Improve CSV export, store timestamps with tombstones, will help to repair/sync volumes later

weed/command/export.go

@@ -53,6 +53,7 @@ var (
 	format      = cmdExport.Flag.String("fileNameFormat", defaultFnFormat, "filename formatted with {{.Mime}} {{.Id}} {{.Name}} {{.Ext}}")
 	newer       = cmdExport.Flag.String("newer", "", "export only files newer than this time, default is all files. Must be specified in RFC3339 without timezone, e.g. 2006-01-02T15:04:05")
 	showDeleted = cmdExport.Flag.Bool("deleted", false, "export deleted files. only applies if -o is not specified")
+	volumeInfo  = cmdExport.Flag.Bool("volumeInfo", false, "show volume info")
 
 	tarOutputFile          *tar.Writer
 	tarHeader              tar.Header
@@ -69,13 +70,13 @@ func printNeedle(vid storage.VolumeId, n *storage.Needle, version storage.Versio
 	if version == storage.Version1 {
 		size = n.Size
 	}
-	fmt.Printf("%s;%s;%d;%t;%s;%d;%s;%t\n",
+	fmt.Printf("\"%s\",\"%s\",%d,%t,%s,%s,%s,%t\n",
 		key,
-		n.Name,
+		strings.Replace(string(n.Name), "\"", "\"\"", -1),
 		size,
 		n.IsGzipped(),
 		n.Mime,
-		n.LastModified,
+		n.LastModifiedString(),
 		n.Ttl.String(),
 		deleted,
 	)
@@ -97,6 +98,11 @@ func runExport(cmd *Command, args []string) bool {
 		return false
 	}
 
+	if (*volumeInfo == true) {
+		showVolumeInfo(*export.dir, *export.collection, *export.volumeId)
+		return true
+	}
+
 	if *output != "" {
 		if *output != "-" && !strings.HasSuffix(*output, ".tar") {
 			fmt.Println("the output file", *output, "should be '-' or end with .tar")
@@ -145,7 +151,7 @@ func runExport(cmd *Command, args []string) bool {
 	var version storage.Version
 
 	if tarOutputFile == nil {
-		fmt.Printf("key;name;size;gzip;mime;modified;ttl;deleted\n")
+		fmt.Printf("key,name,size,gzip,mime,modified,ttl,deleted\n")
 	}
 
 	err = storage.ScanVolumeFile(*export.dir, *export.collection, vid,
@@ -163,12 +169,23 @@ func runExport(cmd *Command, args []string) bool {
 						n.LastModified, newerThanUnix)
 					return nil
 				}
-				return walker(vid, n, version)
+
+				if tarOutputFile != nil {
+					return writeFile(vid, n, version)
+				} else {
+					printNeedle(vid, n, version, false)
+					return nil
+				}
 			}
 			if !ok {
-				if *showDeleted && tarOutputFile == nil && n.DataSize > 0 {
+				if *showDeleted && tarOutputFile == nil {
+					if (n.DataSize > 0) {
+						printNeedle(vid, n, version, true)
+					} else {
+						n.Name = []byte("*tombstone")
 						printNeedle(vid, n, version, true)
 					}
+				}
 
 				glog.V(2).Infof("This seems deleted %d size %d", n.Id, n.Size)
 			} else {
@@ -190,9 +207,42 @@ type nameParams struct {
 	Ext  string
 }
 
-func walker(vid storage.VolumeId, n *storage.Needle, version storage.Version) (err error) {
+func showVolumeInfo(dirname string, collection string, volumeId int) {
+	vid := storage.VolumeId(volumeId)
+	var volume *storage.Volume
+	var err error
+	if volume, err = storage.LoadVolume(dirname, collection, vid, storage.NeedleMapInMemory); err != nil {
+		fmt.Println("Failed to load volume: " + err.Error())
+		return
+	}
+
+	var ttl string
+	if (volume.Ttl.Minutes() > 0) {
+		ttl = volume.Ttl.String()
+	} else {
+		ttl = "none"
+	}
+	fmt.Printf("Collection: %s\n"+
+		"VolumeId: %v\n"+
+		"Content-Size: %v\n"+
+		"Content-Count: %v\n"+
+		"Deleted-Size: %v\n"+
+		"Deleted-Count: %v\n"+
+		"Modification-Time: %s\n"+
+		"TTL: %s\n",
+		volume.Collection,
+		volume.Id,
+		volume.ContentSize(),
+		volume.FileCount(),
+		volume.DeletedSize(),
+		volume.DeletedCount(),
+		volume.LastModifiedString(),
+		ttl,
+	)
+}
+
+func writeFile(vid storage.VolumeId, n *storage.Needle, version storage.Version) (err error) {
 	key := storage.NewFileIdFromNeedle(vid, n).String()
-	if tarOutputFile != nil {
 	fileNameTemplateBuffer.Reset()
 	if err = fileNameTemplate.Execute(fileNameTemplateBuffer,
 		nameParams{
@@ -223,8 +273,5 @@ func walker(vid storage.VolumeId, n *storage.Needle, version storage.Version) (e
 		return err
 	}
 	_, err = tarOutputFile.Write(n.Data)
-	} else {
-		printNeedle(vid, n, version, false)
-	}
 	return
 }

weed/server/volume_server_handlers_write.go

@@ -9,6 +9,8 @@ import (
 	"github.com/chrislusf/seaweedfs/weed/operation"
 	"github.com/chrislusf/seaweedfs/weed/storage"
 	"github.com/chrislusf/seaweedfs/weed/topology"
+	"time"
+	"strconv"
 )
 
 func (vs *VolumeServer) PostHandler(w http.ResponseWriter, r *http.Request) {
@@ -87,6 +89,13 @@ func (vs *VolumeServer) DeleteHandler(w http.ResponseWriter, r *http.Request) {
 		count = chunkManifest.Size
 	}
 
+	n.LastModified = uint64(time.Now().Unix())
+	if len(r.FormValue("time")) > 0 {
+		modifiedTime, err := strconv.ParseInt(r.FormValue("time"), 10, 64)
+		if err == nil {
+			n.LastModified = uint64(modifiedTime)
+		}
+	}
 	_, err := topology.ReplicatedDelete(vs.GetMasterNode(), vs.store, volumeId, n, r)
 
 	if err == nil {
@@ -103,6 +112,7 @@ func (vs *VolumeServer) DeleteHandler(w http.ResponseWriter, r *http.Request) {
 func (vs *VolumeServer) batchDeleteHandler(w http.ResponseWriter, r *http.Request) {
 	r.ParseForm()
 	var ret []operation.DeleteResult
+	now := uint64(time.Now().Unix())
 	for _, fid := range r.Form["fid"] {
 		vid, id_cookie, err := operation.ParseFileId(fid)
 		if err != nil {
@@ -144,6 +154,9 @@ func (vs *VolumeServer) batchDeleteHandler(w http.ResponseWriter, r *http.Reques
 			glog.V(0).Infoln("deleting", fid, "with unmaching cookie from ", r.RemoteAddr, "agent", r.UserAgent())
 			return
 		}
+
+		n.LastModified = now
+
 		if size, err := vs.store.Delete(volumeId, n); err != nil {
 			ret = append(ret, operation.DeleteResult{
 				Fid:    fid,

weed/storage/needle.go

@@ -262,3 +262,7 @@ func ParseKeyHash(key_hash_string string) (uint64, uint32, error) {
 	}
 	return key, uint32(hash), nil
 }
+
+func (n *Needle) LastModifiedString() string {
+	return time.Unix(int64(n.LastModified), 0).Format("2006-01-02T15:04:05")
+}

weed/storage/volume.go

@@ -80,6 +80,26 @@ func (v *Volume) ContentSize() uint64 {
 	return v.nm.ContentSize()
 }
 
+func (v *Volume) DeletedSize() uint64 {
+	return v.nm.DeletedSize()
+}
+
+func (v *Volume) FileCount() int {
+	return v.nm.FileCount()
+}
+
+func (v *Volume) DeletedCount() int {
+	return v.nm.DeletedCount()
+}
+
+func (v *Volume) LastModifiedTime() uint64 {
+	return v.lastModifiedTime
+}
+
+func (v *Volume) LastModifiedString() string {
+	return time.Unix(int64(v.lastModifiedTime), 0).Format("2006-01-02T15:04:05")
+}
+
 // volume is expired if modified time + volume ttl < now
 // except when volume is empty
 // or when the volume does not have a ttl

weed/storage/volume_loading.go

@@ -8,6 +8,14 @@ import (
 	"github.com/chrislusf/seaweedfs/weed/glog"
 )
 
+func LoadVolume(dirname string, collection string, id VolumeId, needleMapKind NeedleMapType) (v *Volume, e error) {
+	v = &Volume{dir: dirname, Collection: collection, Id: id}
+	v.SuperBlock = SuperBlock{}
+	v.needleMapKind = needleMapKind
+	e = v.load(true, false, needleMapKind, 0)
+	return
+}
+
 func loadVolumeWithoutIndex(dirname string, collection string, id VolumeId, needleMapKind NeedleMapType) (v *Volume, e error) {
 	v = &Volume{dir: dirname, Collection: collection, Id: id}
 	v.SuperBlock = SuperBlock{}

weed/topology/store_replicate.go

@@ -102,7 +102,7 @@ func ReplicatedDelete(masterNode string, store *storage.Store,
 	if needToReplicate { //send to other replica locations
 		if r.FormValue("type") != "replicate" {
 			if err = distributedOperation(masterNode, store, volumeId, func(location operation.Location) error {
-				return util.Delete("http://"+location.Url+r.URL.Path+"?type=replicate", jwt)
+				return util.Delete(fmt.Sprintf("http://%s%s?type=replicate&time=%d", location.Url, r.URL.Path, n.LastModified), jwt)
 			}); err != nil {
 				ret = 0
 			}