heal: Remove empty directories (#11354)

Since the introduction of __XLDIR__, an empty directory does not have a
meaning anymore in erasure mode. Make healing removes it wherever it
finds it.

cmd/erasure-object.go

@@ -786,6 +786,31 @@ func (er erasureObjects) deleteObjectVersion(ctx context.Context, bucket, object
 	return reduceWriteQuorumErrs(ctx, g.Wait(), objectOpIgnoredErrs, writeQuorum)
 }
 
+// deleteEmptyDir knows only how to remove an empty directory (not the empty object with a
+// trailing slash), this is called for the healing code to remove such directories.
+func (er erasureObjects) deleteEmptyDir(ctx context.Context, bucket, object string) error {
+	defer ObjectPathUpdated(pathJoin(bucket, object))
+
+	if bucket == minioMetaTmpBucket {
+		return nil
+	}
+
+	disks := er.getDisks()
+	g := errgroup.WithNErrs(len(disks))
+	for index := range disks {
+		index := index
+		g.Go(func() error {
+			if disks[index] == nil {
+				return errDiskNotFound
+			}
+			return disks[index].Delete(ctx, bucket, object, false)
+		}, index)
+	}
+
+	// return errors if any during deletion
+	return reduceWriteQuorumErrs(ctx, g.Wait(), objectOpIgnoredErrs, len(disks)/2+1)
+}
+
 // deleteObject - wrapper for delete object, deletes an object from
 // all the disks in parallel, including `xl.meta` associated with the
 // object.

cmd/erasure-server-pool.go

@@ -1359,6 +1359,14 @@ func (z *erasureServerPools) HealObjects(ctx context.Context, bucket, prefix str
 					break
 				}
 
+				// Remove empty directories if found - they have no meaning.
+				// Can be left over from highly concurrent put/remove.
+				if quorumCount > set.setDriveCount/2 && entry.IsEmptyDir {
+					if !opts.DryRun && opts.Remove {
+						set.deleteEmptyDir(ctx, bucket, entry.Name)
+					}
+				}
+
 				// Indicate that first attempt was a success and subsequent loop
 				// knows that its not our first attempt at 'prefix'
 				err = nil

cmd/storage-datatypes.go

@@ -76,6 +76,8 @@ type FileInfoVersions struct {
 	// Name of the file.
 	Name string
 
+	IsEmptyDir bool
+
 	// Represents the latest mod time of the
 	// latest version.
 	LatestModTime time.Time

cmd/storage-datatypes_gen.go

@@ -737,6 +737,12 @@ func (z *FileInfoVersions) DecodeMsg(dc *msgp.Reader) (err error) {
 				err = msgp.WrapError(err, "Name")
 				return
 			}
+		case "IsEmptyDir":
+			z.IsEmptyDir, err = dc.ReadBool()
+			if err != nil {
+				err = msgp.WrapError(err, "IsEmptyDir")
+				return
+			}
 		case "LatestModTime":
 			z.LatestModTime, err = dc.ReadTime()
 			if err != nil {
@@ -775,9 +781,9 @@ func (z *FileInfoVersions) DecodeMsg(dc *msgp.Reader) (err error) {
 
 // EncodeMsg implements msgp.Encodable
 func (z *FileInfoVersions) EncodeMsg(en *msgp.Writer) (err error) {
-	// map header, size 4
+	// map header, size 5
 	// write "Volume"
-	err = en.Append(0x84, 0xa6, 0x56, 0x6f, 0x6c, 0x75, 0x6d, 0x65)
+	err = en.Append(0x85, 0xa6, 0x56, 0x6f, 0x6c, 0x75, 0x6d, 0x65)
 	if err != nil {
 		return
 	}
@@ -796,6 +802,16 @@ func (z *FileInfoVersions) EncodeMsg(en *msgp.Writer) (err error) {
 		err = msgp.WrapError(err, "Name")
 		return
 	}
+	// write "IsEmptyDir"
+	err = en.Append(0xaa, 0x49, 0x73, 0x45, 0x6d, 0x70, 0x74, 0x79, 0x44, 0x69, 0x72)
+	if err != nil {
+		return
+	}
+	err = en.WriteBool(z.IsEmptyDir)
+	if err != nil {
+		err = msgp.WrapError(err, "IsEmptyDir")
+		return
+	}
 	// write "LatestModTime"
 	err = en.Append(0xad, 0x4c, 0x61, 0x74, 0x65, 0x73, 0x74, 0x4d, 0x6f, 0x64, 0x54, 0x69, 0x6d, 0x65)
 	if err != nil {
@@ -829,13 +845,16 @@ func (z *FileInfoVersions) EncodeMsg(en *msgp.Writer) (err error) {
 // MarshalMsg implements msgp.Marshaler
 func (z *FileInfoVersions) MarshalMsg(b []byte) (o []byte, err error) {
 	o = msgp.Require(b, z.Msgsize())
-	// map header, size 4
+	// map header, size 5
 	// string "Volume"
-	o = append(o, 0x84, 0xa6, 0x56, 0x6f, 0x6c, 0x75, 0x6d, 0x65)
+	o = append(o, 0x85, 0xa6, 0x56, 0x6f, 0x6c, 0x75, 0x6d, 0x65)
 	o = msgp.AppendString(o, z.Volume)
 	// string "Name"
 	o = append(o, 0xa4, 0x4e, 0x61, 0x6d, 0x65)
 	o = msgp.AppendString(o, z.Name)
+	// string "IsEmptyDir"
+	o = append(o, 0xaa, 0x49, 0x73, 0x45, 0x6d, 0x70, 0x74, 0x79, 0x44, 0x69, 0x72)
+	o = msgp.AppendBool(o, z.IsEmptyDir)
 	// string "LatestModTime"
 	o = append(o, 0xad, 0x4c, 0x61, 0x74, 0x65, 0x73, 0x74, 0x4d, 0x6f, 0x64, 0x54, 0x69, 0x6d, 0x65)
 	o = msgp.AppendTime(o, z.LatestModTime)
@@ -882,6 +901,12 @@ func (z *FileInfoVersions) UnmarshalMsg(bts []byte) (o []byte, err error) {
 				err = msgp.WrapError(err, "Name")
 				return
 			}
+		case "IsEmptyDir":
+			z.IsEmptyDir, bts, err = msgp.ReadBoolBytes(bts)
+			if err != nil {
+				err = msgp.WrapError(err, "IsEmptyDir")
+				return
+			}
 		case "LatestModTime":
 			z.LatestModTime, bts, err = msgp.ReadTimeBytes(bts)
 			if err != nil {
@@ -921,7 +946,7 @@ func (z *FileInfoVersions) UnmarshalMsg(bts []byte) (o []byte, err error) {
 
 // Msgsize returns an upper bound estimate of the number of bytes occupied by the serialized message
 func (z *FileInfoVersions) Msgsize() (s int) {
-	s = 1 + 7 + msgp.StringPrefixSize + len(z.Volume) + 5 + msgp.StringPrefixSize + len(z.Name) + 14 + msgp.TimeSize + 9 + msgp.ArrayHeaderSize
+	s = 1 + 7 + msgp.StringPrefixSize + len(z.Volume) + 5 + msgp.StringPrefixSize + len(z.Name) + 11 + msgp.BoolSize + 14 + msgp.TimeSize + 9 + msgp.ArrayHeaderSize
 	for za0001 := range z.Versions {
 		s += z.Versions[za0001].Msgsize()
 	}

cmd/tree-walk.go

@@ -25,6 +25,7 @@ import (
 // TreeWalkResult - Tree walk result carries results of tree walking.
 type TreeWalkResult struct {
 	entry      string
+	isEmptyDir bool
 	end        bool
 }
 
@@ -254,7 +255,7 @@ func doTreeWalk(ctx context.Context, bucket, prefixDir, entryPrefixMatch, marker
 		select {
 		case <-endWalkCh:
 			return false, errWalkAbort
-		case resultCh <- TreeWalkResult{entry: pathJoin(prefixDir, entry), end: isEOF}:
+		case resultCh <- TreeWalkResult{entry: pathJoin(prefixDir, entry), isEmptyDir: leafDir, end: isEOF}:
 		}
 	}
 

cmd/xl-storage.go

@@ -813,6 +813,7 @@ func (s *xlStorage) WalkVersions(ctx context.Context, volume, dirPath, marker st
 				fiv = FileInfoVersions{
 					Volume:     volume,
 					Name:       walkResult.entry,
+					IsEmptyDir: walkResult.isEmptyDir,
 					Versions: []FileInfo{
 						{
 							Volume: volume,