rdblue commented on issue #23608: [SPARK-26682][SQL] Use taskAttemptID instead of attemptNumber for Hadoop. URL: https://github.com/apache/spark/pull/23608#issuecomment-456626927 @vanzin, because the attemptNumber resets to 0 for every stage attempt, the Hadoop TaskAttemptID is not unique. Code that assumes that will be unique (like our committer) can produce incorrect results. That's why the `sparkAttemptNumber` that you pointed out should be set to Spark's unique attempt ID instead of the task's attempt number.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on 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]
