xuanyuanking commented on issue #24110: [SPARK-25341][Core] Support rolling back a shuffle map stage and re-generate the shuffle files URL: https://github.com/apache/spark/pull/24110#issuecomment-479709767 ``` Can you please explain how your PR takes care of multiple write attempts? ``` @ankuriitg Thanks for your question, just as the answer by Wenchen, the general idea is to reuse the stage attempt id as the "generation id" for each indeterminate shuffle map task, the output file will be extended by the "generation id" while multiple write attempts happened, you can check the corresponding code in all shuffle writers, e.g. [SortShuffleWriter](https://github.com/apache/spark/pull/24110/files#diff-1317c86dab6bc6323aeb935a53db6aacR79).
---------------------------------------------------------------- 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]
