minio/cmd/xl-v1-object.go
Harshavardhana 1d8a8c63db Simplify data verification with HashReader. (#5071)
Verify() was being called by caller after the data
has been successfully read after io.EOF. This disconnection
opens a race under concurrent access to such an object.
Verification is not necessary outside of Read() call,
we can simply just do checksum verification right inside
Read() call at io.EOF.

This approach simplifies the usage.
2017-10-22 11:00:34 +05:30

711 lines
23 KiB
Go

/*
* Minio Cloud Storage, (C) 2016, 2017 Minio, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package cmd
import (
"io"
"path"
"strconv"
"strings"
"sync"
"github.com/minio/minio/pkg/hash"
"github.com/minio/minio/pkg/mimedb"
"github.com/minio/minio/pkg/objcache"
)
// list all errors which can be ignored in object operations.
var objectOpIgnoredErrs = append(baseIgnoredErrs, errDiskAccessDenied)
// prepareFile hints the bottom layer to optimize the creation of a new object
func (xl xlObjects) prepareFile(bucket, object string, size int64, onlineDisks []StorageAPI, blockSize int64, dataBlocks int) error {
pErrs := make([]error, len(onlineDisks))
// Calculate the real size of the part in one disk.
actualSize := xl.sizeOnDisk(size, blockSize, dataBlocks)
// Prepare object creation in a all disks
for index, disk := range onlineDisks {
if disk != nil {
if err := disk.PrepareFile(bucket, object, actualSize); err != nil {
// Save error to reduce it later
pErrs[index] = err
// Ignore later access to disk which generated the error
onlineDisks[index] = nil
}
}
}
return reduceWriteQuorumErrs(pErrs, objectOpIgnoredErrs, xl.writeQuorum)
}
/// Object Operations
// CopyObject - copy object source object to destination object.
// if source object and destination object are same we only
// update metadata.
func (xl xlObjects) CopyObject(srcBucket, srcObject, dstBucket, dstObject string, metadata map[string]string) (oi ObjectInfo, e error) {
// Read metadata associated with the object from all disks.
metaArr, errs := readAllXLMetadata(xl.storageDisks, srcBucket, srcObject)
if reducedErr := reduceReadQuorumErrs(errs, objectOpIgnoredErrs, xl.readQuorum); reducedErr != nil {
return oi, toObjectErr(reducedErr, srcBucket, srcObject)
}
// List all online disks.
onlineDisks, modTime := listOnlineDisks(xl.storageDisks, metaArr, errs)
// Pick latest valid metadata.
xlMeta, err := pickValidXLMeta(metaArr, modTime)
if err != nil {
return oi, toObjectErr(err, srcBucket, srcObject)
}
// Reorder online disks based on erasure distribution order.
onlineDisks = shuffleDisks(onlineDisks, xlMeta.Erasure.Distribution)
// Length of the file to read.
length := xlMeta.Stat.Size
// Check if this request is only metadata update.
cpMetadataOnly := isStringEqual(pathJoin(srcBucket, srcObject), pathJoin(dstBucket, dstObject))
if cpMetadataOnly {
xlMeta.Meta = metadata
partsMetadata := make([]xlMetaV1, len(xl.storageDisks))
// Update `xl.json` content on each disks.
for index := range partsMetadata {
partsMetadata[index] = xlMeta
}
tempObj := mustGetUUID()
// Write unique `xl.json` for each disk.
if onlineDisks, err = writeUniqueXLMetadata(onlineDisks, minioMetaTmpBucket, tempObj, partsMetadata, xl.writeQuorum); err != nil {
return oi, toObjectErr(err, srcBucket, srcObject)
}
// Rename atomically `xl.json` from tmp location to destination for each disk.
if _, err = renameXLMetadata(onlineDisks, minioMetaTmpBucket, tempObj, srcBucket, srcObject, xl.writeQuorum); err != nil {
return oi, toObjectErr(err, srcBucket, srcObject)
}
return xlMeta.ToObjectInfo(srcBucket, srcObject), nil
}
// Initialize pipe.
pipeReader, pipeWriter := io.Pipe()
go func() {
var startOffset int64 // Read the whole file.
if gerr := xl.GetObject(srcBucket, srcObject, startOffset, length, pipeWriter); gerr != nil {
errorIf(gerr, "Unable to read %s of the object `%s/%s`.", srcBucket, srcObject)
pipeWriter.CloseWithError(toObjectErr(gerr, srcBucket, srcObject))
return
}
pipeWriter.Close() // Close writer explicitly signalling we wrote all data.
}()
hashReader, err := hash.NewReader(pipeReader, length, "", "")
if err != nil {
return oi, toObjectErr(traceError(err), dstBucket, dstObject)
}
objInfo, err := xl.PutObject(dstBucket, dstObject, hashReader, metadata)
if err != nil {
return oi, toObjectErr(err, dstBucket, dstObject)
}
// Explicitly close the reader.
pipeReader.Close()
return objInfo, nil
}
// GetObject - reads an object erasured coded across multiple
// disks. Supports additional parameters like offset and length
// which are synonymous with HTTP Range requests.
//
// startOffset indicates the starting read location of the object.
// length indicates the total length of the object.
func (xl xlObjects) GetObject(bucket, object string, startOffset int64, length int64, writer io.Writer) error {
if err := checkGetObjArgs(bucket, object); err != nil {
return err
}
// Start offset cannot be negative.
if startOffset < 0 {
return traceError(errUnexpected)
}
// Writer cannot be nil.
if writer == nil {
return traceError(errUnexpected)
}
// Read metadata associated with the object from all disks.
metaArr, errs := readAllXLMetadata(xl.storageDisks, bucket, object)
if reducedErr := reduceReadQuorumErrs(errs, objectOpIgnoredErrs, xl.readQuorum); reducedErr != nil {
return toObjectErr(reducedErr, bucket, object)
}
// List all online disks.
onlineDisks, modTime := listOnlineDisks(xl.storageDisks, metaArr, errs)
// Pick latest valid metadata.
xlMeta, err := pickValidXLMeta(metaArr, modTime)
if err != nil {
return err
}
// Reorder online disks based on erasure distribution order.
onlineDisks = shuffleDisks(onlineDisks, xlMeta.Erasure.Distribution)
// Reorder parts metadata based on erasure distribution order.
metaArr = shufflePartsMetadata(metaArr, xlMeta.Erasure.Distribution)
// For negative length read everything.
if length < 0 {
length = xlMeta.Stat.Size - startOffset
}
// Reply back invalid range if the input offset and length fall out of range.
if startOffset > xlMeta.Stat.Size || startOffset+length > xlMeta.Stat.Size {
return traceError(InvalidRange{startOffset, length, xlMeta.Stat.Size})
}
// Get start part index and offset.
partIndex, partOffset, err := xlMeta.ObjectToPartOffset(startOffset)
if err != nil {
return traceError(InvalidRange{startOffset, length, xlMeta.Stat.Size})
}
// Calculate endOffset according to length
endOffset := startOffset
if length > 0 {
endOffset += length - 1
}
// Get last part index to read given length.
lastPartIndex, _, err := xlMeta.ObjectToPartOffset(endOffset)
if err != nil {
return traceError(InvalidRange{startOffset, length, xlMeta.Stat.Size})
}
// Save the writer.
mw := writer
// Object cache enabled block.
if xlMeta.Stat.Size > 0 && xl.objCacheEnabled {
// Validate if we have previous cache.
var cachedBuffer io.ReaderAt
cachedBuffer, err = xl.objCache.Open(path.Join(bucket, object), modTime)
if err == nil { // Cache hit
// Create a new section reader, starting at an offset with length.
reader := io.NewSectionReader(cachedBuffer, startOffset, length)
// Copy the data out.
if _, err = io.Copy(writer, reader); err != nil {
return traceError(err)
}
// Success.
return nil
} // Cache miss.
// For unknown error, return and error out.
if err != objcache.ErrKeyNotFoundInCache {
return traceError(err)
} // Cache has not been found, fill the cache.
// Cache is only set if whole object is being read.
if startOffset == 0 && length == xlMeta.Stat.Size {
// Proceed to set the cache.
var newBuffer io.WriteCloser
// Create a new entry in memory of length.
newBuffer, err = xl.objCache.Create(path.Join(bucket, object), length)
if err == nil {
// Create a multi writer to write to both memory and client response.
mw = io.MultiWriter(newBuffer, writer)
defer newBuffer.Close()
}
// Ignore error if cache is full, proceed to write the object.
if err != nil && err != objcache.ErrCacheFull {
// For any other error return here.
return toObjectErr(traceError(err), bucket, object)
}
}
}
var totalBytesRead int64
storage, err := NewErasureStorage(onlineDisks, xlMeta.Erasure.DataBlocks, xlMeta.Erasure.ParityBlocks)
if err != nil {
return toObjectErr(err, bucket, object)
}
checksums := make([][]byte, len(storage.disks))
for ; partIndex <= lastPartIndex; partIndex++ {
if length == totalBytesRead {
break
}
// Save the current part name and size.
partName := xlMeta.Parts[partIndex].Name
partSize := xlMeta.Parts[partIndex].Size
readSize := partSize - partOffset
// readSize should be adjusted so that we don't write more data than what was requested.
if readSize > (length - totalBytesRead) {
readSize = length - totalBytesRead
}
// Get the checksums of the current part.
var algorithm BitrotAlgorithm
for index, disk := range storage.disks {
if disk == OfflineDisk {
continue
}
checksumInfo := metaArr[index].Erasure.GetChecksumInfo(partName)
algorithm = checksumInfo.Algorithm
checksums[index] = checksumInfo.Hash
}
file, err := storage.ReadFile(mw, bucket, pathJoin(object, partName), partOffset, readSize, partSize, checksums, algorithm, xlMeta.Erasure.BlockSize)
if err != nil {
errorIf(err, "Unable to read %s of the object `%s/%s`.", partName, bucket, object)
return toObjectErr(err, bucket, object)
}
// Track total bytes read from disk and written to the client.
totalBytesRead += file.Size
// partOffset will be valid only for the first part, hence reset it to 0 for
// the remaining parts.
partOffset = 0
} // End of read all parts loop.
// Return success.
return nil
}
// GetObjectInfo - reads object metadata and replies back ObjectInfo.
func (xl xlObjects) GetObjectInfo(bucket, object string) (oi ObjectInfo, e error) {
if err := checkGetObjArgs(bucket, object); err != nil {
return oi, err
}
info, err := xl.getObjectInfo(bucket, object)
if err != nil {
return oi, toObjectErr(err, bucket, object)
}
return info, nil
}
// getObjectInfo - wrapper for reading object metadata and constructs ObjectInfo.
func (xl xlObjects) getObjectInfo(bucket, object string) (objInfo ObjectInfo, err error) {
// Extracts xlStat and xlMetaMap.
xlStat, xlMetaMap, err := xl.readXLMetaStat(bucket, object)
if err != nil {
return ObjectInfo{}, err
}
objInfo = ObjectInfo{
IsDir: false,
Bucket: bucket,
Name: object,
Size: xlStat.Size,
ModTime: xlStat.ModTime,
ContentType: xlMetaMap["content-type"],
ContentEncoding: xlMetaMap["content-encoding"],
}
// Extract etag.
objInfo.ETag = extractETag(xlMetaMap)
// etag/md5Sum has already been extracted. We need to
// remove to avoid it from appearing as part of
// response headers. e.g, X-Minio-* or X-Amz-*.
objInfo.UserDefined = cleanMetaETag(xlMetaMap)
// Success.
return objInfo, nil
}
func undoRename(disks []StorageAPI, srcBucket, srcEntry, dstBucket, dstEntry string, isDir bool, errs []error) {
var wg = &sync.WaitGroup{}
// Undo rename object on disks where RenameFile succeeded.
// If srcEntry/dstEntry are objects then add a trailing slash to copy
// over all the parts inside the object directory
if isDir {
srcEntry = retainSlash(srcEntry)
dstEntry = retainSlash(dstEntry)
}
for index, disk := range disks {
if disk == nil {
continue
}
// Undo rename object in parallel.
wg.Add(1)
go func(index int, disk StorageAPI) {
defer wg.Done()
if errs[index] != nil {
return
}
_ = disk.RenameFile(dstBucket, dstEntry, srcBucket, srcEntry)
}(index, disk)
}
wg.Wait()
}
// rename - common function that renamePart and renameObject use to rename
// the respective underlying storage layer representations.
func rename(disks []StorageAPI, srcBucket, srcEntry, dstBucket, dstEntry string, isDir bool, quorum int) ([]StorageAPI, error) {
// Initialize sync waitgroup.
var wg = &sync.WaitGroup{}
// Initialize list of errors.
var errs = make([]error, len(disks))
if isDir {
dstEntry = retainSlash(dstEntry)
srcEntry = retainSlash(srcEntry)
}
// Rename file on all underlying storage disks.
for index, disk := range disks {
if disk == nil {
errs[index] = errDiskNotFound
continue
}
wg.Add(1)
go func(index int, disk StorageAPI) {
defer wg.Done()
err := disk.RenameFile(srcBucket, srcEntry, dstBucket, dstEntry)
if err != nil && err != errFileNotFound {
errs[index] = traceError(err)
}
}(index, disk)
}
// Wait for all renames to finish.
wg.Wait()
// We can safely allow RenameFile errors up to len(xl.storageDisks) - xl.writeQuorum
// otherwise return failure. Cleanup successful renames.
err := reduceWriteQuorumErrs(errs, objectOpIgnoredErrs, quorum)
if errorCause(err) == errXLWriteQuorum {
// Undo all the partial rename operations.
undoRename(disks, srcBucket, srcEntry, dstBucket, dstEntry, isDir, errs)
}
return evalDisks(disks, errs), err
}
// renamePart - renames a part of the source object to the destination
// across all disks in parallel. Additionally if we have errors and do
// not have a readQuorum partially renamed files are renamed back to
// its proper location.
func renamePart(disks []StorageAPI, srcBucket, srcPart, dstBucket, dstPart string, quorum int) ([]StorageAPI, error) {
isDir := false
return rename(disks, srcBucket, srcPart, dstBucket, dstPart, isDir, quorum)
}
// renameObject - renames all source objects to destination object
// across all disks in parallel. Additionally if we have errors and do
// not have a readQuorum partially renamed files are renamed back to
// its proper location.
func renameObject(disks []StorageAPI, srcBucket, srcObject, dstBucket, dstObject string, quorum int) ([]StorageAPI, error) {
isDir := true
return rename(disks, srcBucket, srcObject, dstBucket, dstObject, isDir, quorum)
}
// PutObject - creates an object upon reading from the input stream
// until EOF, erasure codes the data across all disk and additionally
// writes `xl.json` which carries the necessary metadata for future
// object operations.
func (xl xlObjects) PutObject(bucket string, object string, data *hash.Reader, metadata map[string]string) (objInfo ObjectInfo, err error) {
// This is a special case with size as '0' and object ends with
// a slash separator, we treat it like a valid operation and
// return success.
if isObjectDir(object, data.Size()) {
// Check if an object is present as one of the parent dir.
// -- FIXME. (needs a new kind of lock).
// -- FIXME (this also causes performance issue when disks are down).
if xl.parentDirIsObject(bucket, path.Dir(object)) {
return ObjectInfo{}, toObjectErr(traceError(errFileAccessDenied), bucket, object)
}
return dirObjectInfo(bucket, object, data.Size(), metadata), nil
}
// Validate put object input args.
if err = checkPutObjectArgs(bucket, object, xl); err != nil {
return ObjectInfo{}, err
}
// Validate input data size and it can never be less than zero.
if data.Size() < 0 {
return ObjectInfo{}, toObjectErr(traceError(errInvalidArgument))
}
// Check if an object is present as one of the parent dir.
// -- FIXME. (needs a new kind of lock).
// -- FIXME (this also causes performance issue when disks are down).
if xl.parentDirIsObject(bucket, path.Dir(object)) {
return ObjectInfo{}, toObjectErr(traceError(errFileAccessDenied), bucket, object)
}
// No metadata is set, allocate a new one.
if metadata == nil {
metadata = make(map[string]string)
}
uniqueID := mustGetUUID()
tempObj := uniqueID
// Limit the reader to its provided size if specified.
var reader io.Reader = data
// Proceed to set the cache.
var newBuffer io.WriteCloser
// If caching is enabled, proceed to set the cache.
if data.Size() > 0 && xl.objCacheEnabled {
// PutObject invalidates any previously cached object in memory.
xl.objCache.Delete(path.Join(bucket, object))
// Create a new entry in memory of size.
newBuffer, err = xl.objCache.Create(path.Join(bucket, object), data.Size())
if err == nil {
// Cache incoming data into a buffer
reader = io.TeeReader(data, newBuffer)
} else {
// Return errors other than ErrCacheFull
if err != objcache.ErrCacheFull {
return ObjectInfo{}, toObjectErr(traceError(err), bucket, object)
}
}
}
// Initialize parts metadata
partsMetadata := make([]xlMetaV1, len(xl.storageDisks))
xlMeta := newXLMetaV1(object, xl.dataBlocks, xl.parityBlocks)
// Initialize xl meta.
for index := range partsMetadata {
partsMetadata[index] = xlMeta
}
// Order disks according to erasure distribution
onlineDisks := shuffleDisks(xl.storageDisks, partsMetadata[0].Erasure.Distribution)
// Delete temporary object in the event of failure.
// If PutObject succeeded there would be no temporary
// object to delete.
defer xl.deleteObject(minioMetaTmpBucket, tempObj)
// Total size of the written object
var sizeWritten int64
storage, err := NewErasureStorage(onlineDisks, xlMeta.Erasure.DataBlocks, xlMeta.Erasure.ParityBlocks)
if err != nil {
return ObjectInfo{}, toObjectErr(err, bucket, object)
}
// Alloc additional space for parity blocks created while erasure coding
buffer := make([]byte, partsMetadata[0].Erasure.BlockSize, 2*partsMetadata[0].Erasure.BlockSize)
// Read data and split into parts - similar to multipart mechanism
for partIdx := 1; ; partIdx++ {
// Compute part name
partName := "part." + strconv.Itoa(partIdx)
// Compute the path of current part
tempErasureObj := pathJoin(uniqueID, partName)
// Calculate the size of the current part.
var curPartSize int64
curPartSize, err = calculatePartSizeFromIdx(data.Size(), globalPutPartSize, partIdx)
if err != nil {
return ObjectInfo{}, toObjectErr(err, bucket, object)
}
// Hint the filesystem to pre-allocate one continuous large block.
// This is only an optimization.
var curPartReader io.Reader
if curPartSize > 0 {
pErr := xl.prepareFile(minioMetaTmpBucket, tempErasureObj, curPartSize, storage.disks, xlMeta.Erasure.BlockSize, xlMeta.Erasure.DataBlocks)
if pErr != nil {
return ObjectInfo{}, toObjectErr(pErr, bucket, object)
}
}
if curPartSize < data.Size() {
curPartReader = io.LimitReader(reader, curPartSize)
} else {
curPartReader = reader
}
file, erasureErr := storage.CreateFile(curPartReader, minioMetaTmpBucket,
tempErasureObj, buffer, DefaultBitrotAlgorithm, xl.writeQuorum)
if erasureErr != nil {
return ObjectInfo{}, toObjectErr(erasureErr, minioMetaTmpBucket, tempErasureObj)
}
// Should return IncompleteBody{} error when reader has fewer bytes
// than specified in request header.
if file.Size < curPartSize {
return ObjectInfo{}, traceError(IncompleteBody{})
}
// Update the total written size
sizeWritten += file.Size
for i := range partsMetadata {
partsMetadata[i].AddObjectPart(partIdx, partName, "", file.Size)
partsMetadata[i].Erasure.AddChecksumInfo(ChecksumInfo{partName, file.Algorithm, file.Checksums[i]})
}
// We wrote everything, break out.
if sizeWritten == data.Size() {
break
}
}
// Save additional erasureMetadata.
modTime := UTCNow()
metadata["etag"] = data.MD5HexString()
// Guess content-type from the extension if possible.
if metadata["content-type"] == "" {
if objectExt := path.Ext(object); objectExt != "" {
if content, ok := mimedb.DB[strings.ToLower(strings.TrimPrefix(objectExt, "."))]; ok {
metadata["content-type"] = content.ContentType
}
}
}
if xl.isObject(bucket, object) {
// Rename if an object already exists to temporary location.
newUniqueID := mustGetUUID()
// Delete successfully renamed object.
defer xl.deleteObject(minioMetaTmpBucket, newUniqueID)
// NOTE: Do not use online disks slice here.
// The reason is that existing object should be purged
// regardless of `xl.json` status and rolled back in case of errors.
_, err = renameObject(xl.storageDisks, bucket, object, minioMetaTmpBucket, newUniqueID, xl.writeQuorum)
if err != nil {
return ObjectInfo{}, toObjectErr(err, bucket, object)
}
}
// Fill all the necessary metadata.
// Update `xl.json` content on each disks.
for index := range partsMetadata {
partsMetadata[index].Meta = metadata
partsMetadata[index].Stat.Size = sizeWritten
partsMetadata[index].Stat.ModTime = modTime
}
// Write unique `xl.json` for each disk.
if onlineDisks, err = writeUniqueXLMetadata(onlineDisks, minioMetaTmpBucket, tempObj, partsMetadata, xl.writeQuorum); err != nil {
return ObjectInfo{}, toObjectErr(err, bucket, object)
}
// Rename the successfully written temporary object to final location.
if _, err = renameObject(onlineDisks, minioMetaTmpBucket, tempObj, bucket, object, xl.writeQuorum); err != nil {
return ObjectInfo{}, toObjectErr(err, bucket, object)
}
// Once we have successfully renamed the object, Close the buffer which would
// save the object on cache.
if sizeWritten > 0 && xl.objCacheEnabled && newBuffer != nil {
newBuffer.Close()
}
// Object info is the same in all disks, so we can pick the first meta
// of the first disk
xlMeta = partsMetadata[0]
objInfo = ObjectInfo{
IsDir: false,
Bucket: bucket,
Name: object,
Size: xlMeta.Stat.Size,
ModTime: xlMeta.Stat.ModTime,
ETag: xlMeta.Meta["etag"],
ContentType: xlMeta.Meta["content-type"],
ContentEncoding: xlMeta.Meta["content-encoding"],
UserDefined: xlMeta.Meta,
}
// Success, return object info.
return objInfo, nil
}
// deleteObject - wrapper for delete object, deletes an object from
// all the disks in parallel, including `xl.json` associated with the
// object.
func (xl xlObjects) deleteObject(bucket, object string) error {
// Initialize sync waitgroup.
var wg = &sync.WaitGroup{}
// Initialize list of errors.
var dErrs = make([]error, len(xl.storageDisks))
for index, disk := range xl.storageDisks {
if disk == nil {
dErrs[index] = traceError(errDiskNotFound)
continue
}
wg.Add(1)
go func(index int, disk StorageAPI) {
defer wg.Done()
err := cleanupDir(disk, bucket, object)
if err != nil && errorCause(err) != errVolumeNotFound {
dErrs[index] = err
}
}(index, disk)
}
// Wait for all routines to finish.
wg.Wait()
return reduceWriteQuorumErrs(dErrs, objectOpIgnoredErrs, xl.writeQuorum)
}
// DeleteObject - deletes an object, this call doesn't necessary reply
// any error as it is not necessary for the handler to reply back a
// response to the client request.
func (xl xlObjects) DeleteObject(bucket, object string) (err error) {
if err = checkDelObjArgs(bucket, object); err != nil {
return err
}
// Validate object exists.
if !xl.isObject(bucket, object) {
return traceError(ObjectNotFound{bucket, object})
} // else proceed to delete the object.
// Delete the object on all disks.
err = xl.deleteObject(bucket, object)
if err != nil {
return toObjectErr(err, bucket, object)
}
if xl.objCacheEnabled {
// Delete from the cache.
xl.objCache.Delete(pathJoin(bucket, object))
}
// Success.
return nil
}