volume: add option to limit compaction speed

weed/command/backup.go

@@ -91,7 +91,7 @@ func runBackup(cmd *Command, args []string) bool {
 	}
 
 	if v.SuperBlock.CompactionRevision < uint16(stats.CompactRevision) {
-		if err = v.Compact(0); err != nil {
+		if err = v.Compact(0, 0); err != nil {
 			fmt.Printf("Compact Volume before synchronizing %v\n", err)
 			return true
 		}

weed/command/compact.go

@@ -43,7 +43,7 @@ func runCompact(cmd *Command, args []string) bool {
 		glog.Fatalf("Load Volume [ERROR] %s\n", err)
 	}
 	if *compactMethod == 0 {
-		if err = v.Compact(preallocate); err != nil {
+		if err = v.Compact(preallocate, 0); err != nil {
 			glog.Fatalf("Compact Volume [ERROR] %s\n", err)
 		}
 	} else {

weed/command/server.go

@@ -94,6 +94,7 @@ func init() {
 	serverOptions.v.indexType = cmdServer.Flag.String("volume.index", "memory", "Choose [memory|leveldb|leveldbMedium|leveldbLarge] mode for memory~performance balance.")
 	serverOptions.v.fixJpgOrientation = cmdServer.Flag.Bool("volume.images.fix.orientation", false, "Adjust jpg orientation when uploading.")
 	serverOptions.v.readRedirect = cmdServer.Flag.Bool("volume.read.redirect", true, "Redirect moved or non-local volumes.")
+	serverOptions.v.compactionMBPerSecond = cmdServer.Flag.Int("volume.compactionMBps", 0, "limit compaction speed in mega bytes per second")
 	serverOptions.v.publicUrl = cmdServer.Flag.String("volume.publicUrl", "", "publicly accessible address")
 
 	s3Options.filerBucketsPath = cmdServer.Flag.String("s3.filer.dir.buckets", "/buckets", "folder on filer to store all buckets")

weed/command/volume.go

@@ -43,6 +43,7 @@ type VolumeServerOptions struct {
 	readRedirect          *bool
 	cpuProfile            *string
 	memProfile            *string
+	compactionMBPerSecond *int
 }
 
 func init() {
@@ -63,6 +64,7 @@ func init() {
 	v.readRedirect = cmdVolume.Flag.Bool("read.redirect", true, "Redirect moved or non-local volumes.")
 	v.cpuProfile = cmdVolume.Flag.String("cpuprofile", "", "cpu profile output file")
 	v.memProfile = cmdVolume.Flag.String("memprofile", "", "memory profile output file")
+	v.compactionMBPerSecond = cmdVolume.Flag.Int("compactionMBps", 0, "limit compaction speed in mega bytes per second")
 }
 
 var cmdVolume = &Command{
@@ -157,6 +159,7 @@ func (v VolumeServerOptions) startVolumeServer(volumeFolders, maxVolumeCounts, v
 		strings.Split(masters, ","), *v.pulseSeconds, *v.dataCenter, *v.rack,
 		v.whiteList,
 		*v.fixJpgOrientation, *v.readRedirect,
+		*v.compactionMBPerSecond,
 	)
 
 	listeningAddress := *v.bindIp + ":" + strconv.Itoa(*v.port)

weed/server/volume_grpc_vacuum.go

@@ -28,7 +28,7 @@ func (vs *VolumeServer) VacuumVolumeCompact(ctx context.Context, req *volume_ser
 
 	resp := &volume_server_pb.VacuumVolumeCompactResponse{}
 
-	err := vs.store.CompactVolume(needle.VolumeId(req.VolumeId), req.Preallocate)
+	err := vs.store.CompactVolume(needle.VolumeId(req.VolumeId), req.Preallocate, vs.compactionBytePerSecond)
 
 	if err != nil {
 		glog.Errorf("compact volume %d: %v", req.VolumeId, err)

weed/server/volume_server.go

@@ -23,6 +23,7 @@ type VolumeServer struct {
 	needleMapKind           storage.NeedleMapType
 	FixJpgOrientation       bool
 	ReadRedirect            bool
+	compactionBytePerSecond int64
 }
 
 func NewVolumeServer(adminMux, publicMux *http.ServeMux, ip string,
@@ -33,7 +34,9 @@ func NewVolumeServer(adminMux, publicMux *http.ServeMux, ip string,
 	dataCenter string, rack string,
 	whiteList []string,
 	fixJpgOrientation bool,
-	readRedirect bool) *VolumeServer {
+	readRedirect bool,
+	compactionMBPerSecond int,
+) *VolumeServer {
 
 	v := viper.GetViper()
 	signingKey := v.GetString("jwt.signing.key")
@@ -47,6 +50,7 @@ func NewVolumeServer(adminMux, publicMux *http.ServeMux, ip string,
 		FixJpgOrientation:       fixJpgOrientation,
 		ReadRedirect:            readRedirect,
 		grpcDialOption:          security.LoadClientTLS(viper.Sub("grpc"), "volume"),
+		compactionBytePerSecond: int64(compactionMBPerSecond) * 1024 * 1024,
 	}
 	vs.MasterNodes = masterNodes
 	vs.store = storage.NewStore(port, ip, publicUrl, folders, maxCounts, vs.needleMapKind)

weed/storage/store_vacuum.go

@@ -14,9 +14,9 @@ func (s *Store) CheckCompactVolume(volumeId needle.VolumeId) (float64, error) {
 	}
 	return 0, fmt.Errorf("volume id %d is not found during check compact", volumeId)
 }
-func (s *Store) CompactVolume(vid needle.VolumeId, preallocate int64) error {
+func (s *Store) CompactVolume(vid needle.VolumeId, preallocate int64, compactionBytePerSecond int64) error {
 	if v := s.findVolume(vid); v != nil {
-		return v.Compact(preallocate)
+		return v.Compact(preallocate, compactionBytePerSecond)
 	}
 	return fmt.Errorf("volume id %d is not found during compact", vid)
 }

weed/storage/volume_vacuum.go

@@ -18,7 +18,7 @@ func (v *Volume) garbageLevel() float64 {
 	return float64(v.nm.DeletedSize()) / float64(v.ContentSize())
 }
 
-func (v *Volume) Compact(preallocate int64) error {
+func (v *Volume) Compact(preallocate int64, compactionBytePerSecond int64) error {
 	glog.V(3).Infof("Compacting volume %d ...", v.Id)
 	//no need to lock for copy on write
 	//v.accessLock.Lock()
@@ -29,7 +29,7 @@ func (v *Volume) Compact(preallocate int64) error {
 	v.lastCompactIndexOffset = v.nm.IndexFileSize()
 	v.lastCompactRevision = v.SuperBlock.CompactionRevision
 	glog.V(3).Infof("creating copies for volume %d ,last offset %d...", v.Id, v.lastCompactIndexOffset)
-	return v.copyDataAndGenerateIndexFile(filePath+".cpd", filePath+".cpx", preallocate)
+	return v.copyDataAndGenerateIndexFile(filePath+".cpd", filePath+".cpx", preallocate, compactionBytePerSecond)
 }
 
 func (v *Volume) Compact2() error {
@@ -242,6 +242,9 @@ type VolumeFileScanner4Vacuum struct {
 	nm                      *NeedleMap
 	newOffset               int64
 	now                     uint64
+	compactionBytePerSecond int64
+	lastSizeCounter         int64
+	lastSizeCheckTime       time.Time
 }
 
 func (scanner *VolumeFileScanner4Vacuum) VisitSuperBlock(superBlock SuperBlock) error {
@@ -269,13 +272,32 @@ func (scanner *VolumeFileScanner4Vacuum) VisitNeedle(n *needle.Needle, offset in
 		if _, _, _, err := n.Append(scanner.dst, scanner.v.Version()); err != nil {
 			return fmt.Errorf("cannot append needle: %s", err)
 		}
-		scanner.newOffset += n.DiskSize(scanner.version)
+		delta := n.DiskSize(scanner.version)
+		scanner.newOffset += delta
+		scanner.maybeSlowdown(delta)
 		glog.V(4).Infoln("saving key", n.Id, "volume offset", offset, "=>", scanner.newOffset, "data_size", n.Size)
 	}
 	return nil
 }
+func (scanner *VolumeFileScanner4Vacuum) maybeSlowdown(delta int64) {
+	if scanner.compactionBytePerSecond > 0 {
+		scanner.lastSizeCounter += delta
+		now := time.Now()
+		elapsedDuration := now.Sub(scanner.lastSizeCheckTime)
+		if elapsedDuration > 100*time.Millisecond {
+			overLimitBytes := scanner.lastSizeCounter - scanner.compactionBytePerSecond/10
+			if overLimitBytes > 0 {
+				overRatio := float64(overLimitBytes) / float64(scanner.compactionBytePerSecond)
+				sleepTime := time.Duration(overRatio*1000) * time.Millisecond
+				// glog.V(0).Infof("currently %d bytes, limit to %d bytes, over by %d bytes, sleeping %v over %.4f", scanner.lastSizeCounter, scanner.compactionBytePerSecond/10, overLimitBytes, sleepTime, overRatio)
+				time.Sleep(sleepTime)
+			}
+			scanner.lastSizeCounter, scanner.lastSizeCheckTime = 0, time.Now()
+		}
+	}
+}
 
-func (v *Volume) copyDataAndGenerateIndexFile(dstName, idxName string, preallocate int64) (err error) {
+func (v *Volume) copyDataAndGenerateIndexFile(dstName, idxName string, preallocate int64, compactionBytePerSecond int64) (err error) {
 	var (
 		dst, idx *os.File
 	)
@@ -294,6 +316,8 @@ func (v *Volume) copyDataAndGenerateIndexFile(dstName, idxName string, prealloca
 		now:                     uint64(time.Now().Unix()),
 		nm:                      NewBtreeNeedleMap(idx),
 		dst:                     dst,
+		compactionBytePerSecond: compactionBytePerSecond,
+		lastSizeCheckTime:       time.Now(),
 	}
 	err = ScanVolumeFile(v.dir, v.Collection, v.Id, v.needleMapKind, scanner)
 	return

weed/storage/volume_vacuum_test.go

@@ -5,6 +5,7 @@ import (
 	"math/rand"
 	"os"
 	"testing"
+	"time"
 
 	"github.com/chrislusf/seaweedfs/weed/storage/needle"
 	"github.com/chrislusf/seaweedfs/weed/storage/types"
@@ -72,8 +73,8 @@ func TestCompaction(t *testing.T) {
 		t.Fatalf("volume creation: %v", err)
 	}
 
-	beforeCommitFileCount := 1000
+	beforeCommitFileCount := 10000
-	afterCommitFileCount := 1000
+	afterCommitFileCount := 10000
 
 	infos := make([]*needleInfo, beforeCommitFileCount+afterCommitFileCount)
 
@@ -81,7 +82,10 @@ func TestCompaction(t *testing.T) {
 		doSomeWritesDeletes(i, v, t, infos)
 	}
 
-	v.Compact(0)
+	startTime := time.Now()
+	v.Compact(0, 1024*1024)
+	speed := float64(v.ContentSize()) / time.Now().Sub(startTime).Seconds()
+	t.Logf("compaction speed: %.2f bytes/s", speed)
 
 	for i := 1; i <= afterCommitFileCount; i++ {
 		doSomeWritesDeletes(i+beforeCommitFileCount, v, t, infos)