You are viewing a plain text version of this content. The canonical link for it is here.

Posted to mapreduce-user@hadoop.apache.org by Павел Мезенцев <pa...@mezentsev.org> on 2015/03/03 03:50:45 UTC

Push or pull in yarn

Hi All!


Resently I read an article about facebook's corona
https://www.facebook.com/notes/facebook-engineering/under-the-hood-scheduling-mapreduce-jobs-more-efficiently-with-corona/10151142560538920

they solved in mr1 pull based task assignment:
task trackers send heartbeat to jobtracker and recive new tasks  in
respond.
This approach wasted 10-20 seconds for each tasks.

My question is about YARN: was this trouble solved in YARN or not?

Best regards,
Mezentsev Pavel