Github user justinrmiller commented on the issue: https://github.com/apache/spark/pull/20572 Otherwise the job crashes, even for gaps of one or two messages on a couple of partitions (and while fail hard is a good philosophy, in this case it's not really practical for us) so it's pretty significant to us. That said, I've got the code copied into my project so when we go to 2.3.0 we could probably just keep using that.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org