FUSE: add chunk cache for recently accessed file chunks

weed/filesys/dirty_page.go

@@ -183,6 +183,7 @@ func (pages *ContinuousDirtyPages) saveToStorage(reader io.Reader, offset int64,
 		glog.V(0).Infof("upload failure %v to %s: %v", pages.f.Name, fileUrl, err)
 		return nil, fmt.Errorf("upload result: %v", uploadResult.Error)
 	}
+	pages.f.wfs.chunkCache.SetChunk(fileId, data)
 
 	return &filer_pb.FileChunk{
 		FileId:    fileId,

weed/filesys/filehandle.go

@@ -92,7 +92,7 @@ func (fh *FileHandle) readFromChunks(buff []byte, offset int64) (int64, error) {
 
 	if fh.f.reader == nil {
 		chunkViews := filer2.ViewFromVisibleIntervals(fh.f.entryViewCache, 0, math.MaxInt32)
-		fh.f.reader = filer2.NewChunkReaderAtFromClient(fh.f.wfs, chunkViews)
+		fh.f.reader = NewChunkReaderAtFromClient(fh.f.wfs, chunkViews, fh.f.wfs.chunkCache)
 	}
 
 	totalRead, err := fh.f.reader.ReadAt(buff, offset)
@@ -153,6 +153,8 @@ func (fh *FileHandle) Release(ctx context.Context, req *fuse.ReleaseRequest) err
 		fh.dirtyPages.releaseResource()
 		fh.f.wfs.ReleaseHandle(fh.f.fullpath(), fuse.HandleID(fh.handle))
 	}
+	fh.f.entryViewCache = nil
+	fh.f.reader = nil
 
 	return nil
 }

weed/filesys/reader_at.go

@@ -1,4 +1,4 @@
-package filer2
+package filesys
 
 import (
 	"bytes"
@@ -7,30 +7,34 @@ import (
 	"io"
 	"sync"
 
+	"github.com/chrislusf/seaweedfs/weed/filer2"
 	"github.com/chrislusf/seaweedfs/weed/glog"
 	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
+	"github.com/chrislusf/seaweedfs/weed/pb/pb_cache"
 	"github.com/chrislusf/seaweedfs/weed/util"
 	"github.com/chrislusf/seaweedfs/weed/wdclient"
 )
 
 type ChunkReadAt struct {
 	masterClient *wdclient.MasterClient
-	chunkViews   []*ChunkView
+	chunkViews   []*filer2.ChunkView
 	buffer       []byte
 	bufferOffset int64
 	lookupFileId func(fileId string) (targetUrl string, err error)
 	readerLock   sync.Mutex
+
+	chunkCache *pb_cache.ChunkCache
 }
 
 // var _ = io.ReaderAt(&ChunkReadAt{})
 
-func NewChunkReaderAtFromClient(filerClient filer_pb.FilerClient, chunkViews   []*ChunkView) *ChunkReadAt {
+func NewChunkReaderAtFromClient(filerClient filer_pb.FilerClient, chunkViews []*filer2.ChunkView, chunkCache *pb_cache.ChunkCache) *ChunkReadAt {
 
 	return &ChunkReadAt{
 		chunkViews: chunkViews,
 		lookupFileId: func(fileId string) (targetUrl string, err error) {
 			err = filerClient.WithFilerClient(func(client filer_pb.SeaweedFilerClient) error {
-				vid := VolumeId(fileId)
+				vid := filer2.VolumeId(fileId)
 				resp, err := client.LookupVolume(context.Background(), &filer_pb.LookupVolumeRequest{
 					VolumeIds: []string{vid},
 				})
@@ -61,7 +65,6 @@ func (c *ChunkReadAt) ReadAt(p []byte, offset int64) (n int, err error) {
 	c.readerLock.Lock()
 	defer c.readerLock.Unlock()
 
-
 	for n < len(p) && err == nil {
 		readCount, readErr := c.doReadAt(p[n:], offset+int64(n))
 		n += readCount
@@ -80,7 +83,8 @@ func (c *ChunkReadAt) doReadAt(p []byte, offset int64) (n int, err error) {
 		if chunk.LogicOffset <= offset && offset < chunk.LogicOffset+int64(chunk.Size) {
 			found = true
 			if c.bufferOffset != chunk.LogicOffset {
-				c.fetchChunkToBuffer(chunk)
+				c.buffer, err = c.fetchChunkData(chunk)
+				c.bufferOffset = chunk.LogicOffset
 			}
 			break
 		}
@@ -97,27 +101,34 @@ func (c *ChunkReadAt) doReadAt(p []byte, offset int64) (n int, err error) {
 
 }
 
-func (c *ChunkReadAt) fetchChunkToBuffer(chunkView *ChunkView) error {
+func (c *ChunkReadAt) fetchChunkData(chunkView *filer2.ChunkView) ([]byte, error) {
 
-	// fmt.Printf("fetching %s [%d,%d)\n", chunkView.FileId, chunkView.LogicOffset, chunkView.LogicOffset+int64(chunkView.Size))
+	fmt.Printf("fetching %s [%d,%d)\n", chunkView.FileId, chunkView.LogicOffset, chunkView.LogicOffset+int64(chunkView.Size))
+
+	chunkData := c.chunkCache.GetChunk(chunkView.FileId)
+	if chunkData != nil {
+		glog.V(3).Infof("cache hit %s [%d,%d)", chunkView.FileId, chunkView.LogicOffset, chunkView.LogicOffset+int64(chunkView.Size))
+		return chunkData, nil
+	}
 
 	urlString, err := c.lookupFileId(chunkView.FileId)
 	if err != nil {
 		glog.V(1).Infof("operation LookupFileId %s failed, err: %v", chunkView.FileId, err)
-		return err
+		return nil, err
 	}
 	var buffer bytes.Buffer
-	err = util.ReadUrlAsStream(urlString, chunkView.CipherKey, chunkView.isGzipped, chunkView.IsFullChunk, chunkView.Offset, int(chunkView.Size), func(data []byte) {
+	err = util.ReadUrlAsStream(urlString, chunkView.CipherKey, chunkView.IsGzipped, chunkView.IsFullChunk, chunkView.Offset, int(chunkView.Size), func(data []byte) {
 		buffer.Write(data)
 	})
 	if err != nil {
 		glog.V(1).Infof("read %s failed, err: %v", chunkView.FileId, err)
-		return err
+		return nil, err
 	}
-	c.buffer = buffer.Bytes()
-	c.bufferOffset = chunkView.LogicOffset
 
 	glog.V(3).Infof("read %s [%d,%d)", chunkView.FileId, chunkView.LogicOffset, chunkView.LogicOffset+int64(chunkView.Size))
 
-	return nil
+	chunkData = buffer.Bytes()
+	c.chunkCache.SetChunk(chunkView.FileId, chunkData)
+
+	return chunkData, nil
 }

weed/filesys/wfs.go

@@ -15,6 +15,7 @@ import (
 	"github.com/chrislusf/seaweedfs/weed/glog"
 	"github.com/chrislusf/seaweedfs/weed/pb"
 	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
+	"github.com/chrislusf/seaweedfs/weed/pb/pb_cache"
 	"github.com/chrislusf/seaweedfs/weed/util"
 	"github.com/seaweedfs/fuse"
 	"github.com/seaweedfs/fuse/fs"
@@ -62,6 +63,8 @@ type WFS struct {
 
 	root        fs.Node
 	fsNodeCache *FsCache
+
+	chunkCache *pb_cache.ChunkCache
 }
 type statsCache struct {
 	filer_pb.StatisticsResponse
@@ -78,6 +81,7 @@ func NewSeaweedFileSystem(option *Option) *WFS {
 				return make([]byte, option.ChunkSizeLimit)
 			},
 		},
+		chunkCache: pb_cache.NewChunkCache(),
 	}
 
 	wfs.root = &Dir{name: wfs.option.FilerMountRootPath, wfs: wfs}

weed/pb/pb_cache/chunk_cache.go

@@ -0,0 +1,32 @@
+package pb_cache
+
+import (
+	"time"
+
+	"github.com/karlseguin/ccache"
+)
+
+// a global cache for recently accessed file chunks
+type ChunkCache struct {
+	cache *ccache.Cache
+}
+
+func NewChunkCache() *ChunkCache {
+	return &ChunkCache{
+		cache: ccache.New(ccache.Configure().MaxSize(1000).ItemsToPrune(100)),
+	}
+}
+
+func (c *ChunkCache) GetChunk(fileId string) []byte {
+	item := c.cache.Get(fileId)
+	if item == nil {
+		return nil
+	}
+	data := item.Value().([]byte)
+	item.Extend(time.Hour)
+	return data
+}
+
+func (c *ChunkCache) SetChunk(fileId string, data []byte) {
+	c.cache.Set(fileId, data, time.Hour)
+}

weed/server/webdav_server.go

@@ -13,9 +13,11 @@ import (
 	"golang.org/x/net/webdav"
 	"google.golang.org/grpc"
 
+	"github.com/chrislusf/seaweedfs/weed/filesys"
 	"github.com/chrislusf/seaweedfs/weed/operation"
 	"github.com/chrislusf/seaweedfs/weed/pb"
 	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
+	"github.com/chrislusf/seaweedfs/weed/pb/pb_cache"
 	"github.com/chrislusf/seaweedfs/weed/util"
 
 	"github.com/chrislusf/seaweedfs/weed/filer2"
@@ -66,6 +68,7 @@ type WebDavFileSystem struct {
 	secret         security.SigningKey
 	filer          *filer2.Filer
 	grpcDialOption grpc.DialOption
+	chunkCache     *pb_cache.ChunkCache
 }
 
 type FileInfo struct {
@@ -96,6 +99,7 @@ type WebDavFile struct {
 func NewWebDavFileSystem(option *WebDavOption) (webdav.FileSystem, error) {
 	return &WebDavFileSystem{
 		option:     option,
+		chunkCache: pb_cache.NewChunkCache(),
 	}, nil
 }
 
@@ -476,7 +480,7 @@ func (f *WebDavFile) Read(p []byte) (readSize int, err error) {
 	}
 	if f.reader == nil {
 		chunkViews := filer2.ViewFromVisibleIntervals(f.entryViewCache, 0, math.MaxInt32)
-		f.reader = filer2.NewChunkReaderAtFromClient(f.fs, chunkViews)
+		f.reader = filesys.NewChunkReaderAtFromClient(f.fs, chunkViews, f.fs.chunkCache)
 	}
 
 	readSize, err = f.reader.ReadAt(p, f.off)