[
https://issues.apache.org/jira/browse/YARN-5077?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15289516#comment-15289516
]
Yufei Gu commented on YARN-5077:
--------------------------------
Thanks [~kasha] for reviewing.
To filter out the non-active queue is done by {{handleFixedFairShares}}, which
is invoked by {{computeSharesInternal}}. But {{handleFixedFairShares}} does
more than that, it also filters out zero-weight queue, and that's one of place
we want to fix in this JIRA. If there are all zero-weight active queues, we
should get them in instead of removing them. So I should know if there is no
non-zero-weight active queue before {{handleFixedFairShares}}, which is done by
func {{noNonZeroWeightActive}}.
I totally agree we should avoid double negation name, but it seems reasonable
here. Of course, the name could be called {{allWeightsZero}} after function
{{handleFixedFairShares}}, so I modify the name after that as your suggestions.
I assume you said YARN-5106. It is fair enough.
> Fix FSLeafQueue#getFairShare() for queues with weight 0.0
> ---------------------------------------------------------
>
> Key: YARN-5077
> URL: https://issues.apache.org/jira/browse/YARN-5077
> Project: Hadoop YARN
> Issue Type: Bug
> Reporter: Yufei Gu
> Assignee: Yufei Gu
> Attachments: YARN-5077.001.patch, YARN-5077.002.patch
>
>
> 1) When a queue's weight is set to 0.0, FSLeafQueue#getFairShare() returns
> <memory:0, vCores:0>
> 2) When a queue's weight is nonzero, FSLeafQueue#getFairShare() returns
> <memory:16384, vCores:8>
> In case 1), that means no container ever gets allocated for an AM because
> from the viewpoint of the RM, there is never any headroom to allocate a
> container on that queue.
> For example, we have a pool with the following weights:
> - root.dev 0.0
> - root.product 1.0
> The root.dev is a best effort pool and should only get resources if
> root.product is not running. In our tests, with no jobs running under
> root.product, jobs started in root.dev queue stay stuck in ACCEPT phase and
> never start.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
