seaweedfs/weed/filesys/dirty_pages_mem_chunk.go
2022-01-17 14:15:10 -08:00

109 lines
2.7 KiB
Go

package filesys
import (
"fmt"
"github.com/chrislusf/seaweedfs/weed/filesys/page_writer"
"github.com/chrislusf/seaweedfs/weed/glog"
"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
"io"
"sync"
"time"
)
type MemoryChunkPages struct {
fh *FileHandle
writeWaitGroup sync.WaitGroup
chunkAddLock sync.Mutex
lastErr error
collection string
replication string
uploadPipeline *page_writer.UploadPipeline
hasWrites bool
}
func newMemoryChunkPages(fh *FileHandle, chunkSize int64) *MemoryChunkPages {
dirtyPages := &MemoryChunkPages{
fh: fh,
}
dirtyPages.uploadPipeline = page_writer.NewUploadPipeline(fh.f.fullpath(),
fh.f.wfs.concurrentWriters, chunkSize, dirtyPages.saveChunkedFileIntevalToStorage)
return dirtyPages
}
func (pages *MemoryChunkPages) AddPage(offset int64, data []byte) {
pages.hasWrites = true
glog.V(4).Infof("%v memory AddPage [%d, %d)", pages.fh.f.fullpath(), offset, offset+int64(len(data)))
pages.uploadPipeline.SaveDataAt(data, offset)
return
}
func (pages *MemoryChunkPages) FlushData() error {
if !pages.hasWrites {
return nil
}
pages.saveChunkedFileToStorage()
pages.writeWaitGroup.Wait()
if pages.lastErr != nil {
return fmt.Errorf("flush data: %v", pages.lastErr)
}
return nil
}
func (pages *MemoryChunkPages) ReadDirtyDataAt(data []byte, startOffset int64) (maxStop int64) {
if !pages.hasWrites {
return
}
return pages.uploadPipeline.MaybeReadDataAt(data, startOffset)
}
func (pages *MemoryChunkPages) GetStorageOptions() (collection, replication string) {
return pages.collection, pages.replication
}
func (pages *MemoryChunkPages) saveChunkedFileToStorage() {
pages.uploadPipeline.FlushAll()
}
func (pages *MemoryChunkPages) saveChunkedFileIntevalToStorage(reader io.Reader, offset int64, size int64, cleanupFn func()) {
mtime := time.Now().UnixNano()
pages.writeWaitGroup.Add(1)
writer := func() {
defer pages.writeWaitGroup.Done()
defer cleanupFn()
chunk, collection, replication, err := pages.fh.f.wfs.saveDataAsChunk(pages.fh.f.fullpath())(reader, pages.fh.f.Name, offset)
if err != nil {
glog.V(0).Infof("%s saveToStorage [%d,%d): %v", pages.fh.f.fullpath(), offset, offset+size, err)
pages.lastErr = err
return
}
chunk.Mtime = mtime
pages.collection, pages.replication = collection, replication
pages.chunkAddLock.Lock()
pages.fh.f.addChunks([]*filer_pb.FileChunk{chunk})
pages.fh.entryViewCache = nil
glog.V(3).Infof("%s saveToStorage %s [%d,%d)", pages.fh.f.fullpath(), chunk.FileId, offset, offset+size)
pages.chunkAddLock.Unlock()
}
if pages.fh.f.wfs.concurrentWriters != nil {
pages.fh.f.wfs.concurrentWriters.Execute(writer)
} else {
go writer()
}
}
func (pages MemoryChunkPages) Destroy() {
pages.uploadPipeline.Shutdown()
}