Github user rezasafi commented on the issue: https://github.com/apache/spark/pull/19388 @mridulm thank you very much for the comment. I was thinking about something similar, but more like sending both stageID and a monotonically increasing jobId to the commit method and then let try the canCommit use the stageId while we use the jobID for hadoop commits. I am testing this right now.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org