[
https://issues.apache.org/jira/browse/YARN-4390?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15259248#comment-15259248
]
Wangda Tan commented on YARN-4390:
----------------------------------
[~eepayne],
I tried this in a SLS cluster:
3 nodes, each one is 8G.
3 queues:
{code}
a.capacity=16.6
b.capacity=16.6
c.capacity=66.8
b/c.disable_preemption=true
{code}
app_1 runs at a, size of each container=1G, it takes over all capacity in the
cluster
app_2 submit to c later, size of each container (including am container) is 2G,
it can preemption resources slowly, looks normal to me.
Few suggestions about preemption configuration:
1) total_preemption_per_round should make sure that, each round needs preempt
enough resource to allocate *one large container*
2) before ver.7, natural_termination_factor should set to 1 to make sure enough
resources will be preempted.
Let me know if it works.
> Do surgical preemption based on reserved container in CapacityScheduler
> -----------------------------------------------------------------------
>
> Key: YARN-4390
> URL: https://issues.apache.org/jira/browse/YARN-4390
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: capacity scheduler
> Affects Versions: 3.0.0, 2.8.0, 2.7.3
> Reporter: Eric Payne
> Assignee: Wangda Tan
> Attachments: QueueNotHittingMax.jpg, YARN-4390-design.1.pdf,
> YARN-4390-test-results.pdf, YARN-4390.1.patch, YARN-4390.2.patch,
> YARN-4390.3.branch-2.patch, YARN-4390.3.patch, YARN-4390.4.patch,
> YARN-4390.5.patch, YARN-4390.6.patch, YARN-4390.7.patch
>
>
> There are multiple reasons why preemption could unnecessarily preempt
> containers. One is that an app could be requesting a large container (say
> 8-GB), and the preemption monitor could conceivably preempt multiple
> containers (say 8, 1-GB containers) in order to fill the large container
> request. These smaller containers would then be rejected by the requesting AM
> and potentially given right back to the preempted app.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
