Fix healing whole file bitrot (#7123)

* Use 0-byte file for bitrot verification of whole-file-bitrot files

Also pass the right checksum information for bitrot verification

* Copy xlMeta info from latest meta except []checksums and []Parts while healing

cmd/bitrot.go

@@ -158,12 +158,13 @@ func bitrotWriterSum(w io.Writer) []byte {
 
 // Verify if a file has bitrot error.
 func bitrotCheckFile(disk StorageAPI, volume string, filePath string, tillOffset int64, algo BitrotAlgorithm, sum []byte, shardSize int64) (err error) {
-	buf := make([]byte, shardSize)
 	if algo != HighwayHash256S {
-		// For whole-file bitrot we don't need to read the entire file as the bitrot verify happens on the server side even if we read small buffer
+		buf := []byte{}
+		// For whole-file bitrot we don't need to read the entire file as the bitrot verify happens on the server side even if we read 0-bytes.
 		_, err = disk.ReadFile(volume, filePath, 0, buf, NewBitrotVerifier(algo, sum))
 		return err
 	}
+	buf := make([]byte, shardSize)
 	r := newStreamingBitrotReader(disk, volume, filePath, tillOffset, algo, shardSize)
 	defer closeBitrotReaders([]io.ReaderAt{r})
 	var offset int64

cmd/bitrot_test.go

@@ -24,7 +24,7 @@ import (
 	"testing"
 )
 
-func TestBitrotReaderWriter(t *testing.T) {
+func testBitrotReaderWriterAlgo(t *testing.T, bitrotAlgo BitrotAlgorithm) {
 	tmpDir, err := ioutil.TempDir("", "")
 	if err != nil {
 		log.Fatal(err)
@@ -41,7 +41,7 @@ func TestBitrotReaderWriter(t *testing.T) {
 
 	disk.MakeVol(volume)
 
-	writer := newBitrotWriter(disk, volume, filePath, 35, HighwayHash256S, 10)
+	writer := newBitrotWriter(disk, volume, filePath, 35, bitrotAlgo, 10)
 
 	_, err = writer.Write([]byte("aaaaaaaaaa"))
 	if err != nil {
@@ -61,7 +61,7 @@ func TestBitrotReaderWriter(t *testing.T) {
 	}
 	writer.(io.Closer).Close()
 
-	reader := newStreamingBitrotReader(disk, volume, filePath, 35, HighwayHash256S, 10)
+	reader := newBitrotReader(disk, volume, filePath, 35, bitrotAlgo, bitrotWriterSum(writer), 10)
 	b := make([]byte, 10)
 	if _, err = reader.ReadAt(b, 0); err != nil {
 		log.Fatal(err)
@@ -76,3 +76,9 @@ func TestBitrotReaderWriter(t *testing.T) {
 		log.Fatal(err)
 	}
 }
+
+func TestAllBitrotAlgorithms(t *testing.T) {
+	for bitrotAlgo := range bitrotAlgorithms {
+		testBitrotReaderWriterAlgo(t, bitrotAlgo)
+	}
+}

cmd/xl-v1-healing.go

@@ -251,7 +251,6 @@ func listAllBuckets(storageDisks []StorageAPI) (buckets map[string]VolInfo,
 // Heals an object by re-writing corrupt/missing erasure blocks.
 func healObject(ctx context.Context, storageDisks []StorageAPI, bucket string, object string,
 	quorum int, dryRun bool) (result madmin.HealResultItem, err error) {
-
 	partsMetadata, errs := readAllXLMetadata(ctx, storageDisks, bucket, object)
 
 	errCount := 0
@@ -432,20 +431,21 @@ func healObject(ctx context.Context, storageDisks []StorageAPI, bucket string, o
 		partActualSize := latestMeta.Parts[partIndex].ActualSize
 		partNumber := latestMeta.Parts[partIndex].Number
 		tillOffset := erasure.ShardFileTillOffset(0, partSize, partSize)
-		checksumInfo := erasureInfo.GetChecksumInfo(partName)
 		readers := make([]io.ReaderAt, len(latestDisks))
+		checksumAlgo := erasureInfo.GetChecksumInfo(partName).Algorithm
 		for i, disk := range latestDisks {
 			if disk == OfflineDisk {
 				continue
 			}
-			readers[i] = newBitrotReader(disk, bucket, pathJoin(object, partName), tillOffset, checksumInfo.Algorithm, checksumInfo.Hash, erasure.ShardSize())
+			checksumInfo := partsMetadata[i].Erasure.GetChecksumInfo(partName)
+			readers[i] = newBitrotReader(disk, bucket, pathJoin(object, partName), tillOffset, checksumAlgo, checksumInfo.Hash, erasure.ShardSize())
 		}
 		writers := make([]io.Writer, len(outDatedDisks))
 		for i, disk := range outDatedDisks {
 			if disk == OfflineDisk {
 				continue
 			}
-			writers[i] = newBitrotWriter(disk, minioMetaTmpBucket, pathJoin(tmpID, partName), tillOffset, checksumInfo.Algorithm, erasure.ShardSize())
+			writers[i] = newBitrotWriter(disk, minioMetaTmpBucket, pathJoin(tmpID, partName), tillOffset, checksumAlgo, erasure.ShardSize())
 		}
 		hErr := erasure.Heal(ctx, readers, writers, partSize)
 		closeBitrotReaders(readers)
@@ -467,7 +467,7 @@ func healObject(ctx context.Context, storageDisks []StorageAPI, bucket string, o
 				continue
 			}
 			partsMetadata[i].AddObjectPart(partNumber, partName, "", partSize, partActualSize)
-			partsMetadata[i].Erasure.AddChecksumInfo(ChecksumInfo{partName, checksumInfo.Algorithm, bitrotWriterSum(writers[i])})
+			partsMetadata[i].Erasure.AddChecksumInfo(ChecksumInfo{partName, checksumAlgo, bitrotWriterSum(writers[i])})
 		}
 
 		// If all disks are having errors, we give up.
@@ -476,14 +476,6 @@ func healObject(ctx context.Context, storageDisks []StorageAPI, bucket string, o
 		}
 	}
 
-	// xl.json should be written to all the healed disks.
-	for index, disk := range outDatedDisks {
-		if disk == nil {
-			continue
-		}
-		partsMetadata[index] = latestMeta
-	}
-
 	// Generate and write `xl.json` generated from other disks.
 	outDatedDisks, aErr := writeUniqueXLMetadata(ctx, outDatedDisks, minioMetaTmpBucket, tmpID,
 		partsMetadata, diskCount(outDatedDisks))

cmd/xl-v1-metadata.go

@@ -189,17 +189,9 @@ func newXLMetaV1(object string, dataBlocks, parityBlocks int) (xlMeta xlMetaV1)
 // Return a new xlMetaV1 initialized using the given xlMetaV1. Used in healing to make sure that we do not copy
 // over any part's checksum info which will differ for different disks.
 func newXLMetaFromXLMeta(meta xlMetaV1) xlMetaV1 {
-	xlMeta := xlMetaV1{}
-	xlMeta.Version = xlMetaVersion
-	xlMeta.Format = xlMetaFormat
-	xlMeta.Minio.Release = ReleaseTag
-	xlMeta.Erasure = ErasureInfo{
-		Algorithm:    meta.Erasure.Algorithm,
-		DataBlocks:   meta.Erasure.DataBlocks,
-		ParityBlocks: meta.Erasure.DataBlocks,
-		BlockSize:    meta.Erasure.BlockSize,
-		Distribution: meta.Erasure.Distribution,
-	}
+	xlMeta := meta
+	xlMeta.Erasure.Checksums = nil
+	xlMeta.Parts = nil
 	return xlMeta
 }