I wanted to know if it was possible to execute/or submit a mapred job to be executed in the hadoop cluster after a given delay from the job before. To explain clearly, jobs are submitted in a sequence and jobs are submitted before the prior mapred job completed execution, Say jobs are submitted to the cluster every 30ms to 1s.

CodePudding user response：

You probably want to use Oozie it's made to order jobs. Most distributions of hadoop (HDP/CDP) usually have it already installed on the cluster. It basically does worklow managment for hadoop. (ie run this job, then that job, then that job.)

(Can be run like a cron job but made to launch on the cluster)