/* * Minio Cloud Storage, (C) 2015 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 xl import ( "bytes" "crypto/md5" "crypto/sha512" "encoding/base64" "encoding/hex" "encoding/xml" "io" "io/ioutil" "math/rand" "runtime/debug" "sort" "strconv" "strings" "time" "github.com/minio/minio/pkg/crypto/sha256" "github.com/minio/minio/pkg/probe" signV4 "github.com/minio/minio/pkg/signature" "github.com/minio/minio/pkg/xl/cache/data" ) /// V2 API functions // NewMultipartUpload - initiate a new multipart session func (xl API) NewMultipartUpload(bucket, key, contentType string) (string, *probe.Error) { xl.lock.Lock() defer xl.lock.Unlock() if !IsValidBucket(bucket) { return "", probe.NewError(BucketNameInvalid{Bucket: bucket}) } if !IsValidObjectName(key) { return "", probe.NewError(ObjectNameInvalid{Object: key}) } // if len(xl.config.NodeDiskMap) > 0 { // return xl.newMultipartUpload(bucket, key, contentType) // } if !xl.storedBuckets.Exists(bucket) { return "", probe.NewError(BucketNotFound{Bucket: bucket}) } storedBucket := xl.storedBuckets.Get(bucket).(storedBucket) objectKey := bucket + "/" + key if _, ok := storedBucket.objectMetadata[objectKey]; ok == true { return "", probe.NewError(ObjectExists{Object: key}) } id := []byte(strconv.Itoa(rand.Int()) + bucket + key + time.Now().UTC().String()) uploadIDSum := sha512.Sum512(id) uploadID := base64.URLEncoding.EncodeToString(uploadIDSum[:])[:47] storedBucket.multiPartSession[key] = MultiPartSession{ UploadID: uploadID, Initiated: time.Now().UTC(), TotalParts: 0, } storedBucket.partMetadata[key] = make(map[int]PartMetadata) multiPartCache := data.NewCache(0) multiPartCache.OnEvicted = xl.evictedPart xl.multiPartObjects[uploadID] = multiPartCache xl.storedBuckets.Set(bucket, storedBucket) return uploadID, nil } // AbortMultipartUpload - abort an incomplete multipart session func (xl API) AbortMultipartUpload(bucket, key, uploadID string) *probe.Error { xl.lock.Lock() defer xl.lock.Unlock() if !IsValidBucket(bucket) { return probe.NewError(BucketNameInvalid{Bucket: bucket}) } if !IsValidObjectName(key) { return probe.NewError(ObjectNameInvalid{Object: key}) } // TODO: multipart support for xl is broken, since we haven't finalized the format in which // it can be stored, disabling this for now until we get the underlying layout stable. // // if len(xl.config.NodeDiskMap) > 0 { // return xl.abortMultipartUpload(bucket, key, uploadID) // } if !xl.storedBuckets.Exists(bucket) { return probe.NewError(BucketNotFound{Bucket: bucket}) } storedBucket := xl.storedBuckets.Get(bucket).(storedBucket) if storedBucket.multiPartSession[key].UploadID != uploadID { return probe.NewError(InvalidUploadID{UploadID: uploadID}) } xl.cleanupMultipartSession(bucket, key, uploadID) return nil } // CreateObjectPart - create a part in a multipart session func (xl API) CreateObjectPart(bucket, key, uploadID string, partID int, contentType, expectedMD5Sum string, size int64, data io.Reader, signature *signV4.Signature) (string, *probe.Error) { xl.lock.Lock() etag, err := xl.createObjectPart(bucket, key, uploadID, partID, "", expectedMD5Sum, size, data, signature) xl.lock.Unlock() // possible free debug.FreeOSMemory() return etag, err.Trace() } // createObject - internal wrapper function called by CreateObjectPart func (xl API) createObjectPart(bucket, key, uploadID string, partID int, contentType, expectedMD5Sum string, size int64, data io.Reader, signature *signV4.Signature) (string, *probe.Error) { if !IsValidBucket(bucket) { return "", probe.NewError(BucketNameInvalid{Bucket: bucket}) } if !IsValidObjectName(key) { return "", probe.NewError(ObjectNameInvalid{Object: key}) } // TODO: multipart support for xl is broken, since we haven't finalized the format in which // it can be stored, disabling this for now until we get the underlying layout stable. // /* if len(xl.config.NodeDiskMap) > 0 { metadata := make(map[string]string) if contentType == "" { contentType = "application/octet-stream" } contentType = strings.TrimSpace(contentType) metadata["contentType"] = contentType if strings.TrimSpace(expectedMD5Sum) != "" { expectedMD5SumBytes, err := base64.StdEncoding.DecodeString(strings.TrimSpace(expectedMD5Sum)) if err != nil { // pro-actively close the connection return "", probe.NewError(InvalidDigest{Md5: expectedMD5Sum}) } expectedMD5Sum = hex.EncodeToString(expectedMD5SumBytes) } partMetadata, err := xl.putObjectPart(bucket, key, expectedMD5Sum, uploadID, partID, data, size, metadata, signature) if err != nil { return "", err.Trace() } return partMetadata.ETag, nil } */ if !xl.storedBuckets.Exists(bucket) { return "", probe.NewError(BucketNotFound{Bucket: bucket}) } strBucket := xl.storedBuckets.Get(bucket).(storedBucket) // Verify upload id if strBucket.multiPartSession[key].UploadID != uploadID { return "", probe.NewError(InvalidUploadID{UploadID: uploadID}) } // get object key parts := strBucket.partMetadata[key] if _, ok := parts[partID]; ok { return parts[partID].ETag, nil } if contentType == "" { contentType = "application/octet-stream" } contentType = strings.TrimSpace(contentType) if strings.TrimSpace(expectedMD5Sum) != "" { expectedMD5SumBytes, err := base64.StdEncoding.DecodeString(strings.TrimSpace(expectedMD5Sum)) if err != nil { // pro-actively close the connection return "", probe.NewError(InvalidDigest{Md5: expectedMD5Sum}) } expectedMD5Sum = hex.EncodeToString(expectedMD5SumBytes) } // calculate md5 hash := md5.New() sha256hash := sha256.New() var totalLength int64 var err error for err == nil { var length int byteBuffer := make([]byte, 1024*1024) length, err = data.Read(byteBuffer) // do not read error return error here, we will handle this error later if length != 0 { hash.Write(byteBuffer[0:length]) sha256hash.Write(byteBuffer[0:length]) ok := xl.multiPartObjects[uploadID].Append(partID, byteBuffer[0:length]) if !ok { return "", probe.NewError(InternalError{}) } totalLength += int64(length) go debug.FreeOSMemory() } } if totalLength != size { xl.multiPartObjects[uploadID].Delete(partID) return "", probe.NewError(IncompleteBody{Bucket: bucket, Object: key}) } if err != io.EOF { return "", probe.NewError(err) } md5SumBytes := hash.Sum(nil) md5Sum := hex.EncodeToString(md5SumBytes) // Verify if the written object is equal to what is expected, only if it is requested as such if strings.TrimSpace(expectedMD5Sum) != "" { if err := isMD5SumEqual(strings.TrimSpace(expectedMD5Sum), md5Sum); err != nil { return "", err.Trace() } } if signature != nil { { ok, err := signature.DoesSignatureMatch(hex.EncodeToString(sha256hash.Sum(nil))) if err != nil { return "", err.Trace() } if !ok { return "", probe.NewError(signV4.SigDoesNotMatch{}) } } } newPart := PartMetadata{ PartNumber: partID, LastModified: time.Now().UTC(), ETag: md5Sum, Size: totalLength, } parts[partID] = newPart strBucket.partMetadata[key] = parts multiPartSession := strBucket.multiPartSession[key] multiPartSession.TotalParts++ strBucket.multiPartSession[key] = multiPartSession xl.storedBuckets.Set(bucket, strBucket) return md5Sum, nil } // cleanupMultipartSession invoked during an abort or complete multipart session to cleanup session from memory func (xl API) cleanupMultipartSession(bucket, key, uploadID string) { storedBucket := xl.storedBuckets.Get(bucket).(storedBucket) for i := 1; i <= storedBucket.multiPartSession[key].TotalParts; i++ { xl.multiPartObjects[uploadID].Delete(i) } delete(storedBucket.multiPartSession, key) delete(storedBucket.partMetadata, key) xl.storedBuckets.Set(bucket, storedBucket) } func (xl API) mergeMultipart(parts *CompleteMultipartUpload, uploadID string, fullObjectWriter *io.PipeWriter) { for _, part := range parts.Part { recvMD5 := part.ETag object, ok := xl.multiPartObjects[uploadID].Get(part.PartNumber) if ok == false { fullObjectWriter.CloseWithError(probe.WrapError(probe.NewError(InvalidPart{}))) return } calcMD5Bytes := md5.Sum(object) // complete multi part request header md5sum per part is hex encoded recvMD5Bytes, err := hex.DecodeString(strings.Trim(recvMD5, "\"")) if err != nil { fullObjectWriter.CloseWithError(probe.WrapError(probe.NewError(InvalidDigest{Md5: recvMD5}))) return } if !bytes.Equal(recvMD5Bytes, calcMD5Bytes[:]) { fullObjectWriter.CloseWithError(probe.WrapError(probe.NewError(BadDigest{}))) return } if _, err := io.Copy(fullObjectWriter, bytes.NewReader(object)); err != nil { fullObjectWriter.CloseWithError(probe.WrapError(probe.NewError(err))) return } object = nil } fullObjectWriter.Close() return } // CompleteMultipartUpload - complete a multipart upload and persist the data func (xl API) CompleteMultipartUpload(bucket, key, uploadID string, data io.Reader, signature *signV4.Signature) (ObjectMetadata, *probe.Error) { xl.lock.Lock() defer xl.lock.Unlock() size := int64(xl.multiPartObjects[uploadID].Stats().Bytes) fullObjectReader, err := xl.completeMultipartUploadV2(bucket, key, uploadID, data, signature) if err != nil { return ObjectMetadata{}, err.Trace() } objectMetadata, err := xl.createObject(bucket, key, "", "", size, fullObjectReader, nil) if err != nil { // No need to call internal cleanup functions here, caller should call AbortMultipartUpload() // which would in-turn cleanup properly in accordance with S3 Spec return ObjectMetadata{}, err.Trace() } xl.cleanupMultipartSession(bucket, key, uploadID) return objectMetadata, nil } func (xl API) completeMultipartUploadV2(bucket, key, uploadID string, data io.Reader, signature *signV4.Signature) (io.Reader, *probe.Error) { if !IsValidBucket(bucket) { return nil, probe.NewError(BucketNameInvalid{Bucket: bucket}) } if !IsValidObjectName(key) { return nil, probe.NewError(ObjectNameInvalid{Object: key}) } // TODO: multipart support for xl is broken, since we haven't finalized the format in which // it can be stored, disabling this for now until we get the underlying layout stable. // // if len(xl.config.NodeDiskMap) > 0 { // xl.lock.Unlock() // return xl.completeMultipartUpload(bucket, key, uploadID, data, signature) // } if !xl.storedBuckets.Exists(bucket) { return nil, probe.NewError(BucketNotFound{Bucket: bucket}) } storedBucket := xl.storedBuckets.Get(bucket).(storedBucket) // Verify upload id if storedBucket.multiPartSession[key].UploadID != uploadID { return nil, probe.NewError(InvalidUploadID{UploadID: uploadID}) } partBytes, err := ioutil.ReadAll(data) if err != nil { return nil, probe.NewError(err) } if signature != nil { partHashBytes := sha256.Sum256(partBytes) ok, err := signature.DoesSignatureMatch(hex.EncodeToString(partHashBytes[:])) if err != nil { return nil, err.Trace() } if !ok { return nil, probe.NewError(signV4.SigDoesNotMatch{}) } } parts := &CompleteMultipartUpload{} if err := xml.Unmarshal(partBytes, parts); err != nil { return nil, probe.NewError(MalformedXML{}) } if !sort.IsSorted(completedParts(parts.Part)) { return nil, probe.NewError(InvalidPartOrder{}) } fullObjectReader, fullObjectWriter := io.Pipe() go xl.mergeMultipart(parts, uploadID, fullObjectWriter) return fullObjectReader, nil } // byKey is a sortable interface for UploadMetadata slice type byKey []*UploadMetadata func (a byKey) Len() int { return len(a) } func (a byKey) Swap(i, j int) { a[i], a[j] = a[j], a[i] } func (a byKey) Less(i, j int) bool { return a[i].Key < a[j].Key } // ListMultipartUploads - list incomplete multipart sessions for a given bucket func (xl API) ListMultipartUploads(bucket string, resources BucketMultipartResourcesMetadata) (BucketMultipartResourcesMetadata, *probe.Error) { // TODO handle delimiter, low priority xl.lock.Lock() defer xl.lock.Unlock() if !IsValidBucket(bucket) { return BucketMultipartResourcesMetadata{}, probe.NewError(BucketNameInvalid{Bucket: bucket}) } // TODO: multipart support for xl is broken, since we haven't finalized the format in which // it can be stored, disabling this for now until we get the underlying layout stable. // // if len(xl.config.NodeDiskMap) > 0 { // return xl.listMultipartUploads(bucket, resources) // } if !xl.storedBuckets.Exists(bucket) { return BucketMultipartResourcesMetadata{}, probe.NewError(BucketNotFound{Bucket: bucket}) } storedBucket := xl.storedBuckets.Get(bucket).(storedBucket) var uploads []*UploadMetadata for key, session := range storedBucket.multiPartSession { if strings.HasPrefix(key, resources.Prefix) { if len(uploads) > resources.MaxUploads { sort.Sort(byKey(uploads)) resources.Upload = uploads resources.NextKeyMarker = key resources.NextUploadIDMarker = session.UploadID resources.IsTruncated = true return resources, nil } // uploadIDMarker is ignored if KeyMarker is empty switch { case resources.KeyMarker != "" && resources.UploadIDMarker == "": if key > resources.KeyMarker { upload := new(UploadMetadata) upload.Key = key upload.UploadID = session.UploadID upload.Initiated = session.Initiated uploads = append(uploads, upload) } case resources.KeyMarker != "" && resources.UploadIDMarker != "": if session.UploadID > resources.UploadIDMarker { if key >= resources.KeyMarker { upload := new(UploadMetadata) upload.Key = key upload.UploadID = session.UploadID upload.Initiated = session.Initiated uploads = append(uploads, upload) } } default: upload := new(UploadMetadata) upload.Key = key upload.UploadID = session.UploadID upload.Initiated = session.Initiated uploads = append(uploads, upload) } } } sort.Sort(byKey(uploads)) resources.Upload = uploads return resources, nil } // partNumber is a sortable interface for Part slice type partNumber []*PartMetadata func (a partNumber) Len() int { return len(a) } func (a partNumber) Swap(i, j int) { a[i], a[j] = a[j], a[i] } func (a partNumber) Less(i, j int) bool { return a[i].PartNumber < a[j].PartNumber } // ListObjectParts - list parts from incomplete multipart session for a given object func (xl API) ListObjectParts(bucket, key string, resources ObjectResourcesMetadata) (ObjectResourcesMetadata, *probe.Error) { // Verify upload id xl.lock.Lock() defer xl.lock.Unlock() if !IsValidBucket(bucket) { return ObjectResourcesMetadata{}, probe.NewError(BucketNameInvalid{Bucket: bucket}) } if !IsValidObjectName(key) { return ObjectResourcesMetadata{}, probe.NewError(ObjectNameInvalid{Object: key}) } // TODO: multipart support for xl is broken, since we haven't finalized the format in which // it can be stored, disabling this for now until we get the underlying layout stable. // // if len(xl.config.NodeDiskMap) > 0 { // return xl.listObjectParts(bucket, key, resources) // } if !xl.storedBuckets.Exists(bucket) { return ObjectResourcesMetadata{}, probe.NewError(BucketNotFound{Bucket: bucket}) } storedBucket := xl.storedBuckets.Get(bucket).(storedBucket) if _, ok := storedBucket.multiPartSession[key]; ok == false { return ObjectResourcesMetadata{}, probe.NewError(ObjectNotFound{Object: key}) } if storedBucket.multiPartSession[key].UploadID != resources.UploadID { return ObjectResourcesMetadata{}, probe.NewError(InvalidUploadID{UploadID: resources.UploadID}) } storedParts := storedBucket.partMetadata[key] objectResourcesMetadata := resources objectResourcesMetadata.Bucket = bucket objectResourcesMetadata.Key = key var parts []*PartMetadata var startPartNumber int switch { case objectResourcesMetadata.PartNumberMarker == 0: startPartNumber = 1 default: startPartNumber = objectResourcesMetadata.PartNumberMarker } for i := startPartNumber; i <= storedBucket.multiPartSession[key].TotalParts; i++ { if len(parts) > objectResourcesMetadata.MaxParts { sort.Sort(partNumber(parts)) objectResourcesMetadata.IsTruncated = true objectResourcesMetadata.Part = parts objectResourcesMetadata.NextPartNumberMarker = i return objectResourcesMetadata, nil } part, ok := storedParts[i] if !ok { return ObjectResourcesMetadata{}, probe.NewError(InvalidPart{}) } parts = append(parts, &part) } sort.Sort(partNumber(parts)) objectResourcesMetadata.Part = parts return objectResourcesMetadata, nil } // evictedPart - call back function called by caching module during individual cache evictions func (xl API) evictedPart(a ...interface{}) { // loop through all buckets buckets := xl.storedBuckets.GetAll() for bucketName, bucket := range buckets { b := bucket.(storedBucket) xl.storedBuckets.Set(bucketName, b) } debug.FreeOSMemory() }