leixm commented on PR #307: URL: https://github.com/apache/incubator-uniffle/pull/307#issuecomment-1305680815
> Spark usually start the task one by one, If we have 700 reduce tasks, 7 shuffle server, we usually start 0 - 99 task first. 0 - 99 reduce will be allocated to shuffle server 0. Will it increase too much pressure for shuffle server 0? So I think origin implement may be useful for some situation. It's better to let users to choose which strategy is most suitable. We should separate the mechanism from strategy, let configuration to decide which way we want. It's ok for me about server's configuration or client's configuration. Good idea. -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
