cloud-fan commented on pull request #29797: URL: https://github.com/apache/spark/pull/29797#issuecomment-695965713
I think the root cause is, we give local shuffle reader expected parallelism, but it's not always satisfied. For example, too many mappers can lead to over-parallelism and many small files. We should add a check and only apply local shuffle reader if expected parallelism can be satisfied. ---------------------------------------------------------------- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
