The official AWS SDK provides an upload manager construct that allows you to upload to S3 from any io.Reader. Using it is straightforward, that is until you need to create and upload a potentially large ZIP file.
The solution: use the upload manager with a pipe.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
| // Open a pipe.
pr, pw := io.Pipe()
errch := make(chan error)
// Upload from pr in a separate Go routine.
go func() {
_, err := s3manager.NewUploader(awssess).Upload(&s3manager.UploadInput{
Bucket: aws.String("brain-bucket"),
Key: aws.String("blobs/very-large.zip"),
Body: pr,
})
errch <- err
}()
// Create a ZIP writer around pw.
zw := zip.NewWriter(pw)
// Add stuff to zip.
zw.Close() // Finishes the ZIP.
pw.Close() // Closes pw, marks EOF in pr.
err := <-errch // If err == nil, success.
|
For convenience, I have wrapped this up in a Go package: github.com/hjr265/s3ow [ GitHub ].