S3Client#getObject response input streams contains more than it should

[mihohren]

Describe the bug

When switching from v2.29.x to v2.30.x I noticed the ResponseInputStream given by S3Client#getObject contains more than it should. The running example will be an object containing abc.

Regression Issue

• Select this option if this issue appears to be a regression.

Expected Behavior

S3Client#getObject#readAllBytes contains only

abc

Current Behavior

S3Client#getObject#readAllBytes contains

3
abc
0
x-amz-checksum-crc32:NSRBwg==
  

Reproduction Steps

Download an existing object through S3Client#getObject.

Possible Solution

Revert to 2.29.x

Additional Information/Context

Discovered through test in LocalStack testcontainer localstack/localstack:0.11.3

AWS Java SDK version used

v2.30.3 (bug confirmed for v2.30.0 - v2.30.3)

JDK version used

openjdk version "21.0.5" 2024-10-15 OpenJDK Runtime Environment (build 21.0.5+11-Ubuntu-1ubuntu122.04) OpenJDK 64-Bit Server VM (build 21.0.5+11-Ubuntu-1ubuntu122.04, mixed mode, sharing)

Operating System and version

Linux 6.8.0-51-generic #52~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Mon Dec 9 15:00:52 UTC 2 x86_64 x86_64 x86_64 GNU/Linux

[bhoradc]

Hello @mihohren,

Thank you for reporting the issue.

The S3Client#putObject request containing extra content (including checksum information like x-amz-checksum-crc32:NSRBwg==) seems related to the changes introduced in AWS SDK Java v2.30.0 - related Discussion Announcement.

As mentioned in the linked announcement, if you need to disable this behavior (especially when working with third-party services like LocalStack that might not support this feature yet), you can do so by setting the config flag to WHEN_REQUIRED, or by using the related AWS shared config file settings or environment variables.

For instance using below sample, works for both LocalStack and AWS S3:

        S3Client s3Client = S3Client.builder()
                .endpointOverride(URI.create("http://localhost:4566"))
                .serviceConfiguration(S3Configuration.builder()
                        .pathStyleAccessEnabled(true)
                        .build())
                 .responseChecksumValidation(ResponseChecksumValidation.WHEN_REQUIRED)
                 .requestChecksumCalculation(RequestChecksumCalculation.WHEN_REQUIRED)
                .build();

contentString retruns just abc using below code snippet:

  try (ResponseInputStream<GetObjectResponse> responseGet = s3Client.getObject(getObjectRequest)) {

            byte[] content = responseGet.readAllBytes();
            String contentString = new String(content);
            System.out.println("Full response content:");
            System.out.println(contentString);
        }

Also, I don't see below additional data for the getObject call. Kindly share the exact code sample you may be using to reproduce this behavior with the S3 getObject API.

3 - chunk size
abc - actual content
0 - marking the end of chunks
x-amz-checksum-crc32:NSRBwg== - trailer

Regards,
Chaitanya

[mihohren]

@bhoradc Thank you for the quick response, adding WHEN_REQUIRED to both requestChecksumCalculation and responseChecksumValidation seems to have fixed the problem.

[github-actions]

This issue is now closed. Comments on closed issues are hard for our team to see.
If you need more assistance, please open a new issue that references this one.