Elek, Marton created HDDS-1213:
----------------------------------

             Summary: Support plain text S3 MPU initialization request
                 Key: HDDS-1213
                 URL: https://issues.apache.org/jira/browse/HDDS-1213
             Project: Hadoop Distributed Data Store
          Issue Type: Bug
          Components: S3
            Reporter: Elek, Marton
            Assignee: Elek, Marton


S3 Multi-Part-Upload (MPU) is implemented recently in the Ozone s3 gateway. We 
have extensive testing with using 'aws s3api' application which is passed.

But it turned out that the more simple `aws s3 cp` command fails with _405 
Media type not supported error_ message

The root cause of this issue is the JAXRS implementation of the multipart 
upload method:

{code}
  @POST
  @Produces(MediaType.APPLICATION_XML)
  public Response multipartUpload(
      @PathParam("bucket") String bucket,
      @PathParam("path") String key,
      @QueryParam("uploads") String uploads,
      @QueryParam("uploadId") @DefaultValue("") String uploadID,
      CompleteMultipartUploadRequest request) throws IOException, OS3Exception {
    if (!uploadID.equals("")) {
      //Complete Multipart upload request.
      return completeMultipartUpload(bucket, key, uploadID, request);
    } else {
      // Initiate Multipart upload request.
      return initiateMultipartUpload(bucket, key);
    }
  }
{code}

Here we have a CompleteMultipartUploadRequest parameter which is created by the 
JAXRS framework based on the media type and the request body. With 
_Content-Type: application/xml_ it's easy: the JAXRS framework uses the 
built-in JAXB serialization. But with plain/text content-type it's not possible 
as there is no serialization support for CompleteMultipartUploadRequest from 
plain/text.





--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org

Reply via email to