DanielLeens commented on issue #9063:
URL: https://github.com/apache/seatunnel/issues/9063#issuecomment-4671707987

   Thanks for raising this.
   
   I checked the current Spark translation path, and this is a real 
compatibility gap rather than a configuration mistake in your job.
   
   The important part is that the CDC incremental reader sends `SourceEvent`s 
back to the enumerator when snapshot splits finish and when it transitions 
phases. But on the Spark batch path, if the source does **not** expose the 
coordinated-source marker, it goes through `ParallelReaderContext`, and that 
context throws exactly the `UnsupportedOperationException` you hit when 
`sendSourceEventToEnumerator(...)` is called.
   
   So the failure is not coming from your Hudi sink config. It is coming from 
the fact that the Spark translation path is still treating this CDC source as a 
plain parallel source even though the reader/enumerator protocol needs 
coordination.
   
   As a practical workaround today, I would run this CDC job on Zeta or Flink 
instead of Spark. The Spark path needs a follow-up fix so CDC sources that emit 
`SourceEvent`s are exposed through the coordinated source path there as well.
   
   This is worth keeping open as a real bug. Contributions are very welcome.


-- 
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