You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

222 lines
6.1 KiB

  1. package erasure_coding
  2. import (
  3. "bytes"
  4. "fmt"
  5. "os"
  6. "testing"
  7. "github.com/chrislusf/seaweedfs/weed/storage"
  8. "github.com/chrislusf/seaweedfs/weed/storage/needle_map"
  9. "github.com/chrislusf/seaweedfs/weed/storage/types"
  10. "github.com/klauspost/reedsolomon"
  11. )
  12. const (
  13. largeBlockSize = 10000
  14. smallBlockSize = 100
  15. )
  16. func TestEncodingDecoding(t *testing.T) {
  17. bufferSize := 50
  18. baseFileName := "1"
  19. err := generateEcFiles(baseFileName, bufferSize, largeBlockSize, smallBlockSize)
  20. if err != nil {
  21. t.Logf("generateEcFiles: %v", err)
  22. }
  23. err = writeSortedEcxFiles(baseFileName)
  24. if err != nil {
  25. t.Logf("writeSortedEcxFiles: %v", err)
  26. }
  27. err = validateFiles(baseFileName)
  28. if err != nil {
  29. t.Logf("writeSortedEcxFiles: %v", err)
  30. }
  31. }
  32. func generateEcFiles(baseFileName string, bufferSize int, largeBlockSize int64, smallBlockSize int64) error {
  33. file, err := os.OpenFile(baseFileName+".dat", os.O_RDONLY, 0)
  34. if err != nil {
  35. return fmt.Errorf("failed to open dat file: %v", err)
  36. }
  37. defer file.Close()
  38. fi, err := file.Stat()
  39. if err != nil {
  40. return fmt.Errorf("failed to stat dat file: %v", err)
  41. }
  42. err = encodeDatFile(fi.Size(), err, baseFileName, bufferSize, largeBlockSize, file, smallBlockSize)
  43. if err != nil {
  44. return fmt.Errorf("encodeDatFile: %v", err)
  45. }
  46. return nil
  47. }
  48. func encodeDatFile(remainingSize int64, err error, baseFileName string, bufferSize int, largeBlockSize int64, file *os.File, smallBlockSize int64) error {
  49. var processedSize int64
  50. enc, err := reedsolomon.New(DataShardsCount, ParityShardsCount)
  51. if err != nil {
  52. return fmt.Errorf("failed to create encoder: %v", err)
  53. }
  54. buffers := make([][]byte, DataShardsCount+ParityShardsCount)
  55. outputs, err := openEcFiles(baseFileName, false)
  56. defer closeEcFiles(outputs)
  57. if err != nil {
  58. return fmt.Errorf("failed to open dat file: %v", err)
  59. }
  60. for i, _ := range buffers {
  61. buffers[i] = make([]byte, bufferSize)
  62. }
  63. for remainingSize > largeBlockSize*DataShardsCount {
  64. err = encodeData(file, enc, processedSize, largeBlockSize, buffers, outputs)
  65. if err != nil {
  66. return fmt.Errorf("failed to encode large chunk data: %v", err)
  67. }
  68. remainingSize -= largeBlockSize * DataShardsCount
  69. processedSize += largeBlockSize * DataShardsCount
  70. }
  71. for remainingSize > 0 {
  72. encodeData(file, enc, processedSize, smallBlockSize, buffers, outputs)
  73. if err != nil {
  74. return fmt.Errorf("failed to encode small chunk data: %v", err)
  75. }
  76. remainingSize -= smallBlockSize * DataShardsCount
  77. processedSize += smallBlockSize * DataShardsCount
  78. }
  79. return nil
  80. }
  81. func writeSortedEcxFiles(baseFileName string) (e error) {
  82. cm, err := readCompactMap(baseFileName)
  83. if err != nil {
  84. return fmt.Errorf("readCompactMap: %v", err)
  85. }
  86. ecxFile, err := os.OpenFile(baseFileName+".ecx", os.O_TRUNC|os.O_CREATE|os.O_WRONLY, 0644)
  87. if err != nil {
  88. return fmt.Errorf("failed to open dat file: %v", err)
  89. }
  90. defer ecxFile.Close()
  91. err = cm.AscendingVisit(func(value needle_map.NeedleValue) error {
  92. bytes := value.ToBytes()
  93. _, writeErr := ecxFile.Write(bytes)
  94. return writeErr
  95. })
  96. if err != nil {
  97. return fmt.Errorf("failed to open dat file: %v", err)
  98. }
  99. return nil
  100. }
  101. func validateFiles(baseFileName string) error {
  102. cm, err := readCompactMap(baseFileName)
  103. if err != nil {
  104. return fmt.Errorf("readCompactMap: %v", err)
  105. }
  106. datFile, err := os.OpenFile(baseFileName+".dat", os.O_RDONLY, 0)
  107. if err != nil {
  108. return fmt.Errorf("failed to open dat file: %v", err)
  109. }
  110. defer datFile.Close()
  111. fi, err := datFile.Stat()
  112. if err != nil {
  113. return fmt.Errorf("failed to stat dat file: %v", err)
  114. }
  115. ecFiles, err := openEcFiles(baseFileName, true)
  116. defer closeEcFiles(ecFiles)
  117. err = cm.AscendingVisit(func(value needle_map.NeedleValue) error {
  118. return assertSame(datFile, fi.Size(), ecFiles, value.Offset, value.Size)
  119. })
  120. if err != nil {
  121. return fmt.Errorf("failed to check ec files: %v", err)
  122. }
  123. return nil
  124. }
  125. func readCompactMap(baseFileName string) (*needle_map.CompactMap, error) {
  126. indexFile, err := os.OpenFile(baseFileName+".idx", os.O_RDONLY, 0644)
  127. if err != nil {
  128. return nil, fmt.Errorf("cannot read Volume Index %s.idx: %v", baseFileName, err)
  129. }
  130. defer indexFile.Close()
  131. cm := needle_map.NewCompactMap()
  132. err = storage.WalkIndexFile(indexFile, func(key types.NeedleId, offset types.Offset, size uint32) error {
  133. if !offset.IsZero() && size != types.TombstoneFileSize {
  134. cm.Set(key, offset, size)
  135. } else {
  136. cm.Delete(key)
  137. }
  138. return nil
  139. })
  140. return cm, err
  141. }
  142. func assertSame(datFile *os.File, datSize int64, ecFiles []*os.File, offset types.Offset, size uint32) error {
  143. data, err := readDatFile(datFile, offset, size)
  144. if err != nil {
  145. return fmt.Errorf("failed to read dat file: %v", err)
  146. }
  147. ecData, err := readEcFile(datSize, ecFiles, offset, size)
  148. if err != nil {
  149. return fmt.Errorf("failed to read ec file: %v", err)
  150. }
  151. if bytes.Compare(data, ecData) != 0 {
  152. return fmt.Errorf("unexpected data read")
  153. }
  154. return nil
  155. }
  156. func readDatFile(datFile *os.File, offset types.Offset, size uint32) ([]byte, error) {
  157. data := make([]byte, size)
  158. n, err := datFile.ReadAt(data, offset.ToAcutalOffset())
  159. if err != nil {
  160. return nil, fmt.Errorf("failed to ReadAt dat file: %v", err)
  161. }
  162. if n != int(size) {
  163. return nil, fmt.Errorf("unexpected read size %d, expected %d", n, size)
  164. }
  165. return data, nil
  166. }
  167. func readEcFile(datSize int64, ecFiles []*os.File, offset types.Offset, size uint32) ([]byte, error) {
  168. return nil, nil
  169. }
  170. func TestLocateData(t *testing.T) {
  171. intervals := locateData(largeBlockSize, smallBlockSize, DataShardsCount*largeBlockSize+1, DataShardsCount*largeBlockSize, 1)
  172. if len(intervals) != 1 {
  173. t.Errorf("unexpected interval size %d", len(intervals))
  174. }
  175. if !intervals[0].sameAs(Interval{0, 0, 1, false}) {
  176. t.Errorf("unexpected interval %+v", intervals[0])
  177. }
  178. intervals = locateData(largeBlockSize, smallBlockSize, DataShardsCount*largeBlockSize+1, DataShardsCount*largeBlockSize/2+100, DataShardsCount*largeBlockSize+1 - DataShardsCount*largeBlockSize/2-100)
  179. fmt.Printf("%+v\n", intervals)
  180. }
  181. func (this Interval) sameAs(that Interval) bool {
  182. return this.isLargeBlock == that.isLargeBlock &&
  183. this.innerBlockOffset == that.innerBlockOffset &&
  184. this.blockIndex == that.blockIndex &&
  185. this.size == that.size
  186. }