rangareddy commented on issue #15198:
URL: https://github.com/apache/hudi/issues/15198#issuecomment-5140107842

   Closing as a configuration issue rather than a Hudi defect.
   
   Root cause: Hudi's Flink writer uses Hadoop's `s3a` filesystem directly, so 
it reads credentials from
   `core-site.xml` on `HADOOP_CONF_DIR`, from `flink.hadoop.*` entries in 
`flink-conf.yaml`, or from
   `hadoop.*`-prefixed table options. Flink's own `flink-s3-fs-hadoop` plugin 
settings (`s3.*`) are not part of
   that set, which is why the `filesystem`-connector CSV write succeeded while 
Hudi failed with `403 Forbidden`
   on `.hoodie` — and why the failure was unchanged with MinIO stopped, since 
Hudi had no endpoint and was not
   contacting MinIO at all. `hoodie.aws.*` configures Amazon CloudWatch and 
Glue, not the filesystem.
   
   The working configurations are given in the comment above. No code change is 
required in Hudi.
   
   The reported versions (Hudi 0.10.1, Flink 1.13.6) are also well out of 
support; if this workload is still
   live, please retest on a current bundle and reopen with details if the 
problem persists.
   
   The Spark-only [AWS S3](https://hudi.apache.org/docs/s3_hoodie) page is a 
genuine gap that contributed to
   this and should be fixed separately.


-- 
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]

Reply via email to