[
https://issues.apache.org/jira/browse/YARN-7494?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16415435#comment-16415435
]
Sunil G commented on YARN-7494:
-------------------------------
Thanks [~cheersyang]
# Removed NodeSortingAlgorithm to make it simpler. Now to get a different
policy, it must be configured in app/queue.
# *CapacityScheduler: line 1803, 1809 unnecessary change* : So for all these
changes, i ll explain a little bit. When a change happens for label (add node
to label, remove node from label, or replace) in CS, i now get a per-label
nodes on all such events and caches it in ClusterNodeTracker. Hence a clear
cached node per partition is now available in ClusterNodeTracker. Now from
MultiNodeSorter, we are resorting cached nodes by getting a copy of the nodes
per partition from ClusterNodeTracker. By this, we no longer need to check with
LabelManager.
# *getCSLeafQueue() did a cast to queue that assumes it is capacity
scheduler.* Since we are doing this from FiCaSchedulerApp, we can definitely
assume its CS. Correct? If I would have done this SchedulerApplicationAttempt,
we had this issue.
# *line 52: {{monitorInterval}} is not initiated.* Could this be more on
per-policy level? In this patch I havent done this. But my idea is to have
getMonitorInterval api in MultiNodePolicy and ask each policy to implement. I
will update this in next patch. Somehow missed in this one.
> Add muti node lookup support for better placement
> -------------------------------------------------
>
> Key: YARN-7494
> URL: https://issues.apache.org/jira/browse/YARN-7494
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: capacity scheduler
> Reporter: Sunil G
> Assignee: Sunil G
> Priority: Major
> Attachments: YARN-7494.001.patch, YARN-7494.002.patch,
> YARN-7494.003.patch, YARN-7494.004.patch, YARN-7494.005.patch,
> YARN-7494.v0.patch, YARN-7494.v1.patch, multi-node-designProposal.png
>
>
> Instead of single node, for effectiveness we can consider a multi node lookup
> based on partition to start with.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
