[
https://issues.apache.org/jira/browse/YARN-4962?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15253975#comment-15253975
]
Daniel Templeton commented on YARN-4962:
----------------------------------------
Got it. I misunderstood your initial problem statement. You want jobs to be
scheduled to fill a node completely before scheduling to the next node to avoid
having the workload spread like peanut butter all over the cluster, making it
hard to schedule a job that needs a full node.
In Grid Engine, the scheduling formula is configurable. The scheduler will
look for the node for which the scheduling formula has the highest value. The
default scheduling formula is essentially "amount of free space." To get the
fill-up behavior you want, you'd set the scheduling formula to "number of
containers."
Which is essentially your original suggestion.
> support filling up containers on node one by one
> -------------------------------------------------
>
> Key: YARN-4962
> URL: https://issues.apache.org/jira/browse/YARN-4962
> Project: Hadoop YARN
> Issue Type: Improvement
> Reporter: sandflee
>
> we had a gpu cluster, jobs with bigger resource request couldn't be satisfied
> for node is running the jobs with smaller resource request. we didn't open
> reserve system because gpu jobs may run days or weeks. we expect scheduler
> allocate containers to fill the node , then there will be resource to run
> jobs with big resource request.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
