Samrat002 commented on code in PR #27788:
URL: https://github.com/apache/flink/pull/27788#discussion_r3214472791
##########
flink-filesystems/flink-s3-fs-native/README.md:
##########
@@ -89,6 +89,29 @@ input.sinkTo(FileSink.forRowFormat(new
Path("s3://my-bucket/output"),
| s3.assume-role.session-name | flink-s3-session | Session name for the
assumed role |
| s3.assume-role.session-duration | 3600 | Session duration in seconds
(900-43200) |
+## Bucket-Level Configuration
+
+The Native S3 FileSystem supports per-bucket configuration overrides, allowing
different S3 buckets to use different connection settings within the same Flink
cluster. This enables scenarios like:
+
+- **Checkpointing to one bucket** with specific credentials
+- **Savepoints to another bucket** with different region/endpoint
+- **Data sinks to third-party buckets** with cross-account IAM roles
+
+### Format
+
+Bucket-level configuration uses the format:
`s3.bucket.<bucket-name>.<property>`
+
+Bucket names containing dots (e.g., `my.company.data`) are fully supported
through longest-suffix matching.
Review Comment:
Added. The README now includes a note under the "Format" section:
> **Note:** AWS recommends avoiding periods (`.`) in bucket names. Buckets
with dots cannot use virtual-hosted-style addressing over HTTPS without custom
certificate validation. If you use dotted bucket names, enable
`path-style-access` for that bucket (see AWS documentation).
The wording was kept concise as suggested, with a link to the AWS bucket
naming rules page.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
