gitlab-com-seaweedfs/seaweedfs
1
0
Fork 0
mirror of https://github.com/seaweedfs/seaweedfs.git synced 2024-01-19 02:48:24 +00:00
Code Issues Projects Releases Wiki Activity

cache local writes before flushing to volume server

Browse source
This commit is contained in:
Chris Lu 2018-05-28 12:30:17 -07:00
parent 07e0d13d2d
commit d0b238d2db
3 changed files with 182 additions and 56 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

152
weed/filesys/dirty_page.go Normal file
View file

@ -0,0 +1,152 @@
package filesys
import (
"sync"
"sort"
"fmt"
"bytes"
"io"
"time"
"context"
"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
"github.com/chrislusf/seaweedfs/weed/operation"
"github.com/chrislusf/seaweedfs/weed/glog"
)
type DirtyPage struct {
Offset int64
Data []byte
}
type ContinuousDirtyPages struct {
sync.Mutex
pages []*DirtyPage
f *File
}
func (pages *ContinuousDirtyPages) AddPage(ctx context.Context, offset int64, data []byte) (chunk *filer_pb.FileChunk, err error) {
pages.Lock()
defer pages.Unlock()
isPerfectAppend := len(pages.pages) == 0
if len(pages.pages) > 0 {
lastPage := pages.pages[len(pages.pages)-1]
if lastPage.Offset+int64(len(lastPage.Data)) == offset {
// write continuous pages
glog.V(3).Infof("%s/%s append [%d,%d)", pages.f.dir.Path, pages.f.Name, offset, offset+int64(len(data)))
isPerfectAppend = true
}
}
isPerfectReplace := false
for _, page := range pages.pages {
if page.Offset == offset && len(page.Data) == len(data) {
// perfect replace
glog.V(3).Infof("%s/%s replace [%d,%d)", pages.f.dir.Path, pages.f.Name, offset, offset+int64(len(data)))
page.Data = data
isPerfectReplace = true
}
}
if isPerfectReplace {
return nil, nil
}
if isPerfectAppend {
pages.pages = append(pages.pages, &DirtyPage{
Offset: offset,
Data: data,
})
return nil, nil
}
chunk, err = pages.saveToStorage(ctx)
glog.V(3).Infof("%s/%s saved [%d,%d)", pages.f.dir.Path, pages.f.Name, chunk.Offset, chunk.Offset+int64(chunk.Size))
pages.pages = []*DirtyPage{&DirtyPage{
Offset: offset,
Data: data,
}}
return
}
func (pages *ContinuousDirtyPages) FlushToStorage(ctx context.Context) (chunk *filer_pb.FileChunk, err error) {
pages.Lock()
defer pages.Unlock()
if chunk, err = pages.saveToStorage(ctx); err == nil {
pages.pages = nil
}
return
}
func (pages *ContinuousDirtyPages) totalSize() (total int64) {
for _, page := range pages.pages {
total += int64(len(page.Data))
}
return
}
func (pages *ContinuousDirtyPages) saveToStorage(ctx context.Context) (*filer_pb.FileChunk, error) {
if len(pages.pages) == 0 {
return nil, nil
}
sort.Slice(pages.pages, func(i, j int) bool {
return pages.pages[i].Offset < pages.pages[j].Offset
})
var fileId, host string
if err := pages.f.wfs.withFilerClient(func(client filer_pb.SeaweedFilerClient) error {
request := &filer_pb.AssignVolumeRequest{
Count: 1,
Replication: "000",
Collection: "",
}
resp, err := client.AssignVolume(ctx, request)
if err != nil {
glog.V(0).Infof("assign volume failure %v: %v", request, err)
return err
}
fileId, host = resp.FileId, resp.Url
return nil
}); err != nil {
return nil, fmt.Errorf("filer assign volume: %v", err)
}
var readers []io.Reader
for _, page := range pages.pages {
readers = append(readers, bytes.NewReader(page.Data))
}
fileUrl := fmt.Sprintf("http://%s/%s", host, fileId)
bufReader := io.MultiReader(readers...)
uploadResult, err := operation.Upload(fileUrl, pages.f.Name, bufReader, false, "application/octet-stream", nil, "")
if err != nil {
glog.V(0).Infof("upload data %v to %s: %v", pages.f.Name, fileUrl, err)
return nil, fmt.Errorf("upload data: %v", err)
}
if uploadResult.Error != "" {
glog.V(0).Infof("upload failure %v to %s: %v", pages.f.Name, fileUrl, err)
return nil, fmt.Errorf("upload result: %v", uploadResult.Error)
}
return &filer_pb.FileChunk{
FileId: fileId,
Offset: pages.pages[0].Offset,
Size: uint64(pages.totalSize()),
Mtime: time.Now().UnixNano(),
}, nil
}

1
weed/filesys/file.go
View file

@ -85,6 +85,7 @@ func (file *File) Open(ctx context.Context, req *fuse.OpenRequest, resp *fuse.Op
return &FileHandle{
f: file,
dirtyPages: &ContinuousDirtyPages{f: file},
RequestId: req.Header.ID,
NodeId: req.Header.Node,
Uid: req.Uid,

73
weed/filesys/filehandle.go
View file

@ -3,22 +3,20 @@ package filesys
import (
"bazil.org/fuse"
"bazil.org/fuse/fs"
"bytes"
"context"
"fmt"
"github.com/chrislusf/seaweedfs/weed/filer2"
"github.com/chrislusf/seaweedfs/weed/glog"
"github.com/chrislusf/seaweedfs/weed/operation"
"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
"github.com/chrislusf/seaweedfs/weed/util"
"strings"
"sync"
"time"
)
type FileHandle struct {
// cache file has been written to
dirty bool
dirtyPages *ContinuousDirtyPages
dirtyMetadata bool
cachePath string
@ -128,56 +126,21 @@ func (fh *FileHandle) Write(ctx context.Context, req *fuse.WriteRequest, resp *f
// write the request to volume servers
glog.V(3).Infof("%+v/%v write fh: %+v", fh.f.dir.Path, fh.f.Name, req)
glog.V(3).Infof("%+v/%v write fh: [%d,%d)", fh.f.dir.Path, fh.f.Name, req.Offset, req.Offset+int64(len(req.Data)))
var fileId, host string
if err := fh.f.wfs.withFilerClient(func(client filer_pb.SeaweedFilerClient) error {
request := &filer_pb.AssignVolumeRequest{
Count: 1,
Replication: "000",
Collection: "",
}
resp, err := client.AssignVolume(ctx, request)
chunk, err := fh.dirtyPages.AddPage(ctx, req.Offset, req.Data)
if err != nil {
glog.V(0).Infof("assign volume failure %v: %v", request, err)
return err
return fmt.Errorf("write %s/%s at [%d,%d): %v", fh.f.dir.Path, fh.f.Name, req.Offset, req.Offset+int64(len(req.Data)), err)
}
fileId, host = resp.FileId, resp.Url
resp.Size = len(req.Data)
return nil
}); err != nil {
return fmt.Errorf("filer assign volume: %v", err)
if chunk != nil {
fh.f.Chunks = append(fh.f.Chunks, chunk)
glog.V(1).Infof("uploaded %s/%s to %s [%d,%d)", fh.f.dir.Path, fh.f.Name, chunk.FileId, chunk.Offset, chunk.Offset+int64(chunk.Size))
fh.dirtyMetadata = true
}
fileUrl := fmt.Sprintf("http://%s/%s", host, fileId)
bufReader := bytes.NewReader(req.Data)
uploadResult, err := operation.Upload(fileUrl, fh.f.Name, bufReader, false, "application/octet-stream", nil, "")
if err != nil {
glog.V(0).Infof("upload data %v to %s: %v", req, fileUrl, err)
return fmt.Errorf("upload data: %v", err)
}
if uploadResult.Error != "" {
glog.V(0).Infof("upload failure %v to %s: %v", req, fileUrl, err)
return fmt.Errorf("upload result: %v", uploadResult.Error)
}
resp.Size = int(uploadResult.Size)
fh.f.Chunks = append(fh.f.Chunks, &filer_pb.FileChunk{
FileId: fileId,
Offset: req.Offset,
Size: uint64(uploadResult.Size),
Mtime: time.Now().UnixNano(),
})
glog.V(1).Infof("uploaded %s/%s to: %v, [%d,%d)", fh.f.dir.Path, fh.f.Name, fileUrl, req.Offset, req.Offset+int64(resp.Size))
fh.dirty = true
return nil
}
@ -197,7 +160,17 @@ func (fh *FileHandle) Flush(ctx context.Context, req *fuse.FlushRequest) error {
// send the data to the OS
glog.V(3).Infof("%s/%s fh flush %v", fh.f.dir.Path, fh.f.Name, req)
if !fh.dirty {
chunk, err := fh.dirtyPages.FlushToStorage(ctx)
if err != nil {
glog.V(0).Infof("flush %s/%s to %s [%d,%d): %v", fh.f.dir.Path, fh.f.Name, chunk.FileId, chunk.Offset, chunk.Offset+int64(chunk.Size), err)
return fmt.Errorf("flush %s/%s to %s [%d,%d): %v", fh.f.dir.Path, fh.f.Name, chunk.FileId, chunk.Offset, chunk.Offset+int64(chunk.Size), err)
}
if chunk != nil {
fh.f.Chunks = append(fh.f.Chunks, chunk)
fh.dirtyMetadata = true
}
if !fh.dirtyMetadata {
return nil
}
@ -206,7 +179,7 @@ func (fh *FileHandle) Flush(ctx context.Context, req *fuse.FlushRequest) error {
return nil
}
err := fh.f.wfs.withFilerClient(func(client filer_pb.SeaweedFilerClient) error {
err = fh.f.wfs.withFilerClient(func(client filer_pb.SeaweedFilerClient) error {
request := &filer_pb.UpdateEntryRequest{
Directory: fh.f.dir.Path,
@ -229,7 +202,7 @@ func (fh *FileHandle) Flush(ctx context.Context, req *fuse.FlushRequest) error {
})
if err == nil {
fh.dirty = false
fh.dirtyMetadata = false
}
return err