seaweedfs/weed/remote_storage/s3/s3_storage_client.go

229 lines
6.6 KiB
Go
Raw Normal View History

2021-07-27 05:53:44 +00:00
package s3
import (
"fmt"
"github.com/aws/aws-sdk-go/aws"
"github.com/aws/aws-sdk-go/aws/credentials"
"github.com/aws/aws-sdk-go/aws/session"
"github.com/aws/aws-sdk-go/service/s3"
"github.com/aws/aws-sdk-go/service/s3/s3iface"
2021-08-01 05:39:38 +00:00
"github.com/aws/aws-sdk-go/service/s3/s3manager"
2021-08-08 08:21:42 +00:00
"github.com/chrislusf/seaweedfs/weed/filer"
2021-07-27 05:53:44 +00:00
"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
2021-08-26 22:18:34 +00:00
"github.com/chrislusf/seaweedfs/weed/pb/remote_pb"
2021-07-27 05:53:44 +00:00
"github.com/chrislusf/seaweedfs/weed/remote_storage"
"github.com/chrislusf/seaweedfs/weed/util"
2021-08-08 08:21:42 +00:00
"io"
2021-08-16 02:27:30 +00:00
"reflect"
2021-07-27 05:53:44 +00:00
)
func init() {
remote_storage.RemoteStorageClientMakers["s3"] = new(s3RemoteStorageMaker)
}
type s3RemoteStorageMaker struct{}
func (s s3RemoteStorageMaker) HasBucket() bool {
return true
}
2021-08-26 22:18:34 +00:00
func (s s3RemoteStorageMaker) Make(conf *remote_pb.RemoteConf) (remote_storage.RemoteStorageClient, error) {
2021-07-27 05:53:44 +00:00
client := &s3RemoteStorageClient{
conf: conf,
}
config := &aws.Config{
Region: aws.String(conf.S3Region),
Endpoint: aws.String(conf.S3Endpoint),
S3ForcePathStyle: aws.Bool(conf.S3ForcePathStyle),
S3DisableContentMD5Validation: aws.Bool(true),
2021-07-27 05:53:44 +00:00
}
if conf.S3AccessKey != "" && conf.S3SecretKey != "" {
config.Credentials = credentials.NewStaticCredentials(conf.S3AccessKey, conf.S3SecretKey, "")
}
sess, err := session.NewSession(config)
if err != nil {
return nil, fmt.Errorf("create aws session: %v", err)
}
sess.Handlers.Build.PushFront(skipSha256PayloadSigning)
2021-07-27 05:53:44 +00:00
client.conn = s3.New(sess)
return client, nil
}
type s3RemoteStorageClient struct {
2021-08-26 22:18:34 +00:00
conf *remote_pb.RemoteConf
2021-07-27 05:53:44 +00:00
conn s3iface.S3API
}
2021-08-08 08:21:42 +00:00
var _ = remote_storage.RemoteStorageClient(&s3RemoteStorageClient{})
2021-08-26 22:18:34 +00:00
func (s *s3RemoteStorageClient) Traverse(remote *remote_pb.RemoteStorageLocation, visitFn remote_storage.VisitFunc) (err error) {
2021-07-27 10:26:35 +00:00
pathKey := remote.Path[1:]
2021-07-27 05:53:44 +00:00
listInput := &s3.ListObjectsV2Input{
Bucket: aws.String(remote.Bucket),
2021-07-27 05:53:44 +00:00
ContinuationToken: nil,
Delimiter: nil, // not aws.String("/"), iterate through all entries
EncodingType: nil,
ExpectedBucketOwner: nil,
FetchOwner: nil,
MaxKeys: nil, // aws.Int64(1000),
2021-07-27 10:32:24 +00:00
Prefix: aws.String(pathKey),
2021-07-27 05:53:44 +00:00
RequestPayer: nil,
StartAfter: nil,
}
isLastPage := false
for !isLastPage && err == nil {
listErr := s.conn.ListObjectsV2Pages(listInput, func(page *s3.ListObjectsV2Output, lastPage bool) bool {
for _, content := range page.Contents {
2021-08-01 05:39:38 +00:00
key := *content.Key
2021-08-15 08:53:46 +00:00
key = "/" + key
2021-07-27 10:26:35 +00:00
dir, name := util.FullPath(key).DirAndName()
2021-07-27 05:53:44 +00:00
if err := visitFn(dir, name, false, &filer_pb.RemoteEntry{
2021-08-09 21:35:18 +00:00
RemoteMtime: (*content.LastModified).Unix(),
RemoteSize: *content.Size,
RemoteETag: *content.ETag,
StorageName: s.conf.Name,
2021-07-27 05:53:44 +00:00
}); err != nil {
return false
}
}
listInput.ContinuationToken = page.NextContinuationToken
isLastPage = lastPage
return true
})
if listErr != nil {
2021-07-27 10:26:35 +00:00
err = fmt.Errorf("list %v: %v", remote, listErr)
2021-07-27 05:53:44 +00:00
}
}
return
}
2021-08-26 22:18:34 +00:00
func (s *s3RemoteStorageClient) ReadFile(loc *remote_pb.RemoteStorageLocation, offset int64, size int64) (data []byte, err error) {
2021-08-01 05:39:38 +00:00
downloader := s3manager.NewDownloaderWithClient(s.conn, func(u *s3manager.Downloader) {
u.PartSize = int64(4 * 1024 * 1024)
u.Concurrency = 1
})
2021-08-08 08:21:42 +00:00
2021-08-01 05:39:38 +00:00
dataSlice := make([]byte, int(size))
writerAt := aws.NewWriteAtBuffer(dataSlice)
_, err = downloader.Download(writerAt, &s3.GetObjectInput{
2021-08-08 08:21:42 +00:00
Bucket: aws.String(loc.Bucket),
Key: aws.String(loc.Path[1:]),
Range: aws.String(fmt.Sprintf("bytes=%d-%d", offset, offset+size-1)),
2021-08-01 05:39:38 +00:00
})
if err != nil {
return nil, fmt.Errorf("failed to download file %s%s: %v", loc.Bucket, loc.Path, err)
}
return writerAt.Bytes(), nil
}
2021-08-08 08:21:42 +00:00
2021-08-26 22:18:34 +00:00
func (s *s3RemoteStorageClient) WriteDirectory(loc *remote_pb.RemoteStorageLocation, entry *filer_pb.Entry) (err error) {
2021-08-09 21:35:18 +00:00
return nil
}
func (s *s3RemoteStorageClient) RemoveDirectory(loc *remote_pb.RemoteStorageLocation) (err error) {
return nil
}
2021-08-26 22:18:34 +00:00
func (s *s3RemoteStorageClient) WriteFile(loc *remote_pb.RemoteStorageLocation, entry *filer_pb.Entry, reader io.Reader) (remoteEntry *filer_pb.RemoteEntry, err error) {
2021-08-08 08:21:42 +00:00
fileSize := int64(filer.FileSize(entry))
partSize := int64(8 * 1024 * 1024) // The minimum/default allowed part size is 5MB
for partSize*1000 < fileSize {
partSize *= 4
}
// Create an uploader with the session and custom options
uploader := s3manager.NewUploaderWithClient(s.conn, func(u *s3manager.Uploader) {
u.PartSize = partSize
u.Concurrency = 1
2021-08-08 08:21:42 +00:00
})
// process tagging
tags := ""
for k, v := range entry.Extended {
if len(tags) > 0 {
tags = tags + "&"
}
tags = tags + k + "=" + string(v)
}
// Upload the file to S3.
_, err = uploader.Upload(&s3manager.UploadInput{
2021-08-23 09:18:56 +00:00
Bucket: aws.String(loc.Bucket),
Key: aws.String(loc.Path[1:]),
Body: reader,
Tagging: aws.String(tags),
StorageClass: aws.String(s.conf.S3StorageClass),
2021-08-08 08:21:42 +00:00
})
//in case it fails to upload
if err != nil {
return nil, fmt.Errorf("upload to %s/%s%s: %v", loc.Name, loc.Bucket, loc.Path, err)
2021-08-08 08:21:42 +00:00
}
// read back the remote entry
return s.readFileRemoteEntry(loc)
2021-08-08 08:21:42 +00:00
}
func toTagging(attributes map[string][]byte) *s3.Tagging {
tagging := &s3.Tagging{}
for k, v := range attributes {
tagging.TagSet = append(tagging.TagSet, &s3.Tag{
Key: aws.String(k),
Value: aws.String(string(v)),
})
}
return tagging
}
2021-08-26 22:18:34 +00:00
func (s *s3RemoteStorageClient) readFileRemoteEntry(loc *remote_pb.RemoteStorageLocation) (*filer_pb.RemoteEntry, error) {
resp, err := s.conn.HeadObject(&s3.HeadObjectInput{
2021-08-09 05:30:36 +00:00
Bucket: aws.String(loc.Bucket),
Key: aws.String(loc.Path[1:]),
})
if err != nil {
return nil, err
}
2021-08-09 05:30:36 +00:00
return &filer_pb.RemoteEntry{
2021-08-09 21:35:18 +00:00
RemoteMtime: resp.LastModified.Unix(),
RemoteSize: *resp.ContentLength,
RemoteETag: *resp.ETag,
StorageName: s.conf.Name,
}, nil
}
2021-08-26 22:18:34 +00:00
func (s *s3RemoteStorageClient) UpdateFileMetadata(loc *remote_pb.RemoteStorageLocation, oldEntry *filer_pb.Entry, newEntry *filer_pb.Entry) (err error) {
2021-08-16 02:27:30 +00:00
if reflect.DeepEqual(oldEntry.Extended, newEntry.Extended) {
return nil
}
tagging := toTagging(newEntry.Extended)
2021-08-08 08:21:42 +00:00
if len(tagging.TagSet) > 0 {
_, err = s.conn.PutObjectTagging(&s3.PutObjectTaggingInput{
Bucket: aws.String(loc.Bucket),
Key: aws.String(loc.Path[1:]),
2021-08-16 02:27:30 +00:00
Tagging: toTagging(newEntry.Extended),
2021-08-08 08:21:42 +00:00
})
} else {
_, err = s.conn.DeleteObjectTagging(&s3.DeleteObjectTaggingInput{
2021-08-09 05:30:36 +00:00
Bucket: aws.String(loc.Bucket),
Key: aws.String(loc.Path[1:]),
2021-08-08 08:21:42 +00:00
})
}
return
}
2021-08-26 22:18:34 +00:00
func (s *s3RemoteStorageClient) DeleteFile(loc *remote_pb.RemoteStorageLocation) (err error) {
2021-08-08 08:21:42 +00:00
_, err = s.conn.DeleteObject(&s3.DeleteObjectInput{
Bucket: aws.String(loc.Bucket),
Key: aws.String(loc.Path[1:]),
})
return
}