aboutsummaryrefslogtreecommitdiff
path: root/weed/storage/erasure_coding
diff options
context:
space:
mode:
Diffstat (limited to 'weed/storage/erasure_coding')
-rw-r--r--weed/storage/erasure_coding/ec_encoder.go128
-rw-r--r--weed/storage/erasure_coding/ec_test.go106
2 files changed, 123 insertions, 111 deletions
diff --git a/weed/storage/erasure_coding/ec_encoder.go b/weed/storage/erasure_coding/ec_encoder.go
index 4b5205dee..f4aa17e9d 100644
--- a/weed/storage/erasure_coding/ec_encoder.go
+++ b/weed/storage/erasure_coding/ec_encoder.go
@@ -6,6 +6,9 @@ import (
"os"
"github.com/chrislusf/seaweedfs/weed/glog"
+ "github.com/chrislusf/seaweedfs/weed/storage"
+ "github.com/chrislusf/seaweedfs/weed/storage/needle_map"
+ "github.com/chrislusf/seaweedfs/weed/storage/types"
"github.com/klauspost/reedsolomon"
)
@@ -16,11 +19,66 @@ const (
ErasureCodingSmallBlockSize = 1024 * 1024 // 1MB
)
+// WriteSortedEcxFile generates .ecx file from existing .idx file
+// all keys are sorted in ascending order
+func WriteSortedEcxFile(baseFileName string) (e error) {
+
+ cm, err := readCompactMap(baseFileName)
+ if err != nil {
+ return fmt.Errorf("readCompactMap: %v", err)
+ }
+
+ ecxFile, err := os.OpenFile(baseFileName+".ecx", os.O_TRUNC|os.O_CREATE|os.O_WRONLY, 0644)
+ if err != nil {
+ return fmt.Errorf("failed to open dat file: %v", err)
+ }
+ defer ecxFile.Close()
+
+ err = cm.AscendingVisit(func(value needle_map.NeedleValue) error {
+ bytes := value.ToBytes()
+ _, writeErr := ecxFile.Write(bytes)
+ return writeErr
+ })
+
+ if err != nil {
+ return fmt.Errorf("failed to open dat file: %v", err)
+ }
+
+ return nil
+}
+
+// WriteEcFiles generates .ec01 ~ .ec14 files
+func WriteEcFiles(baseFileName string) error {
+ return generateEcFiles(baseFileName, 256*1024, ErasureCodingLargeBlockSize, ErasureCodingSmallBlockSize)
+}
+
+func ToExt(ecIndex int) string {
+ return fmt.Sprintf(".ec%02d", ecIndex)
+}
+
+func generateEcFiles(baseFileName string, bufferSize int, largeBlockSize int64, smallBlockSize int64) error {
+ file, err := os.OpenFile(baseFileName+".dat", os.O_RDONLY, 0)
+ if err != nil {
+ return fmt.Errorf("failed to open dat file: %v", err)
+ }
+ defer file.Close()
+
+ fi, err := file.Stat()
+ if err != nil {
+ return fmt.Errorf("failed to stat dat file: %v", err)
+ }
+ err = encodeDatFile(fi.Size(), err, baseFileName, bufferSize, largeBlockSize, file, smallBlockSize)
+ if err != nil {
+ return fmt.Errorf("encodeDatFile: %v", err)
+ }
+ return nil
+}
+
func encodeData(file *os.File, enc reedsolomon.Encoder, startOffset, blockSize int64, buffers [][]byte, outputs []*os.File) error {
bufferSize := int64(len(buffers[0]))
- batchCount := blockSize/bufferSize
- if blockSize%bufferSize!=0 {
+ batchCount := blockSize / bufferSize
+ if blockSize%bufferSize != 0 {
glog.Fatalf("unexpected block size %d buffer size %d", blockSize, bufferSize)
}
@@ -34,10 +92,10 @@ func encodeData(file *os.File, enc reedsolomon.Encoder, startOffset, blockSize i
return nil
}
-func openEcFiles(baseFileName string, forRead bool) (files []*os.File, err error){
- for i := 0; i< DataShardsCount+ParityShardsCount; i++{
- fname := fmt.Sprintf("%s.ec%02d", baseFileName, i+1)
- openOption := os.O_TRUNC|os.O_CREATE|os.O_WRONLY
+func openEcFiles(baseFileName string, forRead bool) (files []*os.File, err error) {
+ for i := 0; i < DataShardsCount+ParityShardsCount; i++ {
+ fname := baseFileName + ToExt(i+1)
+ openOption := os.O_TRUNC | os.O_CREATE | os.O_WRONLY
if forRead {
openOption = os.O_RDONLY
}
@@ -50,15 +108,14 @@ func openEcFiles(baseFileName string, forRead bool) (files []*os.File, err error
return
}
-func closeEcFiles(files []*os.File){
- for _, f := range files{
+func closeEcFiles(files []*os.File) {
+ for _, f := range files {
if f != nil {
f.Close()
}
}
}
-
func encodeDataOneBatch(file *os.File, enc reedsolomon.Encoder, startOffset, blockSize int64, buffers [][]byte, outputs []*os.File) error {
// read data into buffers
@@ -90,3 +147,56 @@ func encodeDataOneBatch(file *os.File, enc reedsolomon.Encoder, startOffset, blo
return nil
}
+
+func encodeDatFile(remainingSize int64, err error, baseFileName string, bufferSize int, largeBlockSize int64, file *os.File, smallBlockSize int64) error {
+ var processedSize int64
+ enc, err := reedsolomon.New(DataShardsCount, ParityShardsCount)
+ if err != nil {
+ return fmt.Errorf("failed to create encoder: %v", err)
+ }
+ buffers := make([][]byte, DataShardsCount+ParityShardsCount)
+ outputs, err := openEcFiles(baseFileName, false)
+ defer closeEcFiles(outputs)
+ if err != nil {
+ return fmt.Errorf("failed to open dat file: %v", err)
+ }
+ for i, _ := range buffers {
+ buffers[i] = make([]byte, bufferSize)
+ }
+ for remainingSize > largeBlockSize*DataShardsCount {
+ err = encodeData(file, enc, processedSize, largeBlockSize, buffers, outputs)
+ if err != nil {
+ return fmt.Errorf("failed to encode large chunk data: %v", err)
+ }
+ remainingSize -= largeBlockSize * DataShardsCount
+ processedSize += largeBlockSize * DataShardsCount
+ }
+ for remainingSize > 0 {
+ encodeData(file, enc, processedSize, smallBlockSize, buffers, outputs)
+ if err != nil {
+ return fmt.Errorf("failed to encode small chunk data: %v", err)
+ }
+ remainingSize -= smallBlockSize * DataShardsCount
+ processedSize += smallBlockSize * DataShardsCount
+ }
+ return nil
+}
+
+func readCompactMap(baseFileName string) (*needle_map.CompactMap, error) {
+ indexFile, err := os.OpenFile(baseFileName+".idx", os.O_RDONLY, 0644)
+ if err != nil {
+ return nil, fmt.Errorf("cannot read Volume Index %s.idx: %v", baseFileName, err)
+ }
+ defer indexFile.Close()
+
+ cm := needle_map.NewCompactMap()
+ err = storage.WalkIndexFile(indexFile, func(key types.NeedleId, offset types.Offset, size uint32) error {
+ if !offset.IsZero() && size != types.TombstoneFileSize {
+ cm.Set(key, offset, size)
+ } else {
+ cm.Delete(key)
+ }
+ return nil
+ })
+ return cm, err
+}
diff --git a/weed/storage/erasure_coding/ec_test.go b/weed/storage/erasure_coding/ec_test.go
index d631471e9..625f4e9a6 100644
--- a/weed/storage/erasure_coding/ec_test.go
+++ b/weed/storage/erasure_coding/ec_test.go
@@ -7,7 +7,6 @@ import (
"os"
"testing"
- "github.com/chrislusf/seaweedfs/weed/storage"
"github.com/chrislusf/seaweedfs/weed/storage/needle_map"
"github.com/chrislusf/seaweedfs/weed/storage/types"
"github.com/klauspost/reedsolomon"
@@ -27,98 +26,20 @@ func TestEncodingDecoding(t *testing.T) {
t.Logf("generateEcFiles: %v", err)
}
- err = writeSortedEcxFiles(baseFileName)
+ err = WriteSortedEcxFile(baseFileName)
if err != nil {
- t.Logf("writeSortedEcxFiles: %v", err)
+ t.Logf("WriteSortedEcxFile: %v", err)
}
err = validateFiles(baseFileName)
if err != nil {
- t.Logf("writeSortedEcxFiles: %v", err)
+ t.Logf("WriteSortedEcxFile: %v", err)
}
removeGeneratedFiles(baseFileName)
}
-func generateEcFiles(baseFileName string, bufferSize int, largeBlockSize int64, smallBlockSize int64) error {
- file, err := os.OpenFile(baseFileName+".dat", os.O_RDONLY, 0)
- if err != nil {
- return fmt.Errorf("failed to open dat file: %v", err)
- }
- defer file.Close()
-
- fi, err := file.Stat()
- if err != nil {
- return fmt.Errorf("failed to stat dat file: %v", err)
- }
- err = encodeDatFile(fi.Size(), err, baseFileName, bufferSize, largeBlockSize, file, smallBlockSize)
- if err != nil {
- return fmt.Errorf("encodeDatFile: %v", err)
- }
- return nil
-}
-
-func encodeDatFile(remainingSize int64, err error, baseFileName string, bufferSize int, largeBlockSize int64, file *os.File, smallBlockSize int64) error {
- var processedSize int64
- enc, err := reedsolomon.New(DataShardsCount, ParityShardsCount)
- if err != nil {
- return fmt.Errorf("failed to create encoder: %v", err)
- }
- buffers := make([][]byte, DataShardsCount+ParityShardsCount)
- outputs, err := openEcFiles(baseFileName, false)
- defer closeEcFiles(outputs)
- if err != nil {
- return fmt.Errorf("failed to open dat file: %v", err)
- }
- for i, _ := range buffers {
- buffers[i] = make([]byte, bufferSize)
- }
- for remainingSize > largeBlockSize*DataShardsCount {
- err = encodeData(file, enc, processedSize, largeBlockSize, buffers, outputs)
- if err != nil {
- return fmt.Errorf("failed to encode large chunk data: %v", err)
- }
- remainingSize -= largeBlockSize * DataShardsCount
- processedSize += largeBlockSize * DataShardsCount
- }
- for remainingSize > 0 {
- encodeData(file, enc, processedSize, smallBlockSize, buffers, outputs)
- if err != nil {
- return fmt.Errorf("failed to encode small chunk data: %v", err)
- }
- remainingSize -= smallBlockSize * DataShardsCount
- processedSize += smallBlockSize * DataShardsCount
- }
- return nil
-}
-
-func writeSortedEcxFiles(baseFileName string) (e error) {
-
- cm, err := readCompactMap(baseFileName)
- if err != nil {
- return fmt.Errorf("readCompactMap: %v", err)
- }
-
- ecxFile, err := os.OpenFile(baseFileName+".ecx", os.O_TRUNC|os.O_CREATE|os.O_WRONLY, 0644)
- if err != nil {
- return fmt.Errorf("failed to open dat file: %v", err)
- }
- defer ecxFile.Close()
-
- err = cm.AscendingVisit(func(value needle_map.NeedleValue) error {
- bytes := value.ToBytes()
- _, writeErr := ecxFile.Write(bytes)
- return writeErr
- })
-
- if err != nil {
- return fmt.Errorf("failed to open dat file: %v", err)
- }
-
- return nil
-}
-
func validateFiles(baseFileName string) error {
cm, err := readCompactMap(baseFileName)
if err != nil {
@@ -148,25 +69,6 @@ func validateFiles(baseFileName string) error {
return nil
}
-func readCompactMap(baseFileName string) (*needle_map.CompactMap, error) {
- indexFile, err := os.OpenFile(baseFileName+".idx", os.O_RDONLY, 0644)
- if err != nil {
- return nil, fmt.Errorf("cannot read Volume Index %s.idx: %v", baseFileName, err)
- }
- defer indexFile.Close()
-
- cm := needle_map.NewCompactMap()
- err = storage.WalkIndexFile(indexFile, func(key types.NeedleId, offset types.Offset, size uint32) error {
- if !offset.IsZero() && size != types.TombstoneFileSize {
- cm.Set(key, offset, size)
- } else {
- cm.Delete(key)
- }
- return nil
- })
- return cm, err
-}
-
func assertSame(datFile *os.File, datSize int64, ecFiles []*os.File, offset types.Offset, size uint32) error {
data, err := readDatFile(datFile, offset, size)
@@ -288,7 +190,7 @@ func removeGeneratedFiles(baseFileName string) {
fname := fmt.Sprintf("%s.ec%02d", baseFileName, i+1)
os.Remove(fname)
}
- os.Remove(baseFileName+".ecx")
+ os.Remove(baseFileName + ".ecx")
}
func TestLocateData(t *testing.T) {
generated by cgit v1.2.3 (git 2.25.1) at 2025-12-25 06:05:55 +0000