[
https://issues.apache.org/jira/browse/YARN-4512?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15097781#comment-15097781
]
Inigo Goiri commented on YARN-4512:
-----------------------------------
bq. I'm not very familiar with ResourceOption, is it only used by RM? I think
it is possible that NM will use its own data to decide how many resources could
be over-allocated.
Right now, the amount of resources for an NM is set in the registration of the
NM. Afterwards, if an admin wants to change the available resources, it has to
use the admin protocol to change it (YARN-312). I think that the overallocation
parameter matches this: (1) we register the node with an overallocation
parameter and (2) we change this parameter using the admin protocol or similar.
In this case, both the NM and the RM are aware of this setting.
> Provide a knob to turn on over-allocation
> -----------------------------------------
>
> Key: YARN-4512
> URL: https://issues.apache.org/jira/browse/YARN-4512
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: nodemanager
> Reporter: Karthik Kambatla
> Assignee: Karthik Kambatla
> Attachments: YARN-4512-YARN-1011.001.patch,
> yarn-4512-yarn-1011.002.patch, yarn-4512-yarn-1011.003.patch
>
>
> We need two configs for overallocation - one to specify the threshold upto
> which it is okay to over-allocate, another to specify the threshold after
> which OPPORTUNISTIC containers should be preempted.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
