Package backups contains all the stand-alone backup-related functionality for juju state. That functionality is encapsulated by the backups.Backups type. The package also exposes a few key helpers and components.

Backups are not a part of juju state nor of normal state operations. However, they certainly are tightly coupled with state (the very subject of backups). This puts backups in an odd position, particularly with regard to the storage of backup metadata and archives.

As noted above backups are about state but not a part of state. So exposing backup-related methods on State would imply the wrong thing. Thus most of the functionality here is defined at a high level without relation to state. A few low-level parts or helpers are exposed as functions to which you pass a state value. Those are kept to a minimum.

Note that state (and juju as a whole) currently does not have a persistence layer abstraction to facilitate separating different persistence needs and implementations. As a consequence, state's data, whether about how an environment should look or about existing resources within an environment, is dumped essentially straight into State's mongo connection. The code in the state package does not make any distinction between the two (nor does the package clearly distinguish between state-related abstractions and state-related data).

Backups add yet another category, merely taking advantage of State's mongo for storage. In the interest of making the distinction clear, among other reasons, backups uses its own database under state's mongo connection.



View Source
const (
	// FilenamePrefix is the prefix used for backup archive files.
	FilenamePrefix = "juju-backup-"

	// FilenameTemplate is used with time.Time.Format to generate a filename.
	FilenameTemplate = FilenamePrefix + "20060102-150405.tar.gz"
View Source
const UnknownString = "<unknown>"

UnknownString is a marker value for string fields with unknown values.


View Source
var UnknownVersion = version.MustParse("9999.9999.9999")

UnknownVersion is a marker value for version fields with unknown values.


func GetFilesToBackUp

func GetFilesToBackUp(rootDir string, paths *Paths, oldmachine string) ([]string, error)

GetFilesToBackUp returns the paths that should be included in the backup archive.

func NewStorage

func NewStorage(st DB) filestorage.FileStorage

NewStorage returns a new FileStorage to use for storing backup archives (and metadata).

func PrepareMachineForRestore

func PrepareMachineForRestore() error

PrepareMachineForRestore deletes all files from the re-bootstrapped machine that are to be replaced by the backup and recreates those directories that are to contain new files; this is to avoid possible mixup from new/old files that lead to an inconsistent restored state machine.

func StoreArchive

func StoreArchive(stor filestorage.FileStorage, meta *Metadata, file io.Reader) error

StoreArchive sends the backup archive and its metadata to storage. It also sets the metadata's ID and Stored values.


type ArchiveData

type ArchiveData struct {
	// contains filtered or unexported fields


ArchiveData is a wrapper around a the uncompressed data in a backup archive file. It provides access to the content of the archive. While ArchiveData provides useful functionality, it may not be appropriate for large archives. The contents of the archive are kept in-memory, so large archives could be too taxing on the host. In that case consider using ArchiveWorkspace instead.

func NewArchiveData

func NewArchiveData(data []byte) *ArchiveData

NewArchiveData builds a new archive data wrapper for the given uncompressed data.

func NewArchiveDataReader

func NewArchiveDataReader(r io.Reader) (*ArchiveData, error)

NewArchiveReader returns a new archive data wrapper for the data in the provided reader. Note that the entire archive will be read into memory and kept there. So for relatively large archives it will often be more appropriate to use ArchiveWorkspace instead.

func (*ArchiveData) Metadata

func (ad *ArchiveData) Metadata() (*Metadata, error)

Metadata returns the metadata stored in the backup archive. If no metadata is there, errors.NotFound is returned.

func (*ArchiveData) NewBuffer

func (ad *ArchiveData) NewBuffer() *bytes.Buffer

NewBuffer wraps the archive data in a Buffer.

func (*ArchiveData) Version

func (ad *ArchiveData) Version() (*version.Number, error)

Version returns the juju version under which the backup archive was created. If no version is found in the archive, it must come from before backup archives included the version. In that case we return version 1.20.

type ArchivePaths

type ArchivePaths struct {
	// ContentDir is the path to the directory within the archive
	// containing all the contents. It is the only file or directory at
	// the top-level of the archive and everything else in the archive
	// is contained in the content directory.
	ContentDir string

	// FilesBundle is the path to the tar file inside the archive
	// containing all the state-related files (with the exception of the
	// DB dump files) gathered in by the backup machinery.
	FilesBundle string

	// DBDumpDir is the path to the directory within the archive
	// contents that contains all the files dumped from the juju state
	// database.
	DBDumpDir string

	// MetadataFile is the path to the metadata file.
	MetadataFile string

ArchivePaths holds the paths to the files and directories in a backup archive.

func NewCanonicalArchivePaths

func NewCanonicalArchivePaths() ArchivePaths

NewCanonicalArchivePaths composes a new ArchivePaths with default values set. These values are relative (un-rooted) and the canonical slash ("/") is the path separator. Thus the paths are suitable for resolving the paths in a backup archive file (which is a tar file).

func NewNonCanonicalArchivePaths

func NewNonCanonicalArchivePaths(rootDir string) ArchivePaths

NonCanonicalArchivePaths builds a new ArchivePaths using default values, rooted at the provided rootDir. The path separator used is platform-dependent. The resulting paths are suitable for locating backup archive contents in a directory into which an archive has been unpacked.

type ArchiveWorkspace

type ArchiveWorkspace struct {
	RootDir string

ArchiveWorkspace is a wrapper around backup archive info that has a concrete root directory and an archive unpacked in it.

func NewArchiveWorkspaceReader

func NewArchiveWorkspaceReader(archive io.Reader) (*ArchiveWorkspace, error)

NewArchiveWorkspaceReader returns a new archive workspace with a new workspace dir populated from the archive. Note that this involves unpacking the entire archive into a directory under the host's "temporary" directory. For relatively large archives this could have adverse effects on hosts with little disk space.

func (*ArchiveWorkspace) Close

func (ws *ArchiveWorkspace) Close() error

Close cleans up the workspace dir.

func (*ArchiveWorkspace) Metadata

func (ws *ArchiveWorkspace) Metadata() (*Metadata, error)

Metadata returns the metadata derived from the JSON file in the archive.

func (*ArchiveWorkspace) OpenBundledFile

func (ws *ArchiveWorkspace) OpenBundledFile(filename string) (io.Reader, error)

OpenBundledFile returns an open ReadCloser for the corresponding file in the archived files bundle.

func (*ArchiveWorkspace) UnpackFilesBundle

func (ws *ArchiveWorkspace) UnpackFilesBundle(targetRoot string) error

UnpackFilesBundle unpacks the archived files bundle into the targeted dir.

type Backups

type Backups interface {
	// Create creates and stores a new juju backup archive. It updates
	// the provided metadata.
	Create(meta *Metadata, paths *Paths, dbInfo *DBInfo) error

	// Add stores the backup archive and returns its new ID.
	Add(archive io.Reader, meta *Metadata) (string, error)

	// Get returns the metadata and archive file associated with the ID.
	Get(id string) (*Metadata, io.ReadCloser, error)

	// List returns the metadata for all stored backups.
	List() ([]*Metadata, error)

	// Remove deletes the backup from storage.
	Remove(id string) error

	// Restore updates juju's state to the contents of the backup archive.
	Restore(backupId string, args params.RestoreArgs) error

Backups is an abstraction around all juju backup-related functionality.

func NewBackups

func NewBackups(stor filestorage.FileStorage) Backups

NewBackups creates a new Backups value using the FileStorage provided.

type DB

type DB interface {
	// MongoSession returns the underlying mongodb session.
	MongoSession() *mgo.Session

	// EnvironTag is the concrete environ tag for this database.
	EnvironTag() names.EnvironTag

DB represents the set of methods required to perform a backup. It exists to break the strict dependency between state and this package, and those that depend on this package.

type DBDumper

type DBDumper interface {
	// Dump something to dumpDir.
	Dump(dumpDir string) error

DBDumper is any type that dumps something to a dump dir.

func NewDBDumper

func NewDBDumper(info *DBInfo) (DBDumper, error)

NewDBDumper returns a new value with a Dump method for dumping the juju state database.

type DBInfo

type DBInfo struct {
	// Address is the DB system's host address.
	Address string
	// Username is used when connecting to the DB system.
	Username string
	// Password is used when connecting to the DB system.
	Password string
	// Targets is a list of databases to dump.
	Targets set.Strings

DBInfo wraps all the DB-specific information backups needs to dump the database. This includes a simplification of the information in authentication.MongoInfo.

func NewDBInfo

func NewDBInfo(mgoInfo *mongo.MongoInfo, session DBSession) (*DBInfo, error)

NewDBInfo returns the information needed by backups to dump the database.

type DBSession

type DBSession interface {
	DatabaseNames() ([]string, error)

type Metadata

type Metadata struct {

	// Started records when the backup was started.
	Started time.Time
	// Finished records when the backup was complete.
	Finished *time.Time
	// Origin identifies where the backup was created.
	Origin Origin
	// Notes is an optional user-supplied annotation.
	Notes string

Metadata contains the metadata for a single state backup archive.

func BuildMetadata

func BuildMetadata(file *os.File) (*Metadata, error)

BuildMetadata generates the metadata for a backup archive file.

func NewMetadata

func NewMetadata() *Metadata

NewMetadata returns a new Metadata for a state backup archive. Only the start time and the version are set.

func NewMetadataJSONReader

func NewMetadataJSONReader(in io.Reader) (*Metadata, error)

NewMetadataJSONReader extracts a new metadata from the JSON file.

func NewMetadataState

func NewMetadataState(db DB, machine string) (*Metadata, error)

NewMetadataState composes a new backup metadata with its origin values set. The environment UUID comes from state. The hostname is retrieved from the OS.

func (*Metadata) AsJSONBuffer

func (m *Metadata) AsJSONBuffer() (io.Reader, error)

AsJSONBuffer returns a bytes.Buffer containing the JSON-ified metadata.

func (*Metadata) MarkComplete

func (m *Metadata) MarkComplete(size int64, checksum string) error

MarkComplete populates the remaining metadata values. The default checksum format is used.

type Origin

type Origin struct {
	Environment string
	Machine     string
	Hostname    string
	Version     version.Number

Origin identifies where a backup archive came from. While it is more about where and Metadata about what and when, that distinction does not merit special consideration. Instead, Origin exists separately from Metadata due to its use as an argument when requesting the creation of a new backup.

func UnknownOrigin

func UnknownOrigin() Origin

UnknownOrigin returns a new backups origin with unknown values.

type Paths

type Paths struct {
	DataDir string
	LogsDir string

Paths holds the paths that backups needs.


Path Synopsis