[
https://issues.apache.org/jira/browse/YARN-6443?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15964939#comment-15964939
]
Wangda Tan commented on YARN-6443:
----------------------------------
[~asuresh],
The proposal to allow reordering of different allocation id with the same
priority looks like a better/cleaner problem definition.
bq. The patch you posted seemed to be doing something similar, but I guess that
will work if there are multiple outstanding reqs againsts the same
SchedulerKey. What we requires is efficient matching across schedulerKeys
I think we should move to an approach which can look up multiple nodes for one
or more resource requests, otherwise it will be hard for locality scheduling in
a large cluster. Setting up missed opportunities, etc. can only alleviate the
problem but cannot solve root cause of the issue.
Looking forward to your patch so we can have more discussions based on that.
> Prune number of SchedulerRequestKeys to examine in the Scheduler inner loop
> to improve performance
> --------------------------------------------------------------------------------------------------
>
> Key: YARN-6443
> URL: https://issues.apache.org/jira/browse/YARN-6443
> Project: Hadoop YARN
> Issue Type: Improvement
> Components: capacity scheduler, fairscheduler
> Reporter: Arun Suresh
> Assignee: Hitesh Sharma
>
> Currently the Schedulers examine an applications pending Requests in Priority
> order. This JIRA proposes to introduce a flag (either via the
> ApplicationMasterService::registerApplication() or via some Scheduler
> configuration) to favor an ordering that is baised to the node that is
> currently heartbeating by relaxing the priority constraint.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
