have you researched the yarn schedulers? namely the capacity and fair schedulers? those are the places where resource limits can be easily defined.
On Mon, Jun 5, 2017 at 9:25 PM, Chang.Wu <583424...@qq.com> wrote: > My Hive engine is MapReduce and Yarn. What my urgent need is to limit the > memory usage of my big sql so that my bigger sql will run a longer time > instead of using up all the resource of queue or even all the resource of > my whole yarn cluster at once. > But I cannot find any solution to define my sql resource-usage upper > threshold. > Any one can give me some suggestions. > > 583424...@qq.com >
