Goal
Change the automatic setting where all jobs submitted to the cluster are under one user account called datameer
or are into one queue.
Learn
To take advantage from scheduling or capacity queues in your cluster back end, you need to define a queue for each user or job within the Custom Properties
.
To address different execution frameworks you can set two parameters:
mapreduce.job.queuename=<Queue_X> tez.queue.name=<Queue_X> spark.yarn.queue=<Queue_X>
If you use Impersonation or Secure Impersonation, no configuration in Datameer is necessary because jobs are submitted to the cluster under the owner of the artifact <Person_X>
. The configuration is done on the cluster only.
Further Information
You can find this information under Scheduling jobs with the cluster and for a Cloudera cluster, under Resource Management.
Comments
0 comments
Please sign in to leave a comment.