtime) RemoveVolume(ctx context.Context, v *Volume, force bool) error

RemoveVolume removes a volumes

func (*Runtime) RenameContainer

func (r *Runtime) RenameContainer(ctx context.Context, ctr *Container, newName string) (*Container, error)

RenameContainer renames the given container. Returns a copy of the container that has been renamed if successful.

func (*Runtime) Reset

func (r *Runtime) Reset(ctx context.Context) error

Reset removes all storage

func (*Runtime) RestoreContainer

func (r *Runtime) RestoreContainer(ctx context.Context, rSpec *spec.Spec, config *ContainerConfig) (*Container, error)

RestoreContainer re-creates a container from an imported checkpoint

func (*Runtime) RunRoot added in v3.3.0

func (r *Runtime) RunRoot() string

RunRoot retrieves the current c/storage temporary directory in use by Libpod.

func (*Runtime) SavePod added in v3.4.0

func (r *Runtime) SavePod(pod *Pod) error

SavePod is a helper function to save the pod state from outside of libpod

func (*Runtime) SecretsManager added in v3.2.3

func (r *Runtime) SecretsManager() (*secrets.SecretsManager, error)

SecretsManager returns the directory that the secrets manager should take

func (*Runtime) Shutdown

func (r *Runtime) Shutdown(force bool) error

Shutdown shuts down the runtime and associated containers and storage If force is true, containers and mounted storage will be shut down before cleaning up; if force is false, an error will be returned if there are still containers running or mounted

func (*Runtime) StorageConfig

func (r *Runtime) StorageConfig() storage.StoreOptions

StorageConfig retrieves the storage options for the container runtime

func (*Runtime) StorageContainer

func (r *Runtime) StorageContainer(idOrName string) (*storage.Container, error)

func (*Runtime) StorageContainers

func (r *Runtime) StorageContainers() ([]storage.Container, error)

StorageContainers returns a list of containers from containers/storage that are not currently known to Podman.

func (*Runtime) SystemContext

func (r *Runtime) SystemContext() *types.SystemContext

SystemContext returns the imagecontext

func (*Runtime) TmpDir

func (r *Runtime) TmpDir() (string, error)

TmpDir gets the current Libpod temporary files directory.

func (*Runtime) UnmountStorageContainer

func (r *Runtime) UnmountStorageContainer(id string, force bool) (bool, error)

UnmountStorageContainer unmounts the storage container's root filesystem

func (*Runtime) Volumes

func (r *Runtime) Volumes(filters ...VolumeFilter) ([]*Volume, error)

Volumes retrieves all volumes Filters can be provided which will determine which volumes are included in the output. If multiple filters are used, a volume will be returned if any of the filters are matched

func (*Runtime) WithPod

func (r *Runtime) WithPod(pod *Pod) CtrCreateOption

WithPod adds the container to a pod. Containers which join a pod can only join the Linux namespaces of other containers in the same pod. Containers can only join pods in the same libpod namespace.

type RuntimeContainerMetadata

type RuntimeContainerMetadata struct {
	// The provided name and the ID of the image that was used to
	// instantiate the container.
	ImageName string `json:"image-name"` // Applicable to both PodSandboxes and Containers
	ImageID   string `json:"image-id"`   // Applicable to both PodSandboxes and Containers
	// The container's name, which for an infrastructure container is usually PodName + "-infra".
	ContainerName string `json:"name"`                 // Applicable to both PodSandboxes and Containers, mandatory
	CreatedAt     int64  `json:"created-at"`           // Applicable to both PodSandboxes and Containers
	MountLabel    string `json:"mountlabel,omitempty"` // Applicable to both PodSandboxes and Containers
}

RuntimeContainerMetadata is the structure that we encode as JSON and store in the metadata field of storage.Container objects. It is used for specifying attributes containers when they are being created, and allows a container's MountLabel, and possibly other values, to be modified in one read/write cycle via calls to storageService.ContainerMetadata, RuntimeContainerMetadata.SetMountLabel, and storageService.SetContainerMetadata.

func (*RuntimeContainerMetadata) SetMountLabel

func (metadata *RuntimeContainerMetadata) SetMountLabel(mountLabel string)

SetMountLabel updates the mount label held by a RuntimeContainerMetadata object.

type RuntimeOption

type RuntimeOption func(*Runtime) error

A RuntimeOption is a functional option which alters the Runtime created by NewRuntime

func WithCNIConfigDir

func WithCNIConfigDir(dir string) RuntimeOption

WithCNIConfigDir sets the CNI configuration directory.

func WithCNIPluginDir

func WithCNIPluginDir(dir string) RuntimeOption

WithCNIPluginDir sets the CNI plugins directory.

func WithCgroupManager

func WithCgroupManager(manager string) RuntimeOption

WithCgroupManager specifies the manager implementation name which is used to handle cgroups for containers. Current valid values are "cgroupfs" and "systemd".

func WithConmonEnv

func WithConmonEnv(environment []string) RuntimeOption

WithConmonEnv specifies the environment variable list for the conmon process.

func WithConmonPath

func WithConmonPath(path string) RuntimeOption

WithConmonPath specifies the path to the conmon binary which manages the runtime.

func WithDefaultInfraCommand

func WithDefaultInfraCommand(cmd string) RuntimeOption

WithDefaultInfraCommand sets the command to run on pause container start up.

func WithDefaultInfraImage

func WithDefaultInfraImage(img string) RuntimeOption

WithDefaultInfraImage sets the infra image for libpod. An infra image is used for inter-container kernel namespace sharing within a pod. Typically, an infra container is lightweight and is there to reap zombie processes within its pid namespace.

func WithDefaultInfraName added in v3.3.0

func WithDefaultInfraName(name string) RuntimeOption

WithDefaultInfraName sets the infra container name for a single pod.

func WithDefaultMountsFile

func WithDefaultMountsFile(mountsFile string) RuntimeOption

WithDefaultMountsFile sets the file to look at for default mounts (mainly secrets). Note we are not saving this in the database as it is for testing purposes only.

func WithDefaultTransport

func WithDefaultTransport(defaultTransport string) RuntimeOption

WithDefaultTransport sets the default transport for retrieving images.

func WithEnableSDNotify

func WithEnableSDNotify() RuntimeOption

WithEnableSDNotify sets a runtime option so we know whether to disable socket/FD listening

func WithEventsLogger

func WithEventsLogger(logger string) RuntimeOption

WithEventsLogger sets the events backend to use. Currently supported values are "file" for file backend and "journald" for journald backend.

func WithHooksDir

func WithHooksDir(hooksDirs ...string) RuntimeOption

WithHooksDir sets the directories to look for OCI runtime hook configuration.

func WithMigrate

func WithMigrate() RuntimeOption

WithMigrate instructs libpod to migrate container configurations to account for changes between Engine versions. All running containers will be stopped during a migration, then restarted after the migration is complete.

func WithMigrateRuntime

func WithMigrateRuntime(requestedRuntime string) RuntimeOption

WithMigrateRuntime instructs Engine to change the default OCI runtime on all containers during a migration. This is not used if `MigrateRuntime()` is not also passed. Engine makes no promises that your containers continue to work with the new runtime - migrations between dissimilar runtimes may well break things. Use with caution.

func WithNamespace

func WithNamespace(ns string) RuntimeOption

WithNamespace sets the namespace for libpod. Namespaces are used to create scopes to separate containers and pods in the state. When namespace is set, libpod will only view containers and pods in the same namespace. All containers and pods created will default to the namespace set here. A namespace of "", the empty string, is equivalent to no namespace, and all containers and pods will be visible.

func WithNetworkCmdPath

func WithNetworkCmdPath(path string) RuntimeOption

WithNetworkCmdPath specifies the path to the slirp4netns binary which manages the runtime.

func WithNoPivotRoot

func WithNoPivotRoot() RuntimeOption

WithNoPivotRoot sets the runtime to use MS_MOVE instead of PIVOT_ROOT when starting containers.

func WithNoStore

func WithNoStore() RuntimeOption

WithNoStore sets a bool on the runtime that we do not need any containers storage.

func WithOCIRuntime

func WithOCIRuntime(runtime string) RuntimeOption

WithOCIRuntime specifies an OCI runtime to use for running containers.

func WithRegistriesConf

func WithRegistriesConf(path string) RuntimeOption

WithRegistriesConf configures the runtime to always use specified registries.conf for image processing.

func WithRenumber

func WithRenumber() RuntimeOption

WithRenumber instructs libpod to perform a lock renumbering while initializing. This will handle migrations from early versions of libpod with file locks to newer versions with SHM locking, as well as changes in the number of configured locks.

func WithRuntimeFlags

func WithRuntimeFlags(runtimeFlags []string) RuntimeOption

WithRuntimeFlags adds the global runtime flags to the container config

func WithSignaturePolicy

func WithSignaturePolicy(path string) RuntimeOption

WithSignaturePolicy specifies the path of a file which decides how trust is managed for images we've pulled. If this is not specified, the system default configuration will be used instead.

func WithStateType

func WithStateType(storeType config.RuntimeStateStore) RuntimeOption

WithStateType sets the backing state implementation for libpod. Please note that information is not portable between backing states. As such, if this differs between two libpods running on the same system, they will not share containers, and unspecified behavior may occur.

func WithStaticDir

func WithStaticDir(dir string) RuntimeOption

WithStaticDir sets the directory that static runtime files which persist across reboots will be stored.

func WithStorageConfig

func WithStorageConfig(config storage.StoreOptions) RuntimeOption

WithStorageConfig uses the given configuration to set up container storage. If this is not specified, the system default configuration will be used instead.

func WithTmpDir

func WithTmpDir(dir string) RuntimeOption

WithTmpDir sets the directory that temporary runtime files which are not expected to survive across reboots will be stored. This should be located on a tmpfs mount (/tmp or /run for example).

func WithVolumePath

func WithVolumePath(volPath string) RuntimeOption

WithVolumePath sets the path under which all named volumes should be created. The path changes based on whether the user is running as root or not.

type State

type State interface {
	// Close performs any pre-exit cleanup (e.g. closing database
	// connections) that may be required
	Close() error

	// Refresh clears container and pod states after a reboot
	Refresh() error

	// GetDBConfig retrieves several paths configured within the database
	// when it was created - namely, Libpod root and tmp dirs, c/storage
	// root and tmp dirs, and c/storage graph driver.
	// This is not implemented by the in-memory state, as it has no need to
	// validate runtime configuration.
	GetDBConfig() (*DBConfig, error)

	// ValidateDBConfig validates the config in the given Runtime struct
	// against paths stored in the configured database.
	// Libpod root and tmp dirs and c/storage root and tmp dirs and graph
	// driver are validated.
	// This is not implemented by the in-memory state, as it has no need to
	// validate runtime configuration that may change over multiple runs of
	// the program.
	ValidateDBConfig(runtime *Runtime) error

	// SetNamespace() sets the namespace for the store, and will determine
	// what containers are retrieved with container and pod retrieval calls.
	// A namespace of "", the empty string, acts as no namespace, and
	// containers and pods in all namespaces will be returned.
	SetNamespace(ns string) error

	// Resolve an ID into a Name. Since Podman names and IDs are globally
	// unique between Pods and Containers, the ID may belong to either a pod
	// or container. Despite this, we will always return ErrNoSuchCtr if the
	// ID does not exist.
	GetName(id string) (string, error)

	// Return a container from the database from its full ID.
	// If the container is not in the set namespace, an error will be
	// returned.
	Container(id string) (*Container, error)
	// Return a container ID from the database by full or partial ID or full
	// name.
	LookupContainerID(idOrName string) (string, error)
	// Return a container from the database by full or partial ID or full
	// name.
	// Containers not in the set namespace will be ignored.
	LookupContainer(idOrName string) (*Container, error)
	// Check if a container with the given full ID exists in the database.
	// If the container exists but is not in the set namespace, false will
	// be returned.
	HasContainer(id string) (bool, error)
	// Adds container to state.
	// The container cannot be part of a pod.
	// The container must have globally unique name and ID - pod names and
	// IDs also conflict with container names and IDs.
	// The container must be in the set namespace if a namespace has been
	// set.
	// All containers this container depends on must be part of the same
	// namespace and must not be joined to a pod.
	AddContainer(ctr *Container) error
	// Removes container from state.
	// Containers that are part of pods must use RemoveContainerFromPod.
	// The container must be part of the set namespace.
	// All dependencies must be removed first.
	// All exec sessions referencing the container must be removed first.
	RemoveContainer(ctr *Container) error
	// UpdateContainer updates a container's state from the backing store.
	// The container must be part of the set namespace.
	UpdateContainer(ctr *Container) error
	// SaveContainer saves a container's current state to the backing store.
	// The container must be part of the set namespace.
	SaveContainer(ctr *Container) error
	// ContainerInUse checks if other containers depend upon a given
	// container.
	// It returns a slice of the IDs of containers which depend on the given
	// container. If the slice is empty, no container depend on the given
	// container.
	// A container cannot be removed if other containers depend on it.
	// The container being checked must be part of the set namespace.
	ContainerInUse(ctr *Container) ([]string, error)
	// Retrieves all containers presently in state.
	// If a namespace is set, only containers within the namespace will be
	// returned.
	AllContainers() ([]*Container, error)

	// Get networks the container is currently connected to.
	GetNetworks(ctr *Container) ([]string, error)
	// Get network aliases for the given container in the given network.
	GetNetworkAliases(ctr *Container, network string) ([]string, error)
	// Get all network aliases for the given container.
	GetAllNetworkAliases(ctr *Container) (map[string][]string, error)
	// Add the container to the given network, adding the given aliases
	// (if present).
	NetworkConnect(ctr *Container, network string, aliases []string) error
	// Remove the container from the given network, removing all aliases for
	// the container in that network in the process.
	NetworkDisconnect(ctr *Container, network string) error

	// Return a container config from the database by full ID
	GetContainerConfig(id string) (*ContainerConfig, error)

	// Add creates a reference to an exec session in the database.
	// The container the exec session is attached to will be recorded.
	// The container state will not be modified.
	// The actual exec session itself is part of the container's state.
	// We assume higher-level callers will add the session by saving the
	// container's state before calling this. This only ensures that the ID
	// of the exec session is associated with the ID of the container.
	// Implementations may, but are not required to, verify that the state
	// of the given container has an exec session with the ID given.
	AddExecSession(ctr *Container, session *ExecSession) error
	// Get retrieves the container a given exec session is attached to.
	GetExecSession(id string) (string, error)
	// Remove a reference to an exec session from the database.
	// This will not modify container state to remove the exec session there
	// and instead only removes the session ID -> container ID reference
	// added by AddExecSession.
	RemoveExecSession(session *ExecSession) error
	// Get the IDs of all exec sessions attached to a given container.
	GetContainerExecSessions(ctr *Container) ([]string, error)
	// Remove all exec sessions for a single container.
	// Usually used as part of removing the container.
	// As with RemoveExecSession, container state will not be modified.
	RemoveContainerExecSessions(ctr *Container) error

	// PLEASE READ FULL DESCRIPTION BEFORE USING.
	// Rewrite a container's configuration.
	// This function breaks libpod's normal prohibition on a read-only
	// configuration, and as such should be used EXTREMELY SPARINGLY and
	// only in very specific circumstances.
	// Specifically, it is ONLY safe to use thing function to make changes
	// that result in a functionally identical configuration (migrating to
	// newer, but identical, configuration fields), or during libpod init
	// WHILE HOLDING THE ALIVE LOCK (to prevent other libpod instances from
	// being initialized).
	// Most things in config can be changed by this, but container ID and
	// name ABSOLUTELY CANNOT BE ALTERED. If you do so, there is a high
	// potential for database corruption.
	// There are a lot of capital letters and conditions here, but the short
	// answer is this: use this only very sparingly, and only if you really
	// know what you're doing.
	RewriteContainerConfig(ctr *Container, newCfg *ContainerConfig) error
	// This is a more limited version of RewriteContainerConfig, though it
	// comes with the added ability to alter a container's name. In exchange
	// it loses the ability to manipulate the container's locks.
	// It is not intended to be as restrictive as RewriteContainerConfig, in
	// that we allow it to be run while other Podman processes are running,
	// and without holding the alive lock.
	// Container ID and pod membership still *ABSOLUTELY CANNOT* be altered.
	// Also, you cannot change a container's dependencies - shared namespace
	// containers or generic dependencies - at present. This is
	// theoretically possible but not yet implemented.
	// If newName is not "" the container will be renamed to the new name.
	// The oldName parameter is only required if newName is given.
	SafeRewriteContainerConfig(ctr *Container, oldName, newName string, newCfg *ContainerConfig) error
	// PLEASE READ THE DESCRIPTION FOR RewriteContainerConfig BEFORE USING.
	// This function is identical to RewriteContainerConfig, save for the
	// fact that it is used with pods instead.
	// It is subject to the same conditions as RewriteContainerConfig.
	// Please do not use this unless you know what you're doing.
	RewritePodConfig(pod *Pod, newCfg *PodConfig) error
	// PLEASE READ THE DESCRIPTION FOR RewriteContainerConfig BEFORE USING.
	// This function is identical to RewriteContainerConfig, save for the
	// fact that it is used with volumes instead.
	// It is subject to the same conditions as RewriteContainerConfig.
	// The exception is that volumes do not have IDs, so only volume name
	// cannot be altered.
	// Please do not use this unless you know what you're doing.
	RewriteVolumeConfig(volume *Volume, newCfg *VolumeConfig) error

	// Accepts full ID of pod.
	// If the pod given is not in the set namespace, an error will be
	// returned.
	Pod(id string) (*Pod, error)
	// Accepts full or partial IDs (as long as they are unique) and names.
	// Pods not in the set namespace are ignored.
	LookupPod(idOrName string) (*Pod, error)
	// Checks if a pod with the given ID is present in the state.
	// If the given pod is not in the set namespace, false is returned.
	HasPod(id string) (bool, error)
	// Check if a pod has a container with the given ID.
	// The pod must be part of the set namespace.
	PodHasContainer(pod *Pod, ctrID string) (bool, error)
	// Get the IDs of all containers in a pod.
	// The pod must be part of the set namespace.
	PodContainersByID(pod *Pod) ([]string, error)
	// Get all the containers in a pod.
	// The pod must be part of the set namespace.
	PodContainers(pod *Pod) ([]*Container, error)
	// Adds pod to state.
	// The pod must be part of the set namespace.
	// The pod's name and ID must be globally unique.
	AddPod(pod *Pod) error
	// Removes pod from state.
	// Only empty pods can be removed from the state.
	// The pod must be part of the set namespace.
	RemovePod(pod *Pod) error
	// Remove all containers from a pod.
	// Used to simultaneously remove containers that might otherwise have
	// dependency issues.
	// Will fail if a dependency outside the pod is encountered.
	// The pod must be part of the set namespace.
	RemovePodContainers(pod *Pod) error
	// AddContainerToPod adds a container to an existing pod.
	// The container given will be added to the state and the pod.
	// The container and its dependencies must be part of the given pod,
	// and the given pod's namespace.
	// The pod must be part of the set namespace.
	// The pod must already exist in the state.
	// The container's name and ID must be globally unique.
	AddContainerToPod(pod *Pod, ctr *Container) error
	// RemoveContainerFromPod removes a container from an existing pod.
	// The container will also be removed from the state.
	// The container must be in the given pod, and the pod must be in the
	// set namespace.
	RemoveContainerFromPod(pod *Pod, ctr *Container) error
	// UpdatePod updates a pod's state from the database.
	// The pod must be in the set namespace.
	UpdatePod(pod *Pod) error
	// SavePod saves a pod's state to the database.
	// The pod must be in the set namespace.
	SavePod(pod *Pod) error
	// Retrieves all pods presently in state.
	// If a namespace has been set, only pods in that namespace will be
	// returned.
	AllPods() ([]*Pod, error)

	// Volume accepts full name of volume
	// If the volume doesn't exist, an error will be returned
	Volume(volName string) (*Volume, error)
	// LookupVolume accepts an unambiguous partial name or full name of a
	// volume. Ambiguous names will result in an error.
	LookupVolume(name string) (*Volume, error)
	// HasVolume returns true if volName exists in the state,
	// otherwise it returns false
	HasVolume(volName string) (bool, error)
	// VolumeInUse goes through the container dependencies of a volume
	// and checks if the volume is being used by any container. If it is
	// a slice of container IDs using the volume is returned
	VolumeInUse(volume *Volume) ([]string, error)
	// AddVolume adds the specified volume to state. The volume's name
	// must be unique within the list of existing volumes
	AddVolume(volume *Volume) error
	// RemoveVolume removes the specified volume.
	// Only volumes that have no container dependencies can be removed
	RemoveVolume(volume *Volume) error
	// UpdateVolume updates the volume's state from the database.
	UpdateVolume(volume *Volume) error
	// SaveVolume saves a volume's state to the database.
	SaveVolume(volume *Volume) error
	// AllVolumes returns all the volumes available in the state
	AllVolumes() ([]*Volume, error)
}

State is a storage backend for libpod's current state. A State is only initialized once per instance of libpod. As such, initialization methods for State implementations may safely assume they will be run as a singleton. For all container and pod retrieval methods, a State must retrieve the Configuration struct of the container or pod and include it in the returned struct. The State of the container or pod may optionally be included as well, but this is not a requirement. As such, all containers and pods must be synced with the database via the UpdateContainer and UpdatePod calls before any state-specific information is retrieved after they are pulled from the database. Generally speaking, the syncContainer() call should be run at the beginning of all API operations, which will silently handle this.

func NewBoltState

func NewBoltState(path string, runtime *Runtime) (State, error)

NewBoltState creates a new bolt-backed state database

type StorageContainer

type StorageContainer struct {
	ID              string
	Names           []string
	Image           string
	CreateTime      time.Time
	PresentInLibpod bool
}

StorageContainer represents a container present in c/storage but not in libpod.

type Volume

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

Volume is a libpod named volume. Named volumes may be shared by multiple containers, and may be created using more complex options than normal bind mounts. They may be backed by a mounted filesystem on the host.

func (*Volume) Anonymous

func (v *Volume) Anonymous() bool

Anonymous returns whether this volume is anonymous. Anonymous volumes were created with a container, and will be removed when that container is removed.

func (*Volume) Config

func (v *Volume) Config() (*VolumeConfig, error)

Config returns the volume's configuration.

func (*Volume) CreatedTime

func (v *Volume) CreatedTime() time.Time

CreatedTime returns the time the volume was created at. It was not tracked for some time, so older volumes may not contain one.

func (*Volume) Driver

func (v *Volume) Driver() string

Driver retrieves the volume's driver.

func (*Volume) GID

func (v *Volume) GID() (int, error)

GID returns the GID the volume will be created as.

func (*Volume) GenerateForKube added in v3.2.0

func (v *Volume) GenerateForKube() *v1.PersistentVolumeClaim

GenerateForKube generates a v1.PersistentVolumeClaim from a libpod volume.

func (*Volume) Inspect

func (v *Volume) Inspect() (*define.InspectVolumeData, error)

Inspect provides detailed information about the configuration of the given volume.

func (*Volume) IsDangling

func (v *Volume) IsDangling() (bool, error)

IsDangling returns whether this volume is dangling (unused by any containers).

func (*Volume) Labels

func (v *Volume) Labels() map[string]string

Labels returns the volume's labels

func (*Volume) MountCount added in v3.4.0

func (v *Volume) MountCount() (uint, error)

MountCount returns the volume's mountcount on the host from state Useful in determining if volume is using plugin or a filesystem mount and its mount

func (*Volume) MountPoint

func (v *Volume) MountPoint() (string, error)

MountPoint returns the volume's mountpoint on the host

func (*Volume) Name

func (v *Volume) Name() string

Name retrieves the volume's name

func (*Volume) Options

func (v *Volume) Options() map[string]string

Options return the volume's options

func (*Volume) Scope

func (v *Volume) Scope() string

Scope retrieves the volume's scope. Libpod does not implement volume scoping, and this is provided solely for Docker compatibility. It returns only "local".

func (*Volume) Size

func (v *Volume) Size() (uint64, error)

Returns the size on disk of volume

func (*Volume) UID

func (v *Volume) UID() (int, error)

UID returns the UID the volume will be created as.

func (*Volume) UsesVolumeDriver

func (v *Volume) UsesVolumeDriver() bool

UsesVolumeDriver determines whether the volume uses a volume driver. Volume drivers are pluggable backends for volumes that will manage the storage and mounting.

func (*Volume) VolumeInUse

func (v *Volume) VolumeInUse() ([]string, error)

VolumeInUse goes through the container dependencies of a volume and checks if the volume is being used by any container.

type VolumeConfig

type VolumeConfig struct {
	// Name of the volume.
	Name string `json:"name"`
	// ID of the volume's lock.
	LockID uint32 `json:"lockID"`
	// Labels for the volume.
	Labels map[string]string `json:"labels"`
	// The volume driver. Empty string or local does not activate a volume
	// driver, all other values will.
	Driver string `json:"volumeDriver"`
	// The location the volume is mounted at.
	MountPoint string `json:"mountPoint"`
	// Time the volume was created.
	CreatedTime time.Time `json:"createdAt,omitempty"`
	// Options to pass to the volume driver. For the local driver, this is
	// a list of mount options. For other drivers, they are passed to the
	// volume driver handling the volume.
	Options map[string]string `json:"volumeOptions,omitempty"`
	// Whether this volume is anonymous (will be removed on container exit)
	IsAnon bool `json:"isAnon"`
	// UID the volume will be created as.
	UID int `json:"uid"`
	// GID the volume will be created as.
	GID int `json:"gid"`
	// Size maximum of the volume.
	Size uint64 `json:"size"`
	// Inodes maximum of the volume.
	Inodes uint64 `json:"inodes"`
}

VolumeConfig holds the volume's immutable configuration.

type VolumeCreateOption

type VolumeCreateOption func(*Volume) error

A VolumeCreateOption is a functional option which alters the Volume created by NewVolume

func WithVolumeDriver

func WithVolumeDriver(driver string) VolumeCreateOption

WithVolumeDriver sets the volume's driver. It is presently not implemented, but will be supported in a future Podman release.

func WithVolumeGID

func WithVolumeGID(gid int) VolumeCreateOption

WithVolumeGID sets the GID that the volume will be created as.

func WithVolumeInodes added in v3.4.0

func WithVolumeInodes(inodes uint64) VolumeCreateOption

WithVolumeInodes sets the maximum inodes of the volume

func WithVolumeLabels

func WithVolumeLabels(labels map[string]string) VolumeCreateOption

WithVolumeLabels sets the labels of the volume.

func WithVolumeName

func WithVolumeName(name string) VolumeCreateOption

WithVolumeName sets the name of the volume.

func WithVolumeNoChown added in v3.2.2

func WithVolumeNoChown() VolumeCreateOption

WithVolumeNoChown prevents the volume from being chowned to the process uid at first use.

func WithVolumeOptions

func WithVolumeOptions(options map[string]string) VolumeCreateOption

WithVolumeOptions sets the options of the volume.

func WithVolumeSize added in v3.4.0

func WithVolumeSize(size uint64) VolumeCreateOption

WithVolumeSize sets the maximum size of the volume

func WithVolumeUID

func WithVolumeUID(uid int) VolumeCreateOption

WithVolumeUID sets the UID that the volume will be created as.

type VolumeFilter

type VolumeFilter func(*Volume) bool

VolumeFilter is a function to determine whether a volume is included in command output. Volumes to be outputted are tested using the function. a true return will include the volume, a false return will exclude it.

type VolumeState

type VolumeState struct {
	// Mountpoint is the location where the volume was mounted.
	// This is only used for volumes using a volume plugin, which will mount
	// at non-standard locations.
	MountPoint string `json:"mountPoint,omitempty"`
	// MountCount is the number of times this volume has been requested to
	// be mounted.
	// It is incremented on mount() and decremented on unmount().
	// On incrementing from 0, the volume will be mounted on the host.
	// On decrementing to 0, the volume will be unmounted on the host.
	MountCount uint `json:"mountCount"`
	// NeedsCopyUp indicates that the next time the volume is mounted into
	// a container, the container will "copy up" the contents of the
	// mountpoint into the volume.
	// This should only be done once. As such, this is set at container
	// create time, then cleared after the copy up is done and never set
	// again.
	NeedsCopyUp bool `json:"notYetMounted,omitempty"`
	// NeedsChown indicates that the next time the volume is mounted into
	// a container, the container will chown the volume to the container process
	// UID/GID.
	NeedsChown bool `json:"notYetChowned,omitempty"`
	// UIDChowned is the UID the volume was chowned to.
	UIDChowned int `json:"uidChowned,omitempty"`
	// GIDChowned is the GID the volume was chowned to.
	GIDChowned int `json:"gidChowned,omitempty"`
}

VolumeState holds the volume's mutable state. Volumes are not guaranteed to have a state. Only volumes using the Local driver that have mount options set will create a state.

type YAMLContainer added in v3.4.2

type YAMLContainer struct {
	v1.Container
	Resources *v1.ResourceRequirements `json:"resources,omitempty"`
}

YAMLContainer represents the same k8s API core Container struct with a small change and that is having Resources as a pointer to k8s API core ResourceRequirements. Because Go doesn't omit empty struct and we want to omit Status in YAML if it's empty. Fixes: GH-11998

type YAMLPod added in v3.4.2

type YAMLPod struct {
	v1.Pod
	Spec   *YAMLPodSpec  `json:"spec,omitempty"`
	Status *v1.PodStatus `json:"status,omitempty"`
}

YAMLPod represents the same k8s API core Pod struct with a small change and that is having Spec as a pointer to YAMLPodSpec and Status as a pointer to k8s API core PodStatus. Because Go doesn't omit empty struct and we want to omit Status in YAML if it's empty. Fixes: GH-11998

func ConvertV1PodToYAMLPod added in v3.4.2

func ConvertV1PodToYAMLPod(pod *v1.Pod) *YAMLPod

ConvertV1PodToYAMLPod takes k8s API core Pod and returns a pointer to YAMLPod

type YAMLPodSpec added in v3.4.2

type YAMLPodSpec struct {
	v1.PodSpec
	Containers []*YAMLContainer `json:"containers"`
}

YAMLPodSpec represents the same k8s API core PodSpec struct with a small change and that is having Containers as a pointer to YAMLContainer. Because Go doesn't omit empty struct and we want to omit Status in YAML if it's empty. Fixes: GH-11998

type YAMLService added in v3.4.2

type YAMLService struct {
	v1.Service
	Status *v1.ServiceStatus `json:"status,omitempty"`
}

YAMLService represents the same k8s API core Service struct with a small change and that is having Status as a pointer to k8s API core ServiceStatus. Because Go doesn't omit empty struct and we want to omit Status in YAML if it's empty. Fixes: GH-11998

func GenerateKubeServiceFromV1Pod

func GenerateKubeServiceFromV1Pod(pod *v1.Pod, servicePorts []v1.ServicePort) (YAMLService, error)

GenerateKubeServiceFromV1Pod creates a v1 service object from a v1 pod object

Directories

Path Synopsis
shm
cni

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL