HDDS-13737. S3 ETag JSON should be quoted

[echonesis]

What changes were proposed in this pull request?

This PR updates the output response during using aws s3api complete-multipart-upload function for Apache Ozone.

When we want to execute multi-part uploading, three steps/functions are included:

• create-multipart-upload
• upload-part
• complete-multipart-upload

We could refer to the doc and figure out the defined response of upload-part should be

{
    "ETag": "\"e868e0f4719e394144ef36531ee6824c\""
}

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-13737

How was this patch tested?

• GitHub Actions CI: https://github.com/echonesis/ozone/actions/runs/19090450103
• Local test like Mint

aws s3api --endpoint $ENDPOINT create-bucket --bucket=$BUCKET

UPLOAD_JSON=$(aws s3api create-multipart-upload \
  --bucket "$BUCKET" \
  --key "$KEY" \
  --endpoint-url "$ENDPOINT")

UPLOAD_ID=$(echo "$UPLOAD_JSON" | jq -r '.UploadId')

UPLOAD_PART_JSON=$(aws s3api upload-part \
  --bucket "$BUCKET" \
  --key "$KEY" \
  --part-number 1 \
  --body "$PART_FILE" \
  --upload-id "$UPLOAD_ID" \
  --endpoint-url "$ENDPOINT")

echo "$UPLOAD_PART_JSON"

[Copilot]

Pull Request Overview

This PR addresses ETag handling inconsistencies in the S3 gateway by ensuring ETags are properly quoted in responses and unquoted when stored or compared internally. The changes standardize the format of ETags across multipart upload operations.

• Adds stripQuotes and wrapInQuotes utility methods for consistent ETag formatting
• Updates multipart upload endpoints to wrap ETags in quotes for HTTP responses
• Modifies tests to strip quotes from ETags before comparisons and XML serialization

Reviewed Changes

Copilot reviewed 5 out of 6 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
ObjectEndpointStreaming.java	Wraps ETag in quotes for multipart upload response headers
ObjectEndpoint.java	Adds stripQuotes/wrapInQuotes utilities and applies proper ETag formatting in multipart operations
AbstractS3SDKV2Tests.java	Updates tests to strip quotes from ETags before comparison and XML construction
AbstractS3SDKV1Tests.java	Updates tests to strip quotes from ETags before XML construction
S3SDKTestUtils.java	Adds stripQuotes utility method for test code reuse
mpu_lib.robot	Strips quotes from ETag before MD5 comparison in Robot Framework tests

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Inconsistent ETag handling: The stripQuotes logic is only applied in the 'else' branch (lines 1070-1073) but not in the 'if (range != null)' branch (lines 1049-1059). Both branches copy source key metadata using putAll(), which means both should handle potentially quoted ETags consistently.

[Copilot]

Duplicate import: PerformanceStringBuilder is imported both as a static import (line 30) and as a regular import (line 115). The regular import should be removed since the static import is sufficient and already used throughout the code.

Suggested change

import org.apache.hadoop.ozone.audit.AuditLogger.PerformanceStringBuilder;

[Copilot]

Duplicate import: CopyDirective is imported both as a static import (line 50) and as a regular import (line 142). The regular import should be removed since the static import is sufficient.

Suggested change

import org.apache.hadoop.ozone.s3.util.S3Consts.CopyDirective;

[jojochuang]

We don't need this helper method. There's one in ObjectEndpoint.

[echonesis]

Thanks @jojochuang !
I remove it as expected.

[jojochuang]

After applying this PR, the same Mint test no longer fails for ETag error. Though it failed later for an unsupported feature.

(3/15) Running awscli tests ... FAILED in 25 seconds
{
"name": "awscli",
"duration": 2108,
"function": "aws --endpoint-url http://host.containers.internal:9878 s3api head-object --bucket awscli-mint-test-bucket-28157 --key datafile-1-kB\n",
"status": "FAIL",
"error": "StorageClass was applied incorrectly"
}

So yes, looks like this PR is functionally correct. As long as the duplicate method and code style issue is fixed, this PR is ready.

[ivandika3]

@echonesis Thanks for the fix, left some comments.

[ivandika3]

@echonesis Thanks for the update. LGTM +1.

[ivandika3]

Thanks @echonesis for the patch and @jojochuang for the review.