skonto edited a comment on issue #24613: [SPARK-27549][SS] Add support for committing kafka offsets per batch for supporting external tooling URL: https://github.com/apache/spark/pull/24613#issuecomment-492766741 @arunmahadevan regarding the other issue about the failing batch I think since the [triggerExecutor](https://github.com/apache/spark/blob/efa303581ac61d6f517aacd08883da2d01530bd2/sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/TriggerExecutor.scala#L52) will block on this call [here](https://github.com/apache/spark/blob/master/sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/MicroBatchExecution.scala#L219) and since there are [no multiple in flight batches](https://github.com/apache/spark/blob/efa303581ac61d6f517aacd08883da2d01530bd2/sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/MicroBatchExecution.scala#L405) we are good because by the time we go back and construct the next batch the previous is done. @gaborgsomogyi could you verify the assumption whether bathes will pile up in case of a failure. Of course offsets go the commit log after the batch is run but then will have to filter the committedOffsets and commit back only the kafka ones [here](https://github.com/apache/spark/blob/efa303581ac61d6f517aacd08883da2d01530bd2/sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/MicroBatchExecution.scala#L571) which is also an option, I tried to comply with the commit API.
---------------------------------------------------------------- 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] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
