[
https://issues.apache.org/jira/browse/YARN-4963?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15281035#comment-15281035
]
Wangda Tan commented on YARN-4963:
----------------------------------
Thanks [~nroberts], few comments:
1) Configuration name:
How about call it:
yarn.scheduler.capacity.per-node-heartbeat.maximum-offswitch-assignments?
offswitch-per-node-limit is a little confusing to me. And in the future we can
add other limits under per-node-heartbeat if needed.
2) We may only need to add getOffSwitchNodeLimit to ParentQueue (instead of
adding it to AbstractCSQueue)
3) (Minor) Logics in ParentQueue:
Add isDebugEnabled for:
{code}
500 LOG.debug("Not assigning more than " + getOffSwitchNodeLimit() +
501 " off-switch containers," +
{code}
And it's better to print offswitchCount with the debug log.
Thoughts?
> capacity scheduler: Make number of OFF_SWITCH assignments per heartbeat
> configurable
> ------------------------------------------------------------------------------------
>
> Key: YARN-4963
> URL: https://issues.apache.org/jira/browse/YARN-4963
> Project: Hadoop YARN
> Issue Type: Improvement
> Components: capacityscheduler
> Affects Versions: 3.0.0, 2.7.2
> Reporter: Nathan Roberts
> Assignee: Nathan Roberts
> Attachments: YARN-4963.001.patch, YARN-4963.002.patch
>
>
> Currently the capacity scheduler will allow exactly 1 OFF_SWITCH assignment
> per heartbeat. With more and more non MapReduce workloads coming along, the
> degree of locality is declining, causing scheduling to be significantly
> slower. It's still important to limit the number of OFF_SWITCH assignments to
> avoid densely packing OFF_SWITCH containers onto nodes.
> Proposal is to add a simple config that makes the number of OFF_SWITCH
> assignments configurable.
> Will upload candidate patch shortly.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
