lsbackup command to list backups.

dgraph/cmd/root_ee.go

@@ -21,6 +21,7 @@ func init() {
 	// subcommands already has the default subcommands, we append to EE ones to that.
 	subcommands = append(subcommands,
 		&backup.Restore,
+		&backup.LsBackup,
 		&acl.CmdAcl,
 	)
 }

ee/backup/backup.go

@@ -84,6 +84,14 @@ type Manifest struct {
 	Groups  []uint32 `json:"groups"`
 }
 
+// ManifestStatus combines a manifest along with other information about it
+// that should not be inside the Manifest struct since it should not be
+// recorded in manifest files.
+type ManifestStatus struct {
+	Manifest *Manifest
+	FileName string
+}
+
 // GoString implements the GoStringer interface for Manifest.
 func (m *Manifest) GoString() string {
 	return fmt.Sprintf(`Manifest{Version: %d, ReadTs: %d, Groups: %v}`,

ee/backup/backup_test.go

@@ -214,7 +214,7 @@ func RestoreFull(t *testing.T, c *dgo.Dgraph) {
 
 	// restore this backup dir (3 files total)
 	t.Logf("--- Restoring from: %q", dirs[0])
-	_, err := runRestore("./data/restore", dirs[0])
+	_, err := runRestoreInternal("./data/restore", dirs[0])
 	require.NoError(t, err)
 
 	// just check p1 which should have the 'movie' predicate (moved during setup)
@@ -277,7 +277,7 @@ func RestoreIncr1(t *testing.T, c *dgo.Dgraph) {
 
 	// restore this backup dir (3 files total)
 	t.Logf("--- Restoring from: %q", dirs[1])
-	_, err := runRestore("./data/restore", dirs[1])
+	_, err := runRestoreInternal("./data/restore", dirs[1])
 	require.NoError(t, err)
 
 	// just check p1 which should have the 'movie' predicate (moved during setup)
@@ -335,7 +335,7 @@ func RestoreIncr2(t *testing.T, c *dgo.Dgraph) {
 
 	// restore this backup dir (3 files total)
 	t.Logf("--- Restoring from: %q", dirs[2])
-	_, err := runRestore("./data/restore", dirs[2])
+	_, err := runRestoreInternal("./data/restore", dirs[2])
 	require.NoError(t, err)
 
 	// just check p1 which should have the 'movie' predicate (moved during setup)

ee/backup/file_handler.go

@@ -137,7 +137,7 @@ func (h *fileHandler) Load(uri *url.URL, fn loadFn) (uint64, error) {
 			continue
 		}
 
-		// Load the backup for each group in manifest.
+		// Check the files for each group in the manifest exist.
 		path := filepath.Dir(manifest)
 		for _, groupId := range m.Groups {
 			file := filepath.Join(path, fmt.Sprintf(backupNameFmt, m.ReadTs, groupId))
@@ -155,6 +155,42 @@ func (h *fileHandler) Load(uri *url.URL, fn loadFn) (uint64, error) {
 	return version, nil
 }
 
+// ListManifests loads the manifests in the locations and returns them.
+func (h *fileHandler) ListManifests(uri *url.URL) ([]*ManifestStatus, error) {
+	if !pathExist(uri.Path) {
+		return nil, x.Errorf("The path %q does not exist or it is inaccessible.", uri.Path)
+	}
+
+	// Get a list of all the manifest files at the location.
+	suffix := filepath.Join(string(filepath.Separator), backupManifest)
+	manifests := x.WalkPathFunc(uri.Path, func(path string, isdir bool) bool {
+		return !isdir && strings.HasSuffix(path, suffix)
+	})
+	if len(manifests) == 0 {
+		return nil, x.Errorf("No manifests found at path: %s", uri.Path)
+	}
+	sort.Strings(manifests)
+	if glog.V(3) {
+		fmt.Printf("Found backup manifest(s): %v\n", manifests)
+	}
+
+	// Process each manifest, first check that they are valid and then confirm the
+	// backup files for each group exist. Each group in manifest must have a backup file.
+	var listedManifests []*ManifestStatus
+	for _, manifest := range manifests {
+		var m Manifest
+		var ms ManifestStatus
+
+		if err := h.readManifest(manifest, &m); err != nil {
+			return nil, x.Wrapf(err, "While reading %q", manifest)
+		}
+		ms.Manifest = &m
+		ms.FileName = manifest
+		listedManifests = append(listedManifests, &ms)
+	}
+	return listedManifests, nil
+}
+
 func (h *fileHandler) Close() error {
 	if h.fp == nil {
 		return nil

ee/backup/handler.go

@@ -72,6 +72,12 @@ type handler interface {
 	// The loadFn receives the files as they are processed by a handler, to do the actual
 	// load to DB.
 	Load(*url.URL, loadFn) (uint64, error)
+
+	// ListManifests will scan the provided URI for backup manifests. This operation
+	// should be read-only.
+	//
+	// The URL object is parsed as described in `newHandler`.
+	ListManifests(*url.URL) ([]*ManifestStatus, error)
 }
 
 // getHandler returns a handler for the URI scheme.
@@ -148,3 +154,18 @@ func Load(l string, fn loadFn) (version uint64, err error) {
 
 	return h.Load(uri, fn)
 }
+
+// ListManifests scans location l for backup files and returns the list of manifests.
+func ListManifests(l string) ([]*ManifestStatus, error) {
+	uri, err := url.Parse(l)
+	if err != nil {
+		return nil, err
+	}
+
+	h := getHandler(uri.Scheme)
+	if h == nil {
+		return nil, x.Errorf("Unsupported URI: %v", uri)
+	}
+
+	return h.ListManifests(uri)
+}

ee/backup/run.go

@@ -31,12 +31,18 @@ import (
 )
 
 var Restore x.SubCommand
+var LsBackup x.SubCommand
 
 var opt struct {
 	location, pdir, zero string
 }
 
 func init() {
+	initRestore()
+	initBackupLs()
+}
+
+func initRestore() {
 	Restore.Cmd = &cobra.Command{
 		Use:   "restore",
 		Short: "Run Dgraph (EE) Restore backup",
@@ -85,7 +91,7 @@ $ dgraph restore -p . -l /var/backups/dgraph -z localhost:5080
 		Args: cobra.NoArgs,
 		Run: func(cmd *cobra.Command, args []string) {
 			defer x.StartProfile(Restore.Conf).Stop()
-			if err := run(); err != nil {
+			if err := runRestoreCmd(); err != nil {
 				fmt.Fprintln(os.Stderr, err)
 				os.Exit(1)
 			}
@@ -102,7 +108,56 @@ $ dgraph restore -p . -l /var/backups/dgraph -z localhost:5080
 	_ = Restore.Cmd.MarkFlagRequired("location")
 }
 
-func run() error {
+func initBackupLs() {
+	LsBackup.Cmd = &cobra.Command{
+		Use:   "lsbackup",
+		Short: "List info on backups in given location",
+		Long: `
+lsbackup looks at a location where backups are stored and prints information about them.
+
+Backups are originated from HTTP at /admin/backup, then can be restored using CLI restore
+command. Restore is intended to be used with new Dgraph clusters in offline state.
+
+The --location flag indicates a source URI with Dgraph backup objects. This URI supports all
+the schemes used for backup.
+
+Source URI formats:
+  [scheme]://[host]/[path]?[args]
+  [scheme]:///[path]?[args]
+  /[path]?[args] (only for local or NFS)
+
+Source URI parts:
+  scheme - service handler, one of: "s3", "minio", "file"
+    host - remote address. ex: "dgraph.s3.amazonaws.com"
+    path - directory, bucket or container at target. ex: "/dgraph/backups/"
+    args - specific arguments that are ok to appear in logs.
+
+Dgraph backup creates a unique backup object for each node group, and restore will create
+a posting directory 'p' matching the backup group ID. Such that a backup file
+named '.../r32-g2.backup' will be loaded to posting dir 'p2'.
+
+Usage examples:
+
+# Run using location in S3:
+$ dgraph lsbackup -l s3://s3.us-west-2.amazonaws.com/srfrog/dgraph
+		`,
+		Args: cobra.NoArgs,
+		Run: func(cmd *cobra.Command, args []string) {
+			defer x.StartProfile(Restore.Conf).Stop()
+			if err := runLsbackupCmd(); err != nil {
+				fmt.Fprintln(os.Stderr, err)
+				os.Exit(1)
+			}
+		},
+	}
+
+	flag := LsBackup.Cmd.Flags()
+	flag.StringVarP(&opt.location, "location", "l", "",
+		"Sets the source location URI (required).")
+	_ = LsBackup.Cmd.MarkFlagRequired("location")
+}
+
+func runRestoreCmd() error {
 	var (
 		start time.Time
 		zc    pb.ZeroClient
@@ -183,3 +238,22 @@ func runRestore(pdir, location string) (uint64, error) {
 		return db.Load(r)
 	})
 }
+
+func runLsbackupCmd() error {
+	fmt.Println("Listing backups from:", opt.location)
+	manifests, err := ListManifests(opt.location)
+	if err != nil {
+		return x.Errorf("Error while listing manifests: %v", err.Error())
+	}
+
+	fmt.Printf("Name\tVersion\tReadTs\tGroups\tValid\n")
+	for _, manifest := range manifests {
+		fmt.Printf("%v\t%v\t%v\t%v\n",
+			manifest.FileName,
+			manifest.Manifest.Version,
+			manifest.Manifest.ReadTs,
+			manifest.Manifest.Groups)
+	}
+
+	return nil
+}

ee/backup/s3_handler.go

@@ -275,6 +275,49 @@ func (h *s3Handler) Load(uri *url.URL, fn loadFn) (uint64, error) {
 	return version, nil
 }
 
+// ListManifests loads the manifests in the locations and returns them.
+func (h *s3Handler) ListManifests(uri *url.URL) ([]*ManifestStatus, error) {
+	mc, err := h.setup(uri)
+	if err != nil {
+		return nil, err
+	}
+
+	var manifests []string
+	var listedManifests []*ManifestStatus
+
+	doneCh := make(chan struct{})
+	defer close(doneCh)
+
+	suffix := "/" + backupManifest
+	for object := range mc.ListObjects(h.bucketName, h.objectPrefix, true, doneCh) {
+		if strings.HasSuffix(object.Key, suffix) {
+			manifests = append(manifests, object.Key)
+		}
+	}
+	if len(manifests) == 0 {
+		return nil, x.Errorf("No manifests found at: %s", uri.String())
+	}
+	sort.Strings(manifests)
+	if glog.V(3) {
+		fmt.Printf("Found backup manifest(s) %s: %v\n", uri.Scheme, manifests)
+	}
+
+	// Process each manifest, first check that they are valid and then confirm the
+	// backup files for each group exist. Each group in manifest must have a backup file.
+	for _, manifest := range manifests {
+		var m Manifest
+		var ms ManifestStatus
+
+		if err := h.readManifest(mc, manifest, &m); err != nil {
+			return nil, x.Wrapf(err, "While reading %q", manifest)
+		}
+		ms.Manifest = &m
+		ms.FileName = manifest
+		listedManifests = append(listedManifests, &ms)
+	}
+	return listedManifests, nil
+}
+
 // upload will block until it's done or an error occurs.
 func (h *s3Handler) upload(mc *minio.Client, object string) error {
 	start := time.Now()