I’m currently working on separating indexing job from fixed yarn cluster. (Setting up with druid)
Because i have multiple tables, and i don’t want several indexing job run in same yarn cluster which is occupy huge disk, network bandwidth.
So i want some of them will go to another idle yarn cluster(separated), and remaining jobs are goes to original yarn cluster.
yarn.resourcemanager.address property into indexing job spec.
Sorry I don’t have the exact answer, but I’d suggest checking out https://hadoop.apache.org/docs/r2.7.2/hadoop-yarn/hadoop-yarn-common/yarn-default.xml and setting as many of the properties involving yarn.resourcemanager.address and yarn.resourcemanager.hostname as possible.
Yeah… i already added as many options in there as i can…
Including yarn.resourcemanager.hostname as well.
I met same error again.
2018년 2월 8일 목요일 오전 10시 54분 51초 UTC+9, Gian Merlino 님의 말: