fix: readData in bulk call using messagepack byte wrappers (#11228)

This PR refactors the way we use buffers for O_DIRECT and
to re-use those buffers for messagepack reader writer.

After some extensive benchmarking found that not all objects
have this benefit, and only objects smaller than 64KiB see
this benefit overall.

Benefits are seen from almost all objects from

1KiB - 32KiB

Beyond this no objects see benefit with bulk call approach
as the latency of bytes sent over the wire v/s streaming
content directly from disk negate each other with no
remarkable benefits.

All other optimizations include reuse of msgp.Reader,
msgp.Writer using sync.Pool's for all internode calls.

cmd/bitrot-streaming.go

@@ -91,6 +91,7 @@ func newStreamingBitrotWriter(disk StorageAPI, volume, filePath string, length i
 // ReadAt() implementation which verifies the bitrot hash available as part of the stream.
 type streamingBitrotReader struct {
 	disk       StorageAPI
+	data       []byte
 	rc         io.Reader
 	volume     string
 	filePath   string
@@ -122,7 +123,11 @@ func (b *streamingBitrotReader) ReadAt(buf []byte, offset int64) (int, error) {
 		// For the first ReadAt() call we need to open the stream for reading.
 		b.currOffset = offset
 		streamOffset := (offset/b.shardSize)*int64(b.h.Size()) + offset
+		if len(b.data) == 0 {
 			b.rc, err = b.disk.ReadFileStream(context.TODO(), b.volume, b.filePath, streamOffset, b.tillOffset-streamOffset)
+		} else {
+			b.rc = io.NewSectionReader(bytes.NewReader(b.data), streamOffset, b.tillOffset-streamOffset)
+		}
 		if err != nil {
 			return 0, err
 		}
@@ -154,10 +159,11 @@ func (b *streamingBitrotReader) ReadAt(buf []byte, offset int64) (int, error) {
 }
 
 // Returns streaming bitrot reader implementation.
-func newStreamingBitrotReader(disk StorageAPI, volume, filePath string, tillOffset int64, algo BitrotAlgorithm, shardSize int64) *streamingBitrotReader {
+func newStreamingBitrotReader(disk StorageAPI, data []byte, volume, filePath string, tillOffset int64, algo BitrotAlgorithm, shardSize int64) *streamingBitrotReader {
 	h := algo.New()
 	return &streamingBitrotReader{
 		disk,
+		data,
 		nil,
 		volume,
 		filePath,

cmd/bitrot.go

@@ -103,9 +103,9 @@ func newBitrotWriter(disk StorageAPI, volume, filePath string, length int64, alg
 	return newWholeBitrotWriter(disk, volume, filePath, algo, shardSize)
 }
 
-func newBitrotReader(disk StorageAPI, bucket string, filePath string, tillOffset int64, algo BitrotAlgorithm, sum []byte, shardSize int64) io.ReaderAt {
+func newBitrotReader(disk StorageAPI, data []byte, bucket string, filePath string, tillOffset int64, algo BitrotAlgorithm, sum []byte, shardSize int64) io.ReaderAt {
 	if algo == HighwayHash256S {
-		return newStreamingBitrotReader(disk, bucket, filePath, tillOffset, algo, shardSize)
+		return newStreamingBitrotReader(disk, data, bucket, filePath, tillOffset, algo, shardSize)
 	}
 	return newWholeBitrotReader(disk, bucket, filePath, algo, tillOffset, sum)
 }

cmd/bitrot_test.go

@@ -62,7 +62,7 @@ func testBitrotReaderWriterAlgo(t *testing.T, bitrotAlgo BitrotAlgorithm) {
 	}
 	writer.(io.Closer).Close()
 
-	reader := newBitrotReader(disk, volume, filePath, 35, bitrotAlgo, bitrotWriterSum(writer), 10)
+	reader := newBitrotReader(disk, nil, volume, filePath, 35, bitrotAlgo, bitrotWriterSum(writer), 10)
 	b := make([]byte, 10)
 	if _, err = reader.ReadAt(b, 0); err != nil {
 		log.Fatal(err)

cmd/erasure-decode_test.go

@@ -134,7 +134,7 @@ func TestErasureDecode(t *testing.T) {
 			}
 			tillOffset := erasure.ShardFileOffset(test.offset, test.length, test.data)
 
-			bitrotReaders[index] = newBitrotReader(disk, "testbucket", "object", tillOffset, writeAlgorithm, bitrotWriterSum(writers[index]), erasure.ShardSize())
+			bitrotReaders[index] = newBitrotReader(disk, nil, "testbucket", "object", tillOffset, writeAlgorithm, bitrotWriterSum(writers[index]), erasure.ShardSize())
 		}
 
 		writer := bytes.NewBuffer(nil)
@@ -164,7 +164,7 @@ func TestErasureDecode(t *testing.T) {
 					continue
 				}
 				tillOffset := erasure.ShardFileOffset(test.offset, test.length, test.data)
-				bitrotReaders[index] = newBitrotReader(disk, "testbucket", "object", tillOffset, writeAlgorithm, bitrotWriterSum(writers[index]), erasure.ShardSize())
+				bitrotReaders[index] = newBitrotReader(disk, nil, "testbucket", "object", tillOffset, writeAlgorithm, bitrotWriterSum(writers[index]), erasure.ShardSize())
 			}
 			for j := range disks[:test.offDisks] {
 				if bitrotReaders[j] == nil {
@@ -270,7 +270,7 @@ func TestErasureDecodeRandomOffsetLength(t *testing.T) {
 				continue
 			}
 			tillOffset := erasure.ShardFileOffset(offset, readLen, length)
-			bitrotReaders[index] = newStreamingBitrotReader(disk, "testbucket", "object", tillOffset, DefaultBitrotAlgorithm, erasure.ShardSize())
+			bitrotReaders[index] = newStreamingBitrotReader(disk, nil, "testbucket", "object", tillOffset, DefaultBitrotAlgorithm, erasure.ShardSize())
 		}
 		err = erasure.Decode(context.Background(), buf, bitrotReaders, offset, readLen, length, nil)
 		closeBitrotReaders(bitrotReaders)
@@ -332,7 +332,7 @@ func benchmarkErasureDecode(data, parity, dataDown, parityDown int, size int64,
 				continue
 			}
 			tillOffset := erasure.ShardFileOffset(0, size, size)
-			bitrotReaders[index] = newStreamingBitrotReader(disk, "testbucket", "object", tillOffset, DefaultBitrotAlgorithm, erasure.ShardSize())
+			bitrotReaders[index] = newStreamingBitrotReader(disk, nil, "testbucket", "object", tillOffset, DefaultBitrotAlgorithm, erasure.ShardSize())
 		}
 		if err = erasure.Decode(context.Background(), bytes.NewBuffer(content[:0]), bitrotReaders, 0, size, size, nil); err != nil {
 			panic(err)

cmd/erasure-heal_test.go

@@ -99,7 +99,7 @@ func TestErasureHeal(t *testing.T) {
 		readers := make([]io.ReaderAt, len(disks))
 		for i, disk := range disks {
 			shardFilesize := erasure.ShardFileSize(test.size)
-			readers[i] = newBitrotReader(disk, "testbucket", "testobject", shardFilesize, test.algorithm, bitrotWriterSum(writers[i]), erasure.ShardSize())
+			readers[i] = newBitrotReader(disk, nil, "testbucket", "testobject", shardFilesize, test.algorithm, bitrotWriterSum(writers[i]), erasure.ShardSize())
 		}
 
 		// setup stale disks for the test case

cmd/erasure-healing-common_test.go

@@ -186,7 +186,7 @@ func TestListOnlineDisks(t *testing.T) {
 			t.Fatalf("Failed to putObject %v", err)
 		}
 
-		partsMetadata, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "")
+		partsMetadata, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
 		fi, err := getLatestFileInfo(ctx, partsMetadata, errs)
 		if err != nil {
 			t.Fatalf("Failed to getLatestFileInfo %v", err)
@@ -287,7 +287,7 @@ func TestDisksWithAllParts(t *testing.T) {
 		t.Fatalf("Failed to putObject %v", err)
 	}
 
-	_, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "")
+	_, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
 	readQuorum := len(erasureDisks) / 2
 	if reducedErr := reduceReadQuorumErrs(ctx, errs, objectOpIgnoredErrs, readQuorum); reducedErr != nil {
 		t.Fatalf("Failed to read xl meta data %v", reducedErr)
@@ -295,7 +295,7 @@ func TestDisksWithAllParts(t *testing.T) {
 
 	// Test that all disks are returned without any failures with
 	// unmodified meta data
-	partsMetadata, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "")
+	partsMetadata, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
 	if err != nil {
 		t.Fatalf("Failed to read xl meta data %v", err)
 	}

cmd/erasure-healing.go

@@ -395,7 +395,7 @@ func (er erasureObjects) healObject(ctx context.Context, bucket string, object s
 				if latestMeta.XLV1 {
 					partPath = pathJoin(object, fmt.Sprintf("part.%d", partNumber))
 				}
-				readers[i] = newBitrotReader(disk, bucket, partPath, tillOffset, checksumAlgo, checksumInfo.Hash, erasure.ShardSize())
+				readers[i] = newBitrotReader(disk, nil, bucket, partPath, tillOffset, checksumAlgo, checksumInfo.Hash, erasure.ShardSize())
 			}
 			writers := make([]io.Writer, len(outDatedDisks))
 			for i, disk := range outDatedDisks {
@@ -811,7 +811,7 @@ func (er erasureObjects) HealObject(ctx context.Context, bucket, object, version
 	storageEndpoints := er.getEndpoints()
 
 	// Read metadata files from all the disks
-	partsMetadata, errs := readAllFileInfo(healCtx, storageDisks, bucket, object, versionID)
+	partsMetadata, errs := readAllFileInfo(healCtx, storageDisks, bucket, object, versionID, false)
 
 	if isAllNotFound(errs) {
 		err = toObjectErr(errFileNotFound, bucket, object)

cmd/erasure-healing_test.go

@@ -67,7 +67,7 @@ func TestHealing(t *testing.T) {
 	}
 
 	disk := er.getDisks()[0]
-	fileInfoPreHeal, err := disk.ReadVersion(context.Background(), bucket, object, "")
+	fileInfoPreHeal, err := disk.ReadVersion(context.Background(), bucket, object, "", false)
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -84,7 +84,7 @@ func TestHealing(t *testing.T) {
 		t.Fatal(err)
 	}
 
-	fileInfoPostHeal, err := disk.ReadVersion(context.Background(), bucket, object, "")
+	fileInfoPostHeal, err := disk.ReadVersion(context.Background(), bucket, object, "", false)
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -113,7 +113,7 @@ func TestHealing(t *testing.T) {
 		t.Fatal(err)
 	}
 
-	fileInfoPostHeal, err = disk.ReadVersion(context.Background(), bucket, object, "")
+	fileInfoPostHeal, err = disk.ReadVersion(context.Background(), bucket, object, "", false)
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -214,7 +214,7 @@ func TestHealObjectCorrupted(t *testing.T) {
 		t.Fatalf("Failed to heal object - %v", err)
 	}
 
-	fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "")
+	fileInfos, errs := readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
 	fi, err := getLatestFileInfo(ctx, fileInfos, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
@@ -239,7 +239,7 @@ func TestHealObjectCorrupted(t *testing.T) {
 		t.Errorf("Expected nil but received %v", err)
 	}
 
-	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "")
+	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
 	nfi, err := getLatestFileInfo(ctx, fileInfos, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)
@@ -265,7 +265,7 @@ func TestHealObjectCorrupted(t *testing.T) {
 		t.Errorf("Expected nil but received %v", err)
 	}
 
-	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "")
+	fileInfos, errs = readAllFileInfo(ctx, erasureDisks, bucket, object, "", false)
 	nfi, err = getLatestFileInfo(ctx, fileInfos, errs)
 	if err != nil {
 		t.Fatalf("Failed to getLatestFileInfo - %v", err)

cmd/erasure-metadata-utils.go

@@ -115,7 +115,7 @@ func hashOrder(key string, cardinality int) []int {
 
 // Reads all `xl.meta` metadata as a FileInfo slice.
 // Returns error slice indicating the failed metadata reads.
-func readAllFileInfo(ctx context.Context, disks []StorageAPI, bucket, object, versionID string) ([]FileInfo, []error) {
+func readAllFileInfo(ctx context.Context, disks []StorageAPI, bucket, object, versionID string, readData bool) ([]FileInfo, []error) {
 	metadataArray := make([]FileInfo, len(disks))
 
 	g := errgroup.WithNErrs(len(disks))
@@ -126,7 +126,7 @@ func readAllFileInfo(ctx context.Context, disks []StorageAPI, bucket, object, ve
 			if disks[index] == nil {
 				return errDiskNotFound
 			}
-			metadataArray[index], err = disks[index].ReadVersion(ctx, bucket, object, versionID)
+			metadataArray[index], err = disks[index].ReadVersion(ctx, bucket, object, versionID, readData)
 			if err != nil {
 				if !IsErr(err, errFileNotFound, errVolumeNotFound, errFileVersionNotFound, errDiskNotFound) {
 					logger.LogOnceIf(ctx, err, disks[index].String())

cmd/erasure-multipart.go

@@ -46,7 +46,7 @@ func (er erasureObjects) checkUploadIDExists(ctx context.Context, bucket, object
 	disks := er.getDisks()
 
 	// Read metadata associated with the object from all disks.
-	metaArr, errs := readAllFileInfo(ctx, disks, minioMetaMultipartBucket, er.getUploadIDDir(bucket, object, uploadID), "")
+	metaArr, errs := readAllFileInfo(ctx, disks, minioMetaMultipartBucket, er.getUploadIDDir(bucket, object, uploadID), "", false)
 
 	readQuorum, _, err := objectQuorumFromMeta(ctx, er, metaArr, errs)
 	if err != nil {
@@ -113,7 +113,7 @@ func (er erasureObjects) cleanupStaleUploadsOnDisk(ctx context.Context, disk Sto
 		}
 		for _, uploadIDDir := range uploadIDDirs {
 			uploadIDPath := pathJoin(shaDir, uploadIDDir)
-			fi, err := disk.ReadVersion(ctx, minioMetaMultipartBucket, uploadIDPath, "")
+			fi, err := disk.ReadVersion(ctx, minioMetaMultipartBucket, uploadIDPath, "", false)
 			if err != nil {
 				continue
 			}
@@ -127,7 +127,7 @@ func (er erasureObjects) cleanupStaleUploadsOnDisk(ctx context.Context, disk Sto
 		return
 	}
 	for _, tmpDir := range tmpDirs {
-		fi, err := disk.ReadVersion(ctx, minioMetaTmpBucket, tmpDir, "")
+		fi, err := disk.ReadVersion(ctx, minioMetaTmpBucket, tmpDir, "", false)
 		if err != nil {
 			continue
 		}
@@ -181,7 +181,7 @@ func (er erasureObjects) ListMultipartUploads(ctx context.Context, bucket, objec
 		if populatedUploadIds.Contains(uploadID) {
 			continue
 		}
-		fi, err := disk.ReadVersion(ctx, minioMetaMultipartBucket, pathJoin(er.getUploadIDDir(bucket, object, uploadID)), "")
+		fi, err := disk.ReadVersion(ctx, minioMetaMultipartBucket, pathJoin(er.getUploadIDDir(bucket, object, uploadID)), "", false)
 		if err != nil {
 			return result, toObjectErr(err, bucket, object)
 		}
@@ -371,7 +371,7 @@ func (er erasureObjects) PutObjectPart(ctx context.Context, bucket, object, uplo
 
 	// Read metadata associated with the object from all disks.
 	partsMetadata, errs = readAllFileInfo(ctx, storageDisks, minioMetaMultipartBucket,
-		uploadIDPath, "")
+		uploadIDPath, "", false)
 
 	// get Quorum for this object
 	_, writeQuorum, err := objectQuorumFromMeta(ctx, er, partsMetadata, errs)
@@ -474,7 +474,7 @@ func (er erasureObjects) PutObjectPart(ctx context.Context, bucket, object, uplo
 	}
 
 	// Read metadata again because it might be updated with parallel upload of another part.
-	partsMetadata, errs = readAllFileInfo(ctx, onlineDisks, minioMetaMultipartBucket, uploadIDPath, "")
+	partsMetadata, errs = readAllFileInfo(ctx, onlineDisks, minioMetaMultipartBucket, uploadIDPath, "", false)
 	reducedErr = reduceWriteQuorumErrs(ctx, errs, objectOpIgnoredErrs, writeQuorum)
 	if reducedErr == errErasureWriteQuorum {
 		return pi, toObjectErr(reducedErr, bucket, object)
@@ -552,7 +552,7 @@ func (er erasureObjects) GetMultipartInfo(ctx context.Context, bucket, object, u
 	storageDisks := er.getDisks()
 
 	// Read metadata associated with the object from all disks.
-	partsMetadata, errs := readAllFileInfo(ctx, storageDisks, minioMetaMultipartBucket, uploadIDPath, opts.VersionID)
+	partsMetadata, errs := readAllFileInfo(ctx, storageDisks, minioMetaMultipartBucket, uploadIDPath, opts.VersionID, false)
 
 	// get Quorum for this object
 	readQuorum, _, err := objectQuorumFromMeta(ctx, er, partsMetadata, errs)
@@ -600,7 +600,7 @@ func (er erasureObjects) ListObjectParts(ctx context.Context, bucket, object, up
 	storageDisks := er.getDisks()
 
 	// Read metadata associated with the object from all disks.
-	partsMetadata, errs := readAllFileInfo(ctx, storageDisks, minioMetaMultipartBucket, uploadIDPath, "")
+	partsMetadata, errs := readAllFileInfo(ctx, storageDisks, minioMetaMultipartBucket, uploadIDPath, "", false)
 
 	// get Quorum for this object
 	_, writeQuorum, err := objectQuorumFromMeta(ctx, er, partsMetadata, errs)
@@ -704,7 +704,7 @@ func (er erasureObjects) CompleteMultipartUpload(ctx context.Context, bucket str
 	storageDisks := er.getDisks()
 
 	// Read metadata associated with the object from all disks.
-	partsMetadata, errs := readAllFileInfo(ctx, storageDisks, minioMetaMultipartBucket, uploadIDPath, "")
+	partsMetadata, errs := readAllFileInfo(ctx, storageDisks, minioMetaMultipartBucket, uploadIDPath, "", false)
 
 	// get Quorum for this object
 	_, writeQuorum, err := objectQuorumFromMeta(ctx, er, partsMetadata, errs)
@@ -889,7 +889,7 @@ func (er erasureObjects) AbortMultipartUpload(ctx context.Context, bucket, objec
 	uploadIDPath := er.getUploadIDDir(bucket, object, uploadID)
 
 	// Read metadata associated with the object from all disks.
-	partsMetadata, errs := readAllFileInfo(ctx, er.getDisks(), minioMetaMultipartBucket, uploadIDPath, "")
+	partsMetadata, errs := readAllFileInfo(ctx, er.getDisks(), minioMetaMultipartBucket, uploadIDPath, "", false)
 
 	// get Quorum for this object
 	_, writeQuorum, err := objectQuorumFromMeta(ctx, er, partsMetadata, errs)

cmd/erasure-object.go

@@ -58,7 +58,7 @@ func (er erasureObjects) CopyObject(ctx context.Context, srcBucket, srcObject, d
 
 	// Read metadata associated with the object from all disks.
 	storageDisks := er.getDisks()
-	metaArr, errs := readAllFileInfo(ctx, storageDisks, srcBucket, srcObject, srcOpts.VersionID)
+	metaArr, errs := readAllFileInfo(ctx, storageDisks, srcBucket, srcObject, srcOpts.VersionID, false)
 
 	// get Quorum for this object
 	readQuorum, writeQuorum, err := objectQuorumFromMeta(ctx, er, metaArr, errs)
@@ -157,7 +157,7 @@ func (er erasureObjects) GetObjectNInfo(ctx context.Context, bucket, object stri
 		unlockOnDefer = true
 	}
 
-	fi, metaArr, onlineDisks, err := er.getObjectFileInfo(ctx, bucket, object, opts)
+	fi, metaArr, onlineDisks, err := er.getObjectFileInfo(ctx, bucket, object, opts, true)
 	if err != nil {
 		return nil, toObjectErr(err, bucket, object)
 	}
@@ -298,7 +298,8 @@ func (er erasureObjects) getObjectWithFileInfo(ctx context.Context, bucket, obje
 			}
 			checksumInfo := metaArr[index].Erasure.GetChecksumInfo(partNumber)
 			partPath := pathJoin(object, metaArr[index].DataDir, fmt.Sprintf("part.%d", partNumber))
-			readers[index] = newBitrotReader(disk, bucket, partPath, tillOffset,
+			data := metaArr[index].Data
+			readers[index] = newBitrotReader(disk, data, bucket, partPath, tillOffset,
 				checksumInfo.Algorithm, checksumInfo.Hash, erasure.ShardSize())
 
 			// Prefer local disks
@@ -337,7 +338,7 @@ func (er erasureObjects) getObjectWithFileInfo(ctx context.Context, bucket, obje
 
 // getObject wrapper for erasure GetObject
 func (er erasureObjects) getObject(ctx context.Context, bucket, object string, startOffset, length int64, writer io.Writer, opts ObjectOptions) error {
-	fi, metaArr, onlineDisks, err := er.getObjectFileInfo(ctx, bucket, object, opts)
+	fi, metaArr, onlineDisks, err := er.getObjectFileInfo(ctx, bucket, object, opts, true)
 	if err != nil {
 		return toObjectErr(err, bucket, object)
 	}
@@ -364,11 +365,11 @@ func (er erasureObjects) GetObjectInfo(ctx context.Context, bucket, object strin
 	return er.getObjectInfo(ctx, bucket, object, opts)
 }
 
-func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object string, opts ObjectOptions) (fi FileInfo, metaArr []FileInfo, onlineDisks []StorageAPI, err error) {
+func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object string, opts ObjectOptions, readData bool) (fi FileInfo, metaArr []FileInfo, onlineDisks []StorageAPI, err error) {
 	disks := er.getDisks()
 
 	// Read metadata associated with the object from all disks.
-	metaArr, errs := readAllFileInfo(ctx, disks, bucket, object, opts.VersionID)
+	metaArr, errs := readAllFileInfo(ctx, disks, bucket, object, opts.VersionID, readData)
 
 	readQuorum, _, err := objectQuorumFromMeta(ctx, er, metaArr, errs)
 	if err != nil {
@@ -410,7 +411,7 @@ func (er erasureObjects) getObjectFileInfo(ctx context.Context, bucket, object s
 
 // getObjectInfo - wrapper for reading object metadata and constructs ObjectInfo.
 func (er erasureObjects) getObjectInfo(ctx context.Context, bucket, object string, opts ObjectOptions) (objInfo ObjectInfo, err error) {
-	fi, _, _, err := er.getObjectFileInfo(ctx, bucket, object, opts)
+	fi, _, _, err := er.getObjectFileInfo(ctx, bucket, object, opts, false)
 	if err != nil {
 		return objInfo, toObjectErr(err, bucket, object)
 	}
@@ -1073,7 +1074,7 @@ func (er erasureObjects) PutObjectTags(ctx context.Context, bucket, object strin
 	disks := er.getDisks()
 
 	// Read metadata associated with the object from all disks.
-	metaArr, errs := readAllFileInfo(ctx, disks, bucket, object, opts.VersionID)
+	metaArr, errs := readAllFileInfo(ctx, disks, bucket, object, opts.VersionID, false)
 
 	readQuorum, writeQuorum, err := objectQuorumFromMeta(ctx, er, metaArr, errs)
 	if err != nil {
@@ -1134,7 +1135,7 @@ func (er erasureObjects) updateObjectMeta(ctx context.Context, bucket, object st
 	disks := er.getDisks()
 
 	// Read metadata associated with the object from all disks.
-	metaArr, errs := readAllFileInfo(ctx, disks, bucket, object, opts.VersionID)
+	metaArr, errs := readAllFileInfo(ctx, disks, bucket, object, opts.VersionID, false)
 
 	readQuorum, writeQuorum, err := objectQuorumFromMeta(ctx, er, metaArr, errs)
 	if err != nil {

cmd/erasure-object_test.go

@@ -19,7 +19,9 @@ package cmd
 import (
 	"bytes"
 	"context"
+	crand "crypto/rand"
 	"errors"
+	"io"
 	"io/ioutil"
 	"os"
 	"testing"
@@ -288,9 +290,13 @@ func TestGetObjectNoQuorum(t *testing.T) {
 	bucket := "bucket"
 	object := "object"
 	opts := ObjectOptions{}
+	buf := make([]byte, 33<<10)
+	if _, err = io.ReadFull(crand.Reader, buf); err != nil {
+		t.Fatal(err)
+	}
 
 	// Test use case 1: All disks are online, xl.meta are present, but data are missing
-	_, err = obj.PutObject(ctx, bucket, object, mustGetPutObjReader(t, bytes.NewReader([]byte("abcd")), int64(len("abcd")), "", ""), opts)
+	_, err = obj.PutObject(ctx, bucket, object, mustGetPutObjReader(t, bytes.NewReader(buf), int64(len(buf)), "", ""), opts)
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -304,7 +310,7 @@ func TestGetObjectNoQuorum(t *testing.T) {
 		}
 	}
 
-	err = xl.GetObject(ctx, bucket, object, 0, int64(len("abcd")), ioutil.Discard, "", opts)
+	err = xl.GetObject(ctx, bucket, object, 0, int64(len(buf)), ioutil.Discard, "", opts)
 	if err != toObjectErr(errFileNotFound, bucket, object) {
 		t.Errorf("Expected GetObject to fail with %v, but failed with %v", toObjectErr(errErasureWriteQuorum, bucket, object), err)
 	}
@@ -315,7 +321,7 @@ func TestGetObjectNoQuorum(t *testing.T) {
 	// invocations, where f - [0,2)
 
 	// Create "object" under "bucket".
-	_, err = obj.PutObject(ctx, bucket, object, mustGetPutObjReader(t, bytes.NewReader([]byte("abcd")), int64(len("abcd")), "", ""), opts)
+	_, err = obj.PutObject(ctx, bucket, object, mustGetPutObjReader(t, bytes.NewReader(buf), int64(len(buf)), "", ""), opts)
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -521,7 +527,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}
 
-	parts1, errs1 := readAllFileInfo(ctx, erasureDisks, bucket, object1, "")
+	parts1, errs1 := readAllFileInfo(ctx, erasureDisks, bucket, object1, "", false)
 
 	parts1SC := globalStorageClass
 
@@ -534,7 +540,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}
 
-	parts2, errs2 := readAllFileInfo(ctx, erasureDisks, bucket, object2, "")
+	parts2, errs2 := readAllFileInfo(ctx, erasureDisks, bucket, object2, "", false)
 	parts2SC := globalStorageClass
 
 	// Object for test case 3 - No StorageClass defined, MetaData in PutObject requesting Standard Storage Class
@@ -546,7 +552,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}
 
-	parts3, errs3 := readAllFileInfo(ctx, erasureDisks, bucket, object3, "")
+	parts3, errs3 := readAllFileInfo(ctx, erasureDisks, bucket, object3, "", false)
 	parts3SC := globalStorageClass
 
 	// Object for test case 4 - Standard StorageClass defined as Parity 6, MetaData in PutObject requesting Standard Storage Class
@@ -564,7 +570,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}
 
-	parts4, errs4 := readAllFileInfo(ctx, erasureDisks, bucket, object4, "")
+	parts4, errs4 := readAllFileInfo(ctx, erasureDisks, bucket, object4, "", false)
 	parts4SC := storageclass.Config{
 		Standard: storageclass.StorageClass{
 			Parity: 6,
@@ -587,7 +593,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}
 
-	parts5, errs5 := readAllFileInfo(ctx, erasureDisks, bucket, object5, "")
+	parts5, errs5 := readAllFileInfo(ctx, erasureDisks, bucket, object5, "", false)
 	parts5SC := storageclass.Config{
 		RRS: storageclass.StorageClass{
 			Parity: 2,
@@ -609,7 +615,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}
 
-	parts6, errs6 := readAllFileInfo(ctx, erasureDisks, bucket, object6, "")
+	parts6, errs6 := readAllFileInfo(ctx, erasureDisks, bucket, object6, "", false)
 	parts6SC := storageclass.Config{
 		RRS: storageclass.StorageClass{
 			Parity: 2,
@@ -632,7 +638,7 @@ func testObjectQuorumFromMeta(obj ObjectLayer, instanceType string, dirs []strin
 		t.Fatalf("Failed to putObject %v", err)
 	}
 
-	parts7, errs7 := readAllFileInfo(ctx, erasureDisks, bucket, object7, "")
+	parts7, errs7 := readAllFileInfo(ctx, erasureDisks, bucket, object7, "", false)
 	parts7SC := storageclass.Config{
 		Standard: storageclass.StorageClass{
 			Parity: 5,

cmd/metacache-set.go

@@ -388,7 +388,7 @@ func (er *erasureObjects) streamMetadataParts(ctx context.Context, o listPathOpt
 				continue
 			}
 
-			_, err := disks[0].ReadVersion(ctx, minioMetaBucket, o.objectPath(0), "")
+			_, err := disks[0].ReadVersion(ctx, minioMetaBucket, o.objectPath(0), "", false)
 			if err != nil {
 				time.Sleep(retryDelay)
 				retries++
@@ -397,7 +397,7 @@ func (er *erasureObjects) streamMetadataParts(ctx context.Context, o listPathOpt
 		}
 
 		// Read metadata associated with the object from all disks.
-		fi, metaArr, onlineDisks, err := er.getObjectFileInfo(ctx, minioMetaBucket, o.objectPath(0), ObjectOptions{})
+		fi, metaArr, onlineDisks, err := er.getObjectFileInfo(ctx, minioMetaBucket, o.objectPath(0), ObjectOptions{}, true)
 		if err != nil {
 			switch toObjectErr(err, minioMetaBucket, o.objectPath(0)).(type) {
 			case ObjectNotFound:
@@ -463,7 +463,7 @@ func (er *erasureObjects) streamMetadataParts(ctx context.Context, o listPathOpt
 						continue
 					}
 
-					_, err := disks[0].ReadVersion(ctx, minioMetaBucket, o.objectPath(partN), "")
+					_, err := disks[0].ReadVersion(ctx, minioMetaBucket, o.objectPath(partN), "", false)
 					if err != nil {
 						time.Sleep(retryDelay)
 						retries++
@@ -471,7 +471,7 @@ func (er *erasureObjects) streamMetadataParts(ctx context.Context, o listPathOpt
 					}
 				}
 				// Load first part metadata...
-				fi, metaArr, onlineDisks, err = er.getObjectFileInfo(ctx, minioMetaBucket, o.objectPath(partN), ObjectOptions{})
+				fi, metaArr, onlineDisks, err = er.getObjectFileInfo(ctx, minioMetaBucket, o.objectPath(partN), ObjectOptions{}, true)
 				if err != nil {
 					time.Sleep(retryDelay)
 					retries++

cmd/naughty-disk_test.go

@@ -252,11 +252,11 @@ func (d *naughtyDisk) DeleteVersion(ctx context.Context, volume, path string, fi
 	return d.disk.DeleteVersion(ctx, volume, path, fi)
 }
 
-func (d *naughtyDisk) ReadVersion(ctx context.Context, volume, path, versionID string) (fi FileInfo, err error) {
+func (d *naughtyDisk) ReadVersion(ctx context.Context, volume, path, versionID string, readData bool) (fi FileInfo, err error) {
 	if err := d.calcError(); err != nil {
 		return FileInfo{}, err
 	}
-	return d.disk.ReadVersion(ctx, volume, path, versionID)
+	return d.disk.ReadVersion(ctx, volume, path, versionID, readData)
 }
 
 func (d *naughtyDisk) WriteAll(ctx context.Context, volume string, path string, b []byte) (err error) {

cmd/storage-datatypes.go

@@ -151,6 +151,8 @@ type FileInfo struct {
 	MarkDeleted                   bool // mark this version as deleted
 	DeleteMarkerReplicationStatus string
 	VersionPurgeStatus            VersionPurgeStatusType
+
+	Data []byte // optionally carries object data
 }
 
 // VersionPurgeStatusKey denotes purge status in metadata

cmd/storage-datatypes_gen.go

@@ -245,8 +245,8 @@ func (z *FileInfo) DecodeMsg(dc *msgp.Reader) (err error) {
 		err = msgp.WrapError(err)
 		return
 	}
-	if zb0001 != 17 {
-		err = msgp.ArrayError{Wanted: 17, Got: zb0001}
+	if zb0001 != 18 {
+		err = msgp.ArrayError{Wanted: 18, Got: zb0001}
 		return
 	}
 	z.Volume, err = dc.ReadString()
@@ -375,13 +375,18 @@ func (z *FileInfo) DecodeMsg(dc *msgp.Reader) (err error) {
 		}
 		z.VersionPurgeStatus = VersionPurgeStatusType(zb0004)
 	}
+	z.Data, err = dc.ReadBytes(z.Data)
+	if err != nil {
+		err = msgp.WrapError(err, "Data")
+		return
+	}
 	return
 }
 
 // EncodeMsg implements msgp.Encodable
 func (z *FileInfo) EncodeMsg(en *msgp.Writer) (err error) {
-	// array header, size 17
-	err = en.Append(0xdc, 0x0, 0x11)
+	// array header, size 18
+	err = en.Append(0xdc, 0x0, 0x12)
 	if err != nil {
 		return
 	}
@@ -489,14 +494,19 @@ func (z *FileInfo) EncodeMsg(en *msgp.Writer) (err error) {
 		err = msgp.WrapError(err, "VersionPurgeStatus")
 		return
 	}
+	err = en.WriteBytes(z.Data)
+	if err != nil {
+		err = msgp.WrapError(err, "Data")
+		return
+	}
 	return
 }
 
 // MarshalMsg implements msgp.Marshaler
 func (z *FileInfo) MarshalMsg(b []byte) (o []byte, err error) {
 	o = msgp.Require(b, z.Msgsize())
-	// array header, size 17
-	o = append(o, 0xdc, 0x0, 0x11)
+	// array header, size 18
+	o = append(o, 0xdc, 0x0, 0x12)
 	o = msgp.AppendString(o, z.Volume)
 	o = msgp.AppendString(o, z.Name)
 	o = msgp.AppendString(o, z.VersionID)
@@ -529,6 +539,7 @@ func (z *FileInfo) MarshalMsg(b []byte) (o []byte, err error) {
 	o = msgp.AppendBool(o, z.MarkDeleted)
 	o = msgp.AppendString(o, z.DeleteMarkerReplicationStatus)
 	o = msgp.AppendString(o, string(z.VersionPurgeStatus))
+	o = msgp.AppendBytes(o, z.Data)
 	return
 }
 
@@ -540,8 +551,8 @@ func (z *FileInfo) UnmarshalMsg(bts []byte) (o []byte, err error) {
 		err = msgp.WrapError(err)
 		return
 	}
-	if zb0001 != 17 {
-		err = msgp.ArrayError{Wanted: 17, Got: zb0001}
+	if zb0001 != 18 {
+		err = msgp.ArrayError{Wanted: 18, Got: zb0001}
 		return
 	}
 	z.Volume, bts, err = msgp.ReadStringBytes(bts)
@@ -670,6 +681,11 @@ func (z *FileInfo) UnmarshalMsg(bts []byte) (o []byte, err error) {
 		}
 		z.VersionPurgeStatus = VersionPurgeStatusType(zb0004)
 	}
+	z.Data, bts, err = msgp.ReadBytesBytes(bts, z.Data)
+	if err != nil {
+		err = msgp.WrapError(err, "Data")
+		return
+	}
 	o = bts
 	return
 }
@@ -687,7 +703,7 @@ func (z *FileInfo) Msgsize() (s int) {
 	for za0003 := range z.Parts {
 		s += z.Parts[za0003].Msgsize()
 	}
-	s += z.Erasure.Msgsize() + msgp.BoolSize + msgp.StringPrefixSize + len(z.DeleteMarkerReplicationStatus) + msgp.StringPrefixSize + len(string(z.VersionPurgeStatus))
+	s += z.Erasure.Msgsize() + msgp.BoolSize + msgp.StringPrefixSize + len(z.DeleteMarkerReplicationStatus) + msgp.StringPrefixSize + len(string(z.VersionPurgeStatus)) + msgp.BytesPrefixSize + len(z.Data)
 	return
 }
 

cmd/storage-interface.go

@@ -58,7 +58,7 @@ type StorageAPI interface {
 	DeleteVersion(ctx context.Context, volume, path string, fi FileInfo) error
 	DeleteVersions(ctx context.Context, volume string, versions []FileInfo) []error
 	WriteMetadata(ctx context.Context, volume, path string, fi FileInfo) error
-	ReadVersion(ctx context.Context, volume, path, versionID string) (FileInfo, error)
+	ReadVersion(ctx context.Context, volume, path, versionID string, readData bool) (FileInfo, error)
 	RenameData(ctx context.Context, srcVolume, srcPath, dataDir, dstVolume, dstPath string) error
 
 	// File operations.

cmd/storage-rest-client.go

@@ -36,6 +36,7 @@ import (
 	"github.com/minio/minio/cmd/logger"
 	"github.com/minio/minio/cmd/rest"
 	xnet "github.com/minio/minio/pkg/net"
+	xbufio "github.com/philhofer/fwd"
 	"github.com/tinylib/msgp/msgp"
 )
 
@@ -226,8 +227,7 @@ func (client *storageRESTClient) DiskInfo(ctx context.Context) (info DiskInfo, e
 				return info, err
 			}
 			defer http.DrainBody(respBody)
-			err = msgp.Decode(respBody, &info)
-			if err != nil {
+			if err = msgp.Decode(respBody, &info); err != nil {
 				return info, err
 			}
 			if info.Error != "" {
@@ -398,18 +398,38 @@ func (client *storageRESTClient) RenameData(ctx context.Context, srcVolume, srcP
 	return err
 }
 
-func (client *storageRESTClient) ReadVersion(ctx context.Context, volume, path, versionID string) (fi FileInfo, err error) {
+// where we keep old *Readers
+var readMsgpReaderPool = sync.Pool{New: func() interface{} { return &msgp.Reader{} }}
+
+// mspNewReader returns a *Reader that reads from the provided reader.
+// The reader will be buffered.
+func msgpNewReader(r io.Reader) *msgp.Reader {
+	p := readMsgpReaderPool.Get().(*msgp.Reader)
+	if p.R == nil {
+		p.R = xbufio.NewReaderSize(r, 8<<10)
+	} else {
+		p.R.Reset(r)
+	}
+	return p
+}
+
+func (client *storageRESTClient) ReadVersion(ctx context.Context, volume, path, versionID string, readData bool) (fi FileInfo, err error) {
 	values := make(url.Values)
 	values.Set(storageRESTVolume, volume)
 	values.Set(storageRESTFilePath, path)
 	values.Set(storageRESTVersionID, versionID)
+	values.Set(storageRESTReadData, strconv.FormatBool(readData))
 
 	respBody, err := client.call(ctx, storageRESTMethodReadVersion, values, nil, -1)
 	if err != nil {
 		return fi, err
 	}
 	defer http.DrainBody(respBody)
-	err = msgp.Decode(respBody, &fi)
+
+	dec := msgpNewReader(respBody)
+	defer readMsgpReaderPool.Put(dec)
+
+	err = fi.DecodeMsg(dec)
 	return fi, err
 }
 
@@ -479,10 +499,12 @@ func (client *storageRESTClient) WalkVersions(ctx context.Context, volume, dirPa
 		defer close(ch)
 		defer http.DrainBody(respBody)
 
-		decoder := msgp.NewReader(respBody)
+		dec := msgpNewReader(respBody)
+		defer readMsgpReaderPool.Put(dec)
+
 		for {
 			var fi FileInfoVersions
-			if gerr := fi.DecodeMsg(decoder); gerr != nil {
+			if gerr := fi.DecodeMsg(dec); gerr != nil {
 				// Upon error return
 				if msgp.Cause(gerr) != io.EOF {
 					logger.LogIf(GlobalContext, gerr)

cmd/storage-rest-common.go

@@ -17,7 +17,7 @@
 package cmd
 
 const (
-	storageRESTVersion       = "v23" // Add small file optimization
+	storageRESTVersion       = "v24" // Add more small file optimization
 	storageRESTVersionPrefix = SlashSeparator + storageRESTVersion
 	storageRESTPrefix        = minioReservedBucketPath + "/storage"
 )
@@ -59,6 +59,7 @@ const (
 	storageRESTDirPath        = "dir-path"
 	storageRESTFilePath       = "file-path"
 	storageRESTVersionID      = "version-id"
+	storageRESTReadData       = "read-data"
 	storageRESTTotalVersions  = "total-versions"
 	storageRESTSrcVolume      = "source-volume"
 	storageRESTSrcPath        = "source-path"

cmd/storage-rest-server.go

@@ -327,12 +327,22 @@ func (s *storageRESTServer) ReadVersionHandler(w http.ResponseWriter, r *http.Re
 	volume := vars[storageRESTVolume]
 	filePath := vars[storageRESTFilePath]
 	versionID := vars[storageRESTVersionID]
-	fi, err := s.storage.ReadVersion(r.Context(), volume, filePath, versionID)
+	readData, err := strconv.ParseBool(vars[storageRESTReadData])
 	if err != nil {
 		s.writeErrorResponse(w, err)
 		return
 	}
-	logger.LogIf(r.Context(), msgp.Encode(w, &fi))
+	fi, err := s.storage.ReadVersion(r.Context(), volume, filePath, versionID, readData)
+	if err != nil {
+		s.writeErrorResponse(w, err)
+		return
+	}
+	bufp := s.storage.rpool.Get().(*[]byte)
+	defer s.storage.rpool.Put(bufp)
+
+	enc := msgp.NewWriterBuf(w, *bufp)
+	logger.LogIf(r.Context(), fi.EncodeMsg(enc))
+	enc.Flush()
 }
 
 // WriteMetadata write new updated metadata.
@@ -1033,7 +1043,7 @@ func registerStorageRESTHandlers(router *mux.Router, endpointServerPools Endpoin
 			subrouter.Methods(http.MethodPost).Path(storageRESTVersionPrefix + storageRESTMethodDeleteVersion).HandlerFunc(httpTraceHdrs(server.DeleteVersionHandler)).
 				Queries(restQueries(storageRESTVolume, storageRESTFilePath)...)
 			subrouter.Methods(http.MethodPost).Path(storageRESTVersionPrefix + storageRESTMethodReadVersion).HandlerFunc(httpTraceHdrs(server.ReadVersionHandler)).
-				Queries(restQueries(storageRESTVolume, storageRESTFilePath, storageRESTVersionID)...)
+				Queries(restQueries(storageRESTVolume, storageRESTFilePath, storageRESTVersionID, storageRESTReadData)...)
 			subrouter.Methods(http.MethodPost).Path(storageRESTVersionPrefix + storageRESTMethodRenameData).HandlerFunc(httpTraceHdrs(server.RenameDataHandler)).
 				Queries(restQueries(storageRESTSrcVolume, storageRESTSrcPath, storageRESTDataDir,
 					storageRESTDstVolume, storageRESTDstPath)...)

cmd/xl-storage-disk-id-check.go

@@ -272,12 +272,12 @@ func (p *xlStorageDiskIDCheck) WriteMetadata(ctx context.Context, volume, path s
 	return p.storage.WriteMetadata(ctx, volume, path, fi)
 }
 
-func (p *xlStorageDiskIDCheck) ReadVersion(ctx context.Context, volume, path, versionID string) (fi FileInfo, err error) {
+func (p *xlStorageDiskIDCheck) ReadVersion(ctx context.Context, volume, path, versionID string, readData bool) (fi FileInfo, err error) {
 	if err = p.checkDiskStale(); err != nil {
 		return fi, err
 	}
 
-	return p.storage.ReadVersion(ctx, volume, path, versionID)
+	return p.storage.ReadVersion(ctx, volume, path, versionID, readData)
 }
 
 func (p *xlStorageDiskIDCheck) ReadAll(ctx context.Context, volume string, path string) (buf []byte, err error) {

cmd/xl-storage.go

@@ -65,6 +65,10 @@ const (
 	// Size of each buffer.
 	readAheadBufSize = 1 << 20
 
+	// Small file threshold below which data accompanies metadata
+	// from storage layer.
+	smallFileThreshold = 32 * humanize.KiByte
+
 	// XL metadata file carries per object metadata.
 	xlStorageFormatFile = "xl.meta"
 )
@@ -1082,7 +1086,14 @@ func (s *xlStorage) renameLegacyMetadata(volume, path string) error {
 }
 
 // ReadVersion - reads metadata and returns FileInfo at path `xl.meta`
-func (s *xlStorage) ReadVersion(ctx context.Context, volume, path, versionID string) (fi FileInfo, err error) {
+// for all objects less than `32KiB` this call returns data as well
+// along with metadata.
+func (s *xlStorage) ReadVersion(ctx context.Context, volume, path, versionID string, readData bool) (fi FileInfo, err error) {
+	volumeDir, err := s.getVolDir(volume)
+	if err != nil {
+		return fi, err
+	}
+
 	buf, err := s.ReadAll(ctx, volume, pathJoin(path, xlStorageFormatFile))
 	if err != nil {
 		if err == errFileNotFound {
@@ -1117,7 +1128,65 @@ func (s *xlStorage) ReadVersion(ctx context.Context, volume, path, versionID str
 		return fi, errFileNotFound
 	}
 
-	return getFileInfo(buf, volume, path, versionID)
+	fi, err = getFileInfo(buf, volume, path, versionID)
+	if err != nil {
+		return fi, err
+	}
+
+	if readData {
+		// Reading data for small objects when
+		// - object has not yet transitioned
+		// - object size lesser than 32KiB
+		// - object has maximum of 1 parts
+		if fi.TransitionStatus == "" && fi.DataDir != "" && fi.Size <= smallFileThreshold && len(fi.Parts) == 1 {
+			fi.Data, err = s.readAllData(volumeDir, pathJoin(volumeDir, path, fi.DataDir, fmt.Sprintf("part.%d", fi.Parts[0].Number)))
+			if err != nil {
+				return FileInfo{}, err
+			}
+		}
+	}
+
+	return fi, nil
+}
+
+func (s *xlStorage) readAllData(volumeDir, filePath string) (buf []byte, err error) {
+	var f *os.File
+	if globalStorageClass.GetDMA() == storageclass.DMAReadWrite {
+		f, err = disk.OpenFileDirectIO(filePath, os.O_RDONLY, 0666)
+	} else {
+		f, err = os.Open(filePath)
+	}
+	if err != nil {
+		if osIsNotExist(err) {
+			// Check if the object doesn't exist because its bucket
+			// is missing in order to return the correct error.
+			_, err = os.Stat(volumeDir)
+			if err != nil && osIsNotExist(err) {
+				return nil, errVolumeNotFound
+			}
+			return nil, errFileNotFound
+		} else if osIsPermission(err) {
+			return nil, errFileAccessDenied
+		} else if isSysErrNotDir(err) || isSysErrIsDir(err) {
+			return nil, errFileNotFound
+		} else if isSysErrHandleInvalid(err) {
+			// This case is special and needs to be handled for windows.
+			return nil, errFileNotFound
+		} else if isSysErrIO(err) {
+			return nil, errFaultyDisk
+		} else if isSysErrTooManyFiles(err) {
+			return nil, errTooManyOpenFiles
+		} else if isSysErrInvalidArg(err) {
+			return nil, errFileNotFound
+		}
+		return nil, err
+	}
+
+	atomic.AddInt32(&s.activeIOCount, 1)
+	rd := &odirectReader{f, nil, nil, true, s, nil}
+	defer rd.Close()
+
+	return ioutil.ReadAll(rd)
 }
 
 // ReadAll reads from r until an error or EOF and returns the data it read.
@@ -1405,35 +1474,13 @@ func (s *xlStorage) ReadFileStream(ctx context.Context, volume, path string, off
 		return nil, err
 	}
 
+	var file *os.File
 	if offset == 0 && globalStorageClass.GetDMA() == storageclass.DMAReadWrite {
-		file, err := disk.OpenFileDirectIO(filePath, os.O_RDONLY, 0666)
-		if err != nil {
-			switch {
-			case osIsNotExist(err):
-				_, err = os.Stat(volumeDir)
-				if err != nil && osIsNotExist(err) {
-					return nil, errVolumeNotFound
-				}
-				return nil, errFileNotFound
-			case osIsPermission(err):
-				return nil, errFileAccessDenied
-			case isSysErrNotDir(err):
-				return nil, errFileAccessDenied
-			case isSysErrIO(err):
-				return nil, errFaultyDisk
-			case isSysErrTooManyFiles(err):
-				return nil, errTooManyOpenFiles
-			default:
-				return nil, err
-			}
-		}
-
-		atomic.AddInt32(&s.activeIOCount, 1)
-		return &odirectReader{file, nil, nil, true, s, nil}, nil
+		file, err = disk.OpenFileDirectIO(filePath, os.O_RDONLY, 0666)
+	} else {
+		// Open the fileile fileor reading.
+		file, err = os.Open(filePath)
 	}
-
-	// Open the file for reading.
-	file, err := os.Open(filePath)
 	if err != nil {
 		switch {
 		case osIsNotExist(err):
@@ -1466,11 +1513,17 @@ func (s *xlStorage) ReadFileStream(ctx context.Context, volume, path string, off
 		return nil, errIsNotRegular
 	}
 
+	atomic.AddInt32(&s.activeIOCount, 1)
+	if offset == 0 && globalStorageClass.GetDMA() == storageclass.DMAReadWrite {
+		return &odirectReader{file, nil, nil, true, s, nil}, nil
+	}
+
+	if offset > 0 {
 		if _, err = file.Seek(offset, io.SeekStart); err != nil {
 			return nil, err
 		}
+	}
 
-	atomic.AddInt32(&s.activeIOCount, 1)
 	r := struct {
 		io.Reader
 		io.Closer
@@ -1517,6 +1570,7 @@ func (s *xlStorage) CreateFile(ctx context.Context, volume, path string, fileSiz
 	if err != nil {
 		return err
 	}
+
 	// Stat a volume entry.
 	_, err = os.Stat(volumeDir)
 	if err != nil {

go.mod

@@ -62,11 +62,10 @@ require (
 	github.com/ncw/directio v1.0.5
 	github.com/nsqio/go-nsq v1.0.8
 	github.com/olivere/elastic/v7 v7.0.22
-	github.com/philhofer/fwd v1.1.1 // indirect
+	github.com/philhofer/fwd v1.1.1
 	github.com/pierrec/lz4 v2.5.2+incompatible
 	github.com/pkg/errors v0.9.1
 	github.com/prometheus/client_golang v1.8.0
-	github.com/quasilyte/go-ruleguard/dsl/fluent v0.0.0-20201222093424-5d7e62a465d3 // indirect
 	github.com/rjeczalik/notify v0.9.2
 	github.com/rs/cors v1.7.0
 	github.com/secure-io/sio-go v0.3.0

go.sum

@@ -539,10 +539,6 @@ github.com/prometheus/procfs v0.0.8/go.mod h1:7Qr8sr6344vo1JqZ6HhLceV9o3AJ1Ff+Gx
 github.com/prometheus/procfs v0.1.3/go.mod h1:lV6e/gmhEcM9IjHGsFOCxxuZ+z1YqCvr4OA4YeYWdaU=
 github.com/prometheus/procfs v0.2.0 h1:wH4vA7pcjKuZzjF7lM8awk4fnuJO6idemZXoKnULUx4=
 github.com/prometheus/procfs v0.2.0/go.mod h1:lV6e/gmhEcM9IjHGsFOCxxuZ+z1YqCvr4OA4YeYWdaU=
-github.com/quasilyte/go-ruleguard v0.2.1 h1:56eRm0daAyny9UhJnmtJW/UyLZQusukBAB8oT8AHKHo=
-github.com/quasilyte/go-ruleguard/dsl v0.0.0-20201222101508-986133edf04e h1:gXFs5pU/5pxy0nw9QoV2dAhGXI+jKSN0GJEL8TMKf4A=
-github.com/quasilyte/go-ruleguard/dsl/fluent v0.0.0-20201222093424-5d7e62a465d3 h1:eL7x4/zMnlquMxYe7V078BD7MGskZ0daGln+SJCVzuY=
-github.com/quasilyte/go-ruleguard/dsl/fluent v0.0.0-20201222093424-5d7e62a465d3/go.mod h1:P7JlQWFT7jDcFZMtUPQbtGzzzxva3rBn6oIF+LPwFcM=
 github.com/rcrowley/go-metrics v0.0.0-20181016184325-3113b8401b8a h1:9ZKAASQSHhDYGoxY8uLVpewe1GDZ2vu2Tr/vTdVAkFQ=
 github.com/rcrowley/go-metrics v0.0.0-20181016184325-3113b8401b8a/go.mod h1:bCqnVzQkZxMG4s8nGwiZ5l3QUCyqpo9Y+/ZMZ9VjZe4=
 github.com/rcrowley/go-metrics v0.0.0-20200313005456-10cdbea86bc0 h1:MkV+77GLUNo5oJ0jf870itWm3D0Sjh7+Za9gazKc5LQ=