[
https://issues.apache.org/jira/browse/YARN-10486?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17233298#comment-17233298
]
Peter Bacsko commented on YARN-10486:
-------------------------------------
Thanks [~bteke] let's proceed with approach #2.
I made several changes in patch #5:
* Renamed variables
* Renamed the property to relax-capacity-validation
* Added unit tests to TestParentQueue.java
* Fixed validations in ParentQueue.java
Note that there are two places in ParentQueue.java where you have to switch
between normal and relaxed validation. The previous patch was incorrect. The
"if" conditions might not be completely obvious, let me know if it's confusing
I'll and try to rewrite it.
Another thing is that we can't just let the queues to have arbitrary capacity.
It would be good but the capacity values propagate towards the top (root) and
eventually somewhere along the line it will fail. So basically you'd need to
set the new property to "true" for a given hierarchy, which is really not
convenient. So the sum of capacities must be either 0% or 100%. But it's still
more lenient than the current code.
> FS-CS converter: handle case when weight=0
> ------------------------------------------
>
> Key: YARN-10486
> URL: https://issues.apache.org/jira/browse/YARN-10486
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: yarn
> Reporter: Peter Bacsko
> Assignee: Peter Bacsko
> Priority: Major
> Labels: fs2cs
> Attachments: YARN-10486-001.patch, YARN-10486-002.patch,
> YARN-10486-003-approach2.patch, YARN-10486-004-approach2.patch,
> YARN-10486-005.patch
>
>
> We can encounter an ArithmeticException if there is a single or multiple
> queues under a parent with zero weight.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
