rdblue commented on issue #23608: [SPARK-26682][SQL] Use taskAttemptID instead of attemptNumber for Hadoop. URL: https://github.com/apache/spark/pull/23608#issuecomment-456997104 @zsxwing, changing Spark's behavior in HadoopMapReduceCommitProtocol isn't related to the problem here. If you think that is broken and will collide as well, please open an issue for it. The problem is that some Hadoop committer implementations expect the TaskAttemptID to be unique because stages are re-attempted in separate Hadoop MR jobs. Spark should ensure that the TaskAttemptIDs that Spark creates are not reused.
---------------------------------------------------------------- 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]
