optimization: improve random range query for large files

This commit is contained in:
Chris Lu 2021-07-19 23:07:22 -07:00
parent b938df97a2
commit 7ab389e7ec
7 changed files with 22 additions and 16 deletions

View file

@ -39,9 +39,14 @@ func SeparateManifestChunks(chunks []*filer_pb.FileChunk) (manifestChunks, nonMa
return
}
func ResolveChunkManifest(lookupFileIdFn wdclient.LookupFileIdFunctionType, chunks []*filer_pb.FileChunk) (dataChunks, manifestChunks []*filer_pb.FileChunk, manifestResolveErr error) {
func ResolveChunkManifest(lookupFileIdFn wdclient.LookupFileIdFunctionType, chunks []*filer_pb.FileChunk, startOffset, stopOffset int64) (dataChunks, manifestChunks []*filer_pb.FileChunk, manifestResolveErr error) {
// TODO maybe parallel this
for _, chunk := range chunks {
if max(chunk.Offset, startOffset) >= min(chunk.Offset+int64(chunk.Size), stopOffset) {
continue
}
if !chunk.IsChunkManifest {
dataChunks = append(dataChunks, chunk)
continue
@ -54,7 +59,7 @@ func ResolveChunkManifest(lookupFileIdFn wdclient.LookupFileIdFunctionType, chun
manifestChunks = append(manifestChunks, chunk)
// recursive
dchunks, mchunks, subErr := ResolveChunkManifest(lookupFileIdFn, resolvedChunks)
dchunks, mchunks, subErr := ResolveChunkManifest(lookupFileIdFn, resolvedChunks, startOffset, stopOffset)
if subErr != nil {
return chunks, nil, subErr
}

View file

@ -53,7 +53,7 @@ func ETagChunks(chunks []*filer_pb.FileChunk) (etag string) {
func CompactFileChunks(lookupFileIdFn wdclient.LookupFileIdFunctionType, chunks []*filer_pb.FileChunk) (compacted, garbage []*filer_pb.FileChunk) {
visibles, _ := NonOverlappingVisibleIntervals(lookupFileIdFn, chunks)
visibles, _ := NonOverlappingVisibleIntervals(lookupFileIdFn, chunks, 0, math.MaxInt64)
fileIds := make(map[string]bool)
for _, interval := range visibles {
@ -72,11 +72,11 @@ func CompactFileChunks(lookupFileIdFn wdclient.LookupFileIdFunctionType, chunks
func MinusChunks(lookupFileIdFn wdclient.LookupFileIdFunctionType, as, bs []*filer_pb.FileChunk) (delta []*filer_pb.FileChunk, err error) {
aData, aMeta, aErr := ResolveChunkManifest(lookupFileIdFn, as)
aData, aMeta, aErr := ResolveChunkManifest(lookupFileIdFn, as, 0, math.MaxInt64)
if aErr != nil {
return nil, aErr
}
bData, bMeta, bErr := ResolveChunkManifest(lookupFileIdFn, bs)
bData, bMeta, bErr := ResolveChunkManifest(lookupFileIdFn, bs, 0, math.MaxInt64)
if bErr != nil {
return nil, bErr
}
@ -117,7 +117,7 @@ func (cv *ChunkView) IsFullChunk() bool {
func ViewFromChunks(lookupFileIdFn wdclient.LookupFileIdFunctionType, chunks []*filer_pb.FileChunk, offset int64, size int64) (views []*ChunkView) {
visibles, _ := NonOverlappingVisibleIntervals(lookupFileIdFn, chunks)
visibles, _ := NonOverlappingVisibleIntervals(lookupFileIdFn, chunks, offset, offset+size)
return ViewFromVisibleIntervals(visibles, offset, size)
@ -221,9 +221,9 @@ func MergeIntoVisibles(visibles []VisibleInterval, chunk *filer_pb.FileChunk) (n
// NonOverlappingVisibleIntervals translates the file chunk into VisibleInterval in memory
// If the file chunk content is a chunk manifest
func NonOverlappingVisibleIntervals(lookupFileIdFn wdclient.LookupFileIdFunctionType, chunks []*filer_pb.FileChunk) (visibles []VisibleInterval, err error) {
func NonOverlappingVisibleIntervals(lookupFileIdFn wdclient.LookupFileIdFunctionType, chunks []*filer_pb.FileChunk, startOffset int64, stopOffset int64) (visibles []VisibleInterval, err error) {
chunks, _, err = ResolveChunkManifest(lookupFileIdFn, chunks)
chunks, _, err = ResolveChunkManifest(lookupFileIdFn, chunks, startOffset, stopOffset)
sort.Slice(chunks, func(i, j int) bool {
if chunks[i].Mtime == chunks[j].Mtime {

View file

@ -90,7 +90,7 @@ func TestRandomFileChunksCompact(t *testing.T) {
}
}
visibles, _ := NonOverlappingVisibleIntervals(nil, chunks)
visibles, _ := NonOverlappingVisibleIntervals(nil, chunks, 0, math.MaxInt64)
for _, v := range visibles {
for x := v.start; x < v.stop; x++ {
@ -227,7 +227,7 @@ func TestIntervalMerging(t *testing.T) {
for i, testcase := range testcases {
log.Printf("++++++++++ merged test case %d ++++++++++++++++++++", i)
intervals, _ := NonOverlappingVisibleIntervals(nil, testcase.Chunks)
intervals, _ := NonOverlappingVisibleIntervals(nil, testcase.Chunks, 0, math.MaxInt64)
for x, interval := range intervals {
log.Printf("test case %d, interval %d, start=%d, stop=%d, fileId=%s",
i, x, interval.start, interval.stop, interval.fileId)

View file

@ -130,7 +130,7 @@ func (fh *FileHandle) readFromChunks(buff []byte, offset int64) (int64, error) {
var chunkResolveErr error
if fh.entryViewCache == nil {
fh.entryViewCache, chunkResolveErr = filer.NonOverlappingVisibleIntervals(fh.f.wfs.LookupFn(), entry.Chunks)
fh.entryViewCache, chunkResolveErr = filer.NonOverlappingVisibleIntervals(fh.f.wfs.LookupFn(), entry.Chunks, 0, math.MaxInt64)
if chunkResolveErr != nil {
return 0, fmt.Errorf("fail to resolve chunk manifest: %v", chunkResolveErr)
}

View file

@ -5,6 +5,7 @@ import (
"fmt"
"github.com/chrislusf/seaweedfs/weed/pb"
"github.com/chrislusf/seaweedfs/weed/wdclient"
"math"
"google.golang.org/grpc"
@ -228,11 +229,11 @@ func (fs *FilerSink) UpdateEntry(key string, oldEntry *filer_pb.Entry, newParent
}
func compareChunks(lookupFileIdFn wdclient.LookupFileIdFunctionType, oldEntry, newEntry *filer_pb.Entry) (deletedChunks, newChunks []*filer_pb.FileChunk, err error) {
aData, aMeta, aErr := filer.ResolveChunkManifest(lookupFileIdFn, oldEntry.Chunks)
aData, aMeta, aErr := filer.ResolveChunkManifest(lookupFileIdFn, oldEntry.Chunks, 0, math.MaxInt64)
if aErr != nil {
return nil, nil, aErr
}
bData, bMeta, bErr := filer.ResolveChunkManifest(lookupFileIdFn, newEntry.Chunks)
bData, bMeta, bErr := filer.ResolveChunkManifest(lookupFileIdFn, newEntry.Chunks, 0, math.MaxInt64)
if bErr != nil {
return nil, nil, bErr
}

View file

@ -532,7 +532,7 @@ func (f *WebDavFile) Read(p []byte) (readSize int, err error) {
return 0, io.EOF
}
if f.entryViewCache == nil {
f.entryViewCache, _ = filer.NonOverlappingVisibleIntervals(filer.LookupFn(f.fs), f.entry.Chunks)
f.entryViewCache, _ = filer.NonOverlappingVisibleIntervals(filer.LookupFn(f.fs), f.entry.Chunks, 0, math.MaxInt64)
f.reader = nil
}
if f.reader == nil {

View file

@ -164,7 +164,7 @@ func (c *commandVolumeFsck) collectFilerFileIdAndPaths(volumeIdToServer map[uint
if verbose && entry.Entry.IsDirectory {
fmt.Fprintf(writer, "checking directory %s\n", util.NewFullPath(entry.Dir, entry.Entry.Name))
}
dChunks, mChunks, resolveErr := filer.ResolveChunkManifest(filer.LookupFn(c.env), entry.Entry.Chunks)
dChunks, mChunks, resolveErr := filer.ResolveChunkManifest(filer.LookupFn(c.env), entry.Entry.Chunks, 0, math.MaxInt64)
if resolveErr != nil {
return nil
}
@ -311,7 +311,7 @@ func (c *commandVolumeFsck) collectFilerFileIds(tempFolder string, volumeIdToSer
files[i.vid].Write(buffer)
}
}, func(entry *filer_pb.FullEntry, outputChan chan interface{}) (err error) {
dChunks, mChunks, resolveErr := filer.ResolveChunkManifest(filer.LookupFn(c.env), entry.Entry.Chunks)
dChunks, mChunks, resolveErr := filer.ResolveChunkManifest(filer.LookupFn(c.env), entry.Entry.Chunks, 0, math.MaxInt64)
if resolveErr != nil {
return nil
}