1
0
Fork 0
mirror of https://github.com/chrislusf/seaweedfs synced 2024-05-21 10:52:16 +02:00
seaweedfs/weed/filer/filechunks.go
Chris Lu d4566d4aaa
more solid weed mount (#4089)
* compare chunks by timestamp

* fix slab clearing error

* fix test compilation

* move oldest chunk to sealed, instead of by fullness

* lock on fh.entryViewCache

* remove verbose logs

* revert slat clearing

* less logs

* less logs

* track write and read by timestamp

* remove useless logic

* add entry lock on file handle release

* use mem chunk only, swap file chunk has problems

* comment out code that maybe used later

* add debug mode to compare data read and write

* more efficient readResolvedChunks with linked list

* small optimization

* fix test compilation

* minor fix on writer

* add SeparateGarbageChunks

* group chunks into sections

* turn off debug mode

* fix tests

* fix tests

* tmp enable swap file chunk

* Revert "tmp enable swap file chunk"

This reverts commit 985137ec47.

* simple refactoring

* simple refactoring

* do not re-use swap file chunk. Sealed chunks should not be re-used.

* comment out debugging facilities

* either mem chunk or swap file chunk is fine now

* remove orderedMutex  as *semaphore.Weighted

not found impactful

* optimize size calculation for changing large files

* optimize performance to avoid going through the long list of chunks

* still problems with swap file chunk

* rename

* tiny optimization

* swap file chunk save only successfully read data

* fix

* enable both mem and swap file chunk

* resolve chunks with range

* rename

* fix chunk interval list

* also change file handle chunk group when adding chunks

* pick in-active chunk with time-decayed counter

* fix compilation

* avoid nil with empty fh.entry

* refactoring

* rename

* rename

* refactor visible intervals to *list.List

* refactor chunkViews to *list.List

* add IntervalList for generic interval list

* change visible interval to use IntervalList in generics

* cahnge chunkViews to *IntervalList[*ChunkView]

* use NewFileChunkSection to create

* rename variables

* refactor

* fix renaming leftover

* renaming

* renaming

* add insert interval

* interval list adds lock

* incrementally add chunks to readers

Fixes:
1. set start and stop offset for the value object
2. clone the value object
3. use pointer instead of copy-by-value when passing to interval.Value
4. use insert interval since adding chunk could be out of order

* fix tests compilation

* fix tests compilation
2023-01-02 23:20:45 -08:00

310 lines
8.3 KiB
Go

package filer
import (
"bytes"
"fmt"
"github.com/seaweedfs/seaweedfs/weed/wdclient"
"math"
"github.com/seaweedfs/seaweedfs/weed/pb/filer_pb"
"github.com/seaweedfs/seaweedfs/weed/util"
)
func TotalSize(chunks []*filer_pb.FileChunk) (size uint64) {
for _, c := range chunks {
t := uint64(c.Offset + int64(c.Size))
if size < t {
size = t
}
}
return
}
func FileSize(entry *filer_pb.Entry) (size uint64) {
if entry == nil || entry.Attributes == nil {
return 0
}
fileSize := entry.Attributes.FileSize
if entry.RemoteEntry != nil {
if entry.RemoteEntry.RemoteMtime > entry.Attributes.Mtime {
fileSize = maxUint64(fileSize, uint64(entry.RemoteEntry.RemoteSize))
}
}
return maxUint64(TotalSize(entry.GetChunks()), fileSize)
}
func ETag(entry *filer_pb.Entry) (etag string) {
if entry.Attributes == nil || entry.Attributes.Md5 == nil {
return ETagChunks(entry.GetChunks())
}
return fmt.Sprintf("%x", entry.Attributes.Md5)
}
func ETagEntry(entry *Entry) (etag string) {
if entry.IsInRemoteOnly() {
return entry.Remote.RemoteETag
}
if entry.Attr.Md5 == nil {
return ETagChunks(entry.GetChunks())
}
return fmt.Sprintf("%x", entry.Attr.Md5)
}
func ETagChunks(chunks []*filer_pb.FileChunk) (etag string) {
if len(chunks) == 1 {
return fmt.Sprintf("%x", util.Base64Md5ToBytes(chunks[0].ETag))
}
var md5Digests [][]byte
for _, c := range chunks {
md5Digests = append(md5Digests, util.Base64Md5ToBytes(c.ETag))
}
return fmt.Sprintf("%x-%d", util.Md5(bytes.Join(md5Digests, nil)), len(chunks))
}
func CompactFileChunks(lookupFileIdFn wdclient.LookupFileIdFunctionType, chunks []*filer_pb.FileChunk) (compacted, garbage []*filer_pb.FileChunk) {
visibles, _ := NonOverlappingVisibleIntervals(lookupFileIdFn, chunks, 0, math.MaxInt64)
compacted, garbage = SeparateGarbageChunks(visibles, chunks)
return
}
func SeparateGarbageChunks(visibles *IntervalList[*VisibleInterval], chunks []*filer_pb.FileChunk) (compacted []*filer_pb.FileChunk, garbage []*filer_pb.FileChunk) {
fileIds := make(map[string]bool)
for x := visibles.Front(); x != nil; x = x.Next {
interval := x.Value
fileIds[interval.fileId] = true
}
for _, chunk := range chunks {
if _, found := fileIds[chunk.GetFileIdString()]; found {
compacted = append(compacted, chunk)
} else {
garbage = append(garbage, chunk)
}
}
return compacted, garbage
}
func MinusChunks(lookupFileIdFn wdclient.LookupFileIdFunctionType, as, bs []*filer_pb.FileChunk) (delta []*filer_pb.FileChunk, err error) {
aData, aMeta, aErr := ResolveChunkManifest(lookupFileIdFn, as, 0, math.MaxInt64)
if aErr != nil {
return nil, aErr
}
bData, bMeta, bErr := ResolveChunkManifest(lookupFileIdFn, bs, 0, math.MaxInt64)
if bErr != nil {
return nil, bErr
}
delta = append(delta, DoMinusChunks(aData, bData)...)
delta = append(delta, DoMinusChunks(aMeta, bMeta)...)
return
}
func DoMinusChunks(as, bs []*filer_pb.FileChunk) (delta []*filer_pb.FileChunk) {
fileIds := make(map[string]bool)
for _, interval := range bs {
fileIds[interval.GetFileIdString()] = true
}
for _, chunk := range as {
if _, found := fileIds[chunk.GetFileIdString()]; !found {
delta = append(delta, chunk)
}
}
return
}
func DoMinusChunksBySourceFileId(as, bs []*filer_pb.FileChunk) (delta []*filer_pb.FileChunk) {
fileIds := make(map[string]bool)
for _, interval := range bs {
fileIds[interval.GetFileIdString()] = true
fileIds[interval.GetSourceFileId()] = true
}
for _, chunk := range as {
_, sourceFileIdFound := fileIds[chunk.GetSourceFileId()]
_, fileIdFound := fileIds[chunk.GetFileId()]
if !sourceFileIdFound && !fileIdFound {
delta = append(delta, chunk)
}
}
return
}
type ChunkView struct {
FileId string
OffsetInChunk int64 // offset within the chunk
ViewSize uint64
ViewOffset int64 // actual offset in the file, for the data specified via [offset, offset+size) in current chunk
ChunkSize uint64
CipherKey []byte
IsGzipped bool
ModifiedTsNs int64
}
func (cv *ChunkView) SetStartStop(start, stop int64) {
cv.OffsetInChunk += start - cv.ViewOffset
cv.ViewOffset = start
cv.ViewSize = uint64(stop - start)
}
func (cv *ChunkView) Clone() IntervalValue {
return &ChunkView{
FileId: cv.FileId,
OffsetInChunk: cv.OffsetInChunk,
ViewSize: cv.ViewSize,
ViewOffset: cv.ViewOffset,
ChunkSize: cv.ChunkSize,
CipherKey: cv.CipherKey,
IsGzipped: cv.IsGzipped,
ModifiedTsNs: cv.ModifiedTsNs,
}
}
func (cv *ChunkView) IsFullChunk() bool {
return cv.ViewSize == cv.ChunkSize
}
func ViewFromChunks(lookupFileIdFn wdclient.LookupFileIdFunctionType, chunks []*filer_pb.FileChunk, offset int64, size int64) (chunkViews *IntervalList[*ChunkView]) {
visibles, _ := NonOverlappingVisibleIntervals(lookupFileIdFn, chunks, offset, offset+size)
return ViewFromVisibleIntervals(visibles, offset, size)
}
func ViewFromVisibleIntervals(visibles *IntervalList[*VisibleInterval], offset int64, size int64) (chunkViews *IntervalList[*ChunkView]) {
stop := offset + size
if size == math.MaxInt64 {
stop = math.MaxInt64
}
if stop < offset {
stop = math.MaxInt64
}
chunkViews = NewIntervalList[*ChunkView]()
for x := visibles.Front(); x != nil; x = x.Next {
chunk := x.Value
chunkStart, chunkStop := max(offset, chunk.start), min(stop, chunk.stop)
if chunkStart < chunkStop {
chunkView := &ChunkView{
FileId: chunk.fileId,
OffsetInChunk: chunkStart - chunk.start + chunk.offsetInChunk,
ViewSize: uint64(chunkStop - chunkStart),
ViewOffset: chunkStart,
ChunkSize: chunk.chunkSize,
CipherKey: chunk.cipherKey,
IsGzipped: chunk.isGzipped,
ModifiedTsNs: chunk.modifiedTsNs,
}
chunkViews.AppendInterval(&Interval[*ChunkView]{
StartOffset: chunkStart,
StopOffset: chunkStop,
TsNs: chunk.modifiedTsNs,
Value: chunkView,
Prev: nil,
Next: nil,
})
}
}
return chunkViews
}
func MergeIntoVisibles(visibles *IntervalList[*VisibleInterval], start int64, stop int64, chunk *filer_pb.FileChunk) {
newV := &VisibleInterval{
start: start,
stop: stop,
fileId: chunk.GetFileIdString(),
modifiedTsNs: chunk.ModifiedTsNs,
offsetInChunk: start - chunk.Offset, // the starting position in the chunk
chunkSize: chunk.Size, // size of the chunk
cipherKey: chunk.CipherKey,
isGzipped: chunk.IsCompressed,
}
visibles.InsertInterval(start, stop, chunk.ModifiedTsNs, newV)
}
func MergeIntoChunkViews(chunkViews *IntervalList[*ChunkView], start int64, stop int64, chunk *filer_pb.FileChunk) {
chunkView := &ChunkView{
FileId: chunk.GetFileIdString(),
OffsetInChunk: start - chunk.Offset,
ViewSize: uint64(stop - start),
ViewOffset: start,
ChunkSize: chunk.Size,
CipherKey: chunk.CipherKey,
IsGzipped: chunk.IsCompressed,
ModifiedTsNs: chunk.ModifiedTsNs,
}
chunkViews.InsertInterval(start, stop, chunk.ModifiedTsNs, chunkView)
}
// 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, startOffset int64, stopOffset int64) (visibles *IntervalList[*VisibleInterval], err error) {
chunks, _, err = ResolveChunkManifest(lookupFileIdFn, chunks, startOffset, stopOffset)
if err != nil {
return
}
visibles2 := readResolvedChunks(chunks, 0, math.MaxInt64)
return visibles2, err
}
// find non-overlapping visible intervals
// visible interval map to one file chunk
type VisibleInterval struct {
start int64
stop int64
modifiedTsNs int64
fileId string
offsetInChunk int64
chunkSize uint64
cipherKey []byte
isGzipped bool
}
func (v *VisibleInterval) SetStartStop(start, stop int64) {
v.offsetInChunk += start - v.start
v.start, v.stop = start, stop
}
func (v *VisibleInterval) Clone() IntervalValue {
return &VisibleInterval{
start: v.start,
stop: v.stop,
modifiedTsNs: v.modifiedTsNs,
fileId: v.fileId,
offsetInChunk: v.offsetInChunk,
chunkSize: v.chunkSize,
cipherKey: v.cipherKey,
isGzipped: v.isGzipped,
}
}
func min(x, y int64) int64 {
if x <= y {
return x
}
return y
}
func max(x, y int64) int64 {
if x <= y {
return y
}
return x
}