minio/object-api-multipart.go

578 lines
16 KiB
Go

/*
* Minio Cloud Storage, (C) 2016 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 main
import (
"crypto/md5"
"encoding/hex"
"fmt"
"io"
"io/ioutil"
"path"
"strconv"
"strings"
"github.com/minio/minio/pkg/probe"
"github.com/skyrings/skyring-common/tools/uuid"
)
const (
// Minio meta volume.
minioMetaVolume = ".minio"
)
// checks whether bucket exists.
func (o objectAPI) isBucketExist(bucketName string) (bool, error) {
// Check whether bucket exists.
if _, e := o.storage.StatVol(bucketName); e != nil {
if e == errVolumeNotFound {
return false, nil
}
return false, e
}
return true, nil
}
// checkLeafDirectory - verifies if a given path is leaf directory if
// yes returns all the files inside it.
func (o objectAPI) checkLeafDirectory(prefixPath string) (isLeaf bool, fis []FileInfo) {
var allFileInfos []FileInfo
var markerPath string
for {
fileInfos, eof, e := o.storage.ListFiles(minioMetaVolume, prefixPath, markerPath, false, 1000)
if e != nil {
break
}
allFileInfos = append(allFileInfos, fileInfos...)
if eof {
break
}
markerPath = allFileInfos[len(allFileInfos)-1].Name
}
for _, fileInfo := range allFileInfos {
if fileInfo.Mode.IsDir() {
isLeaf = false
return isLeaf, nil
}
fileName := path.Base(fileInfo.Name)
if !strings.Contains(fileName, ".") {
fis = append(fis, fileInfo)
}
}
isLeaf = true
return isLeaf, fis
}
// ListMultipartUploads - list multipart uploads.
func (o objectAPI) ListMultipartUploads(bucket, prefix, keyMarker, uploadIDMarker, delimiter string, maxUploads int) (ListMultipartsInfo, *probe.Error) {
result := ListMultipartsInfo{}
// Verify if bucket is valid.
if !IsValidBucketName(bucket) {
return ListMultipartsInfo{}, probe.NewError(BucketNameInvalid{Bucket: bucket})
}
if !IsValidObjectPrefix(prefix) {
return ListMultipartsInfo{}, probe.NewError(ObjectNameInvalid{Bucket: bucket, Object: prefix})
}
// Verify if delimiter is anything other than '/', which we do not support.
if delimiter != "" && delimiter != slashSeparator {
return ListMultipartsInfo{}, probe.NewError(UnsupportedDelimiter{
Delimiter: delimiter,
})
}
// Verify if marker has prefix.
if keyMarker != "" && !strings.HasPrefix(keyMarker, prefix) {
return ListMultipartsInfo{}, probe.NewError(InvalidMarkerPrefixCombination{
Marker: keyMarker,
Prefix: prefix,
})
}
if uploadIDMarker != "" {
if strings.HasSuffix(keyMarker, slashSeparator) {
return result, probe.NewError(InvalidUploadIDKeyCombination{
UploadIDMarker: uploadIDMarker,
KeyMarker: keyMarker,
})
}
id, e := uuid.Parse(uploadIDMarker)
if e != nil {
return result, probe.NewError(e)
}
if id.IsZero() {
return result, probe.NewError(MalformedUploadID{
UploadID: uploadIDMarker,
})
}
}
recursive := true
if delimiter == slashSeparator {
recursive = false
}
result.IsTruncated = true
result.MaxUploads = maxUploads
newMaxUploads := 0
// Not using path.Join() as it strips off the trailing '/'.
// Also bucket should always be followed by '/' even if prefix is empty.
prefixPath := pathJoin(bucket, prefix)
if recursive {
keyMarkerPath := ""
if keyMarker != "" {
keyMarkerPath = path.Join(bucket, keyMarker, uploadIDMarker)
}
outerLoop:
for {
fileInfos, eof, e := o.storage.ListFiles(minioMetaVolume, prefixPath, keyMarkerPath, recursive, maxUploads-newMaxUploads)
if e != nil {
return ListMultipartsInfo{}, probe.NewError(e)
}
for _, fi := range fileInfos {
keyMarkerPath = fi.Name
// fi.Name will look like bucket/object/uploadID, extract object and uploadID.
uploadID := path.Base(fi.Name)
objectName := strings.TrimPrefix(path.Dir(fi.Name), retainSlash(bucket))
if strings.Contains(uploadID, ".") {
// Contains partnumber and md5sum info, skip this.
continue
}
result.Uploads = append(result.Uploads, uploadMetadata{
Object: objectName,
UploadID: uploadID,
Initiated: fi.ModTime,
})
result.NextKeyMarker = objectName
result.NextUploadIDMarker = uploadID
newMaxUploads++
if newMaxUploads == maxUploads {
if eof {
result.IsTruncated = false
}
break outerLoop
}
}
if eof {
result.IsTruncated = false
break
}
}
if !result.IsTruncated {
result.NextKeyMarker = ""
result.NextUploadIDMarker = ""
}
return result, nil
}
var fileInfos []FileInfo
// read all the "fileInfos" in the prefix
for {
marker := ""
fis, eof, e := o.storage.ListFiles(minioMetaVolume, prefixPath, marker, recursive, 1000)
if e != nil {
return ListMultipartsInfo{}, probe.NewError(e)
}
for _, fi := range fis {
marker = fi.Name
if fi.Mode.IsDir() {
fileInfos = append(fileInfos, fi)
}
}
if eof {
break
}
}
// Create "uploads" slice from "fileInfos" slice.
var uploads []uploadMetadata
for _, fi := range fileInfos {
leaf, entries := o.checkLeafDirectory(fi.Name)
objectName := strings.TrimPrefix(fi.Name, retainSlash(bucket))
if leaf {
for _, entry := range entries {
if strings.Contains(path.Base(entry.Name), ".") {
continue
}
uploads = append(uploads, uploadMetadata{
Object: path.Dir(objectName),
UploadID: path.Base(entry.Name),
Initiated: entry.ModTime,
})
}
continue
}
uploads = append(uploads, uploadMetadata{
Object: objectName,
})
}
index := 0
for i, upload := range uploads {
index = i
if upload.Object > keyMarker {
break
}
if uploads[index].Object == keyMarker && upload.UploadID > uploadIDMarker {
break
}
}
for ; index < len(uploads); index++ {
if (len(result.Uploads) + len(result.CommonPrefixes)) == maxUploads {
break
}
result.NextKeyMarker = uploads[index].Object
if strings.HasSuffix(uploads[index].Object, slashSeparator) {
// for a directory entry
result.CommonPrefixes = append(result.CommonPrefixes, uploads[index].Object)
continue
}
result.NextUploadIDMarker = uploads[index].UploadID
result.Uploads = append(result.Uploads, uploads[index])
}
if index == len(uploads) {
result.IsTruncated = false
result.NextKeyMarker = ""
result.NextUploadIDMarker = ""
}
return result, nil
}
func (o objectAPI) NewMultipartUpload(bucket, object string) (string, *probe.Error) {
// Verify if bucket name is valid.
if !IsValidBucketName(bucket) {
return "", probe.NewError(BucketNameInvalid{Bucket: bucket})
}
// Verify if object name is valid.
if !IsValidObjectName(object) {
return "", probe.NewError(ObjectNameInvalid{Bucket: bucket, Object: object})
}
// Verify whether the bucket exists.
isExist, err := o.isBucketExist(bucket)
if err != nil {
return "", probe.NewError(err)
}
if !isExist {
return "", probe.NewError(BucketNotFound{Bucket: bucket})
}
if _, e := o.storage.StatVol(minioMetaVolume); e != nil {
if e == errVolumeNotFound {
e = o.storage.MakeVol(minioMetaVolume)
if e != nil {
if e == errDiskFull {
return "", probe.NewError(StorageFull{})
}
return "", probe.NewError(e)
}
}
}
for {
uuid, e := uuid.New()
if e != nil {
return "", probe.NewError(e)
}
uploadID := uuid.String()
uploadIDPath := path.Join(bucket, object, uploadID)
if _, e = o.storage.StatFile(minioMetaVolume, uploadIDPath); e != nil {
if e != errFileNotFound {
return "", probe.NewError(toObjectErr(e, minioMetaVolume, uploadIDPath))
}
// uploadIDPath doesn't exist, so create empty file to reserve the name
var w io.WriteCloser
if w, e = o.storage.CreateFile(minioMetaVolume, uploadIDPath); e == nil {
// Close the writer.
if e = w.Close(); e != nil {
return "", probe.NewError(e)
}
} else {
return "", probe.NewError(toObjectErr(e, minioMetaVolume, uploadIDPath))
}
return uploadID, nil
}
// uploadIDPath already exists.
// loop again to try with different uuid generated.
}
}
// isUploadIDExists - verify if a given uploadID exists and is valid.
func (o objectAPI) isUploadIDExists(bucket, object, uploadID string) (bool, error) {
uploadIDPath := path.Join(bucket, object, uploadID)
st, e := o.storage.StatFile(minioMetaVolume, uploadIDPath)
if e != nil {
// Upload id does not exist.
if e == errFileNotFound {
return false, nil
}
return false, e
}
// Upload id exists and is a regular file.
return st.Mode.IsRegular(), nil
}
// PutObjectPart - writes the multipart upload chunks.
func (o objectAPI) PutObjectPart(bucket, object, uploadID string, partID int, size int64, data io.Reader, md5Hex string) (string, *probe.Error) {
// Verify if bucket is valid.
if !IsValidBucketName(bucket) {
return "", probe.NewError(BucketNameInvalid{Bucket: bucket})
}
if !IsValidObjectName(object) {
return "", probe.NewError(ObjectNameInvalid{Bucket: bucket, Object: object})
}
// Verify whether the bucket exists.
isExist, err := o.isBucketExist(bucket)
if err != nil {
return "", probe.NewError(err)
}
if !isExist {
return "", probe.NewError(BucketNotFound{Bucket: bucket})
}
if status, e := o.isUploadIDExists(bucket, object, uploadID); e != nil {
return "", probe.NewError(e)
} else if !status {
return "", probe.NewError(InvalidUploadID{UploadID: uploadID})
}
partSuffix := fmt.Sprintf("%s.%d.%s", uploadID, partID, md5Hex)
fileWriter, e := o.storage.CreateFile(minioMetaVolume, path.Join(bucket, object, partSuffix))
if e != nil {
return "", probe.NewError(toObjectErr(e, bucket, object))
}
// Initialize md5 writer.
md5Writer := md5.New()
// Instantiate a new multi writer.
multiWriter := io.MultiWriter(md5Writer, fileWriter)
// Instantiate checksum hashers and create a multiwriter.
if size > 0 {
if _, e = io.CopyN(multiWriter, data, size); e != nil {
safeCloseAndRemove(fileWriter)
return "", probe.NewError(toObjectErr(e))
}
// Reader shouldn't have more data what mentioned in size argument.
// reading one more byte from the reader to validate it.
// expected to fail, success validates existence of more data in the reader.
if _, e = io.CopyN(ioutil.Discard, data, 1); e == nil {
safeCloseAndRemove(fileWriter)
return "", probe.NewError(UnExpectedDataSize{Size: int(size)})
}
} else {
if _, e = io.Copy(multiWriter, data); e != nil {
safeCloseAndRemove(fileWriter)
return "", probe.NewError(toObjectErr(e))
}
}
newMD5Hex := hex.EncodeToString(md5Writer.Sum(nil))
if md5Hex != "" {
if newMD5Hex != md5Hex {
safeCloseAndRemove(fileWriter)
return "", probe.NewError(BadDigest{md5Hex, newMD5Hex})
}
}
e = fileWriter.Close()
if e != nil {
return "", probe.NewError(e)
}
return newMD5Hex, nil
}
func (o objectAPI) ListObjectParts(bucket, object, uploadID string, partNumberMarker, maxParts int) (ListPartsInfo, *probe.Error) {
// Verify if bucket is valid.
if !IsValidBucketName(bucket) {
return ListPartsInfo{}, probe.NewError(BucketNameInvalid{Bucket: bucket})
}
if !IsValidObjectName(object) {
return ListPartsInfo{}, probe.NewError(ObjectNameInvalid{Bucket: bucket, Object: object})
}
if status, e := o.isUploadIDExists(bucket, object, uploadID); e != nil {
return ListPartsInfo{}, probe.NewError(e)
} else if !status {
return ListPartsInfo{}, probe.NewError(InvalidUploadID{UploadID: uploadID})
}
result := ListPartsInfo{}
var markerPath string
nextPartNumberMarker := 0
uploadIDPath := path.Join(bucket, object, uploadID)
// Figure out the marker for the next subsequent calls, if the
// partNumberMarker is already set.
if partNumberMarker > 0 {
partNumberMarkerPath := uploadIDPath + "." + strconv.Itoa(partNumberMarker) + "."
fileInfos, _, e := o.storage.ListFiles(minioMetaVolume, partNumberMarkerPath, "", false, 1)
if e != nil {
return result, probe.NewError(toObjectErr(e, minioMetaVolume, partNumberMarkerPath))
}
if len(fileInfos) == 0 {
return result, probe.NewError(InvalidPart{})
}
markerPath = fileInfos[0].Name
}
uploadIDPrefix := uploadIDPath + "."
fileInfos, eof, e := o.storage.ListFiles(minioMetaVolume, uploadIDPrefix, markerPath, false, maxParts)
if e != nil {
return result, probe.NewError(InvalidPart{})
}
for _, fileInfo := range fileInfos {
fileName := path.Base(fileInfo.Name)
splitResult := strings.Split(fileName, ".")
partNum, e := strconv.Atoi(splitResult[1])
if e != nil {
return result, probe.NewError(e)
}
md5sum := splitResult[2]
result.Parts = append(result.Parts, partInfo{
PartNumber: partNum,
LastModified: fileInfo.ModTime,
ETag: md5sum,
Size: fileInfo.Size,
})
nextPartNumberMarker = partNum
}
result.Bucket = bucket
result.Object = object
result.UploadID = uploadID
result.PartNumberMarker = partNumberMarker
result.NextPartNumberMarker = nextPartNumberMarker
result.MaxParts = maxParts
result.IsTruncated = !eof
return result, nil
}
// Create an s3 compatible MD5sum for complete multipart transaction.
func makeS3MD5(md5Strs ...string) (string, *probe.Error) {
var finalMD5Bytes []byte
for _, md5Str := range md5Strs {
md5Bytes, e := hex.DecodeString(md5Str)
if e != nil {
return "", probe.NewError(e)
}
finalMD5Bytes = append(finalMD5Bytes, md5Bytes...)
}
md5Hasher := md5.New()
md5Hasher.Write(finalMD5Bytes)
s3MD5 := fmt.Sprintf("%s-%d", hex.EncodeToString(md5Hasher.Sum(nil)), len(md5Strs))
return s3MD5, nil
}
func (o objectAPI) CompleteMultipartUpload(bucket string, object string, uploadID string, parts []completePart) (string, *probe.Error) {
// Verify if bucket is valid.
if !IsValidBucketName(bucket) {
return "", probe.NewError(BucketNameInvalid{Bucket: bucket})
}
if !IsValidObjectName(object) {
return "", probe.NewError(ObjectNameInvalid{
Bucket: bucket,
Object: object,
})
}
if status, e := o.isUploadIDExists(bucket, object, uploadID); e != nil {
return "", probe.NewError(e)
} else if !status {
return "", probe.NewError(InvalidUploadID{UploadID: uploadID})
}
fileWriter, e := o.storage.CreateFile(bucket, object)
if e != nil {
return "", probe.NewError(toObjectErr(e, bucket, object))
}
var md5Sums []string
for _, part := range parts {
// Construct part suffix.
partSuffix := fmt.Sprintf("%s.%d.%s", uploadID, part.PartNumber, part.ETag)
var fileReader io.ReadCloser
fileReader, e = o.storage.ReadFile(minioMetaVolume, path.Join(bucket, object, partSuffix), 0)
if e != nil {
if e == errFileNotFound {
return "", probe.NewError(InvalidPart{})
}
return "", probe.NewError(e)
}
_, e = io.Copy(fileWriter, fileReader)
if e != nil {
return "", probe.NewError(e)
}
e = fileReader.Close()
if e != nil {
return "", probe.NewError(e)
}
md5Sums = append(md5Sums, part.ETag)
}
e = fileWriter.Close()
if e != nil {
return "", probe.NewError(e)
}
// Save the s3 md5.
s3MD5, err := makeS3MD5(md5Sums...)
if err != nil {
return "", err.Trace(md5Sums...)
}
// Cleanup all the parts.
o.removeMultipartUpload(bucket, object, uploadID)
// Return md5sum.
return s3MD5, nil
}
func (o objectAPI) removeMultipartUpload(bucket, object, uploadID string) *probe.Error {
// Verify if bucket is valid.
if !IsValidBucketName(bucket) {
return probe.NewError(BucketNameInvalid{Bucket: bucket})
}
if !IsValidObjectName(object) {
return probe.NewError(ObjectNameInvalid{Bucket: bucket, Object: object})
}
marker := ""
for {
uploadIDPath := path.Join(bucket, object, uploadID)
fileInfos, eof, e := o.storage.ListFiles(minioMetaVolume, uploadIDPath, marker, false, 1000)
if e != nil {
return probe.NewError(InvalidUploadID{UploadID: uploadID})
}
for _, fileInfo := range fileInfos {
o.storage.DeleteFile(minioMetaVolume, fileInfo.Name)
marker = fileInfo.Name
}
if eof {
break
}
}
return nil
}
func (o objectAPI) AbortMultipartUpload(bucket, object, uploadID string) *probe.Error {
// Verify if bucket is valid.
if !IsValidBucketName(bucket) {
return probe.NewError(BucketNameInvalid{Bucket: bucket})
}
if !IsValidObjectName(object) {
return probe.NewError(ObjectNameInvalid{Bucket: bucket, Object: object})
}
if status, e := o.isUploadIDExists(bucket, object, uploadID); e != nil {
return probe.NewError(e)
} else if !status {
return probe.NewError(InvalidUploadID{UploadID: uploadID})
}
err := o.removeMultipartUpload(bucket, object, uploadID)
if err != nil {
return err.Trace(bucket, object, uploadID)
}
return nil
}