gaborgsomogyi commented on pull request #29729:
URL: https://github.com/apache/spark/pull/29729#issuecomment-702748456


   > with and without this patch all three operations are used.
   
   It's true when we take a look at it globally (driver + executors).
   
   > That said, the behavior of not checking group.id in executors (#28623 
(comment)) is weird to me. Probably it's either Kafka bug or on group resource 
read operation doesn't leverage group.id.
   
   That said this more like a feature (though weird to me as well). When one 
uses `assign` operation `KafkaConsumer` is not joining to any group. As a 
consequence no `group.id` authorization performed. The only exception (which 
Spark not uses) is when `KafkaConsumer` uses `assign` operation and either auto 
or manual offset commit initiated. Such case `group.id` authorization would be 
performed. I don't think Kafka guys can be convinced on this that it's a bug...
   


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
[email protected]



---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to