[
https://issues.apache.org/jira/browse/YARN-7612?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16300858#comment-16300858
]
Arun Suresh edited comment on YARN-7612 at 12/22/17 1:58 AM:
-------------------------------------------------------------
Updating patch (v009) with some minor refactoring.
* Removed the tag indexed map from the BatchedRequest - The algorithm should be
free to index/sort its Input as it see fit, so don't want to pre-sort/index its
input for it.
* We now feedback to the algorithm a blacklist. Essentially, if the scheduler
had rejected a placed request, and if retry attempts > 1, then when the
Algorthim tries to re-place the request, The node on which it had previously
placed (and was rejected by the scheduler) is available to the Algorithm as a
set of blacklisted nodes.
* I have also added a testcasse to show re-placement. The placement algorithm
places on a node that has no capacity, and in the next attempt, it will place
on a different node using the blacklist. [~cheersyang], this should demonstrate
something similar to what you were asking (although using anti-affinity, not
cardinality)
was (Author: asuresh):
Updating patch (v009) with some minor refactoring.
* Removed the tag indexed map from the BatchedRequest - The algorithm is free
to index its Input as it see fit.
* We now feedback to the algorithm a blacklist. Essentially, if the scheduler
had rejected a placed request, and if retry attempts > 1, then when the
Algorthim tries to re-place the request, The node on which it had previously
placed (and was rejected by the scheduler) is available to the Algorithm as a
set of blacklisted nodes.
* I have also added a testcasse to show re-placement. The placement algorithm
places on a node that has no capacity, and in the next attempt, it will place
on a different node using the blacklist. [~cheersyang], this should demonstrate
something similar to what you were asking (although using anti-affinity, not
cardinality)
> Add Placement Processor Framework
> ---------------------------------
>
> Key: YARN-7612
> URL: https://issues.apache.org/jira/browse/YARN-7612
> Project: Hadoop YARN
> Issue Type: Sub-task
> Reporter: Arun Suresh
> Assignee: Arun Suresh
> Attachments: YARN-7612-YARN-6592.001.patch,
> YARN-7612-YARN-6592.002.patch, YARN-7612-YARN-6592.003.patch,
> YARN-7612-YARN-6592.004.patch, YARN-7612-YARN-6592.005.patch,
> YARN-7612-YARN-6592.006.patch, YARN-7612-YARN-6592.007.patch,
> YARN-7612-YARN-6592.008.patch, YARN-7612-YARN-6592.009.patch,
> YARN-7612-v2.wip.patch, YARN-7612.wip.patch
>
>
> This introduces a Placement Processor and a Planning algorithm framework to
> handle placement constraints and scheduling requests from an app and places
> them on nodes.
> The actual planning algorithm(s) will be handled in a YARN-7613.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
