hustfeiwang commented on issue #23711: [SPARK-26795][SHUFFLE]Retry remote fileSegmentManagedBuffer when creating inputStream failed during shuffle read phase#SPARK-26795 URL: https://github.com/apache/spark/pull/23711#issuecomment-459609193 @squito There are few IOExceptions occur when creating inputStream. But I think it is a complete mechanism to refetch these remote corrupt block just like when the shuffleFetcherItertor detect that the block is corrupt later. If it is just as a fetchFailed to deal with, it would remove the relative mapout put and recompute previous relative tasks, which is expensive in a multi-stage job especially.
---------------------------------------------------------------- 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]
