[
https://issues.apache.org/jira/browse/YARN-5395?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15381829#comment-15381829
]
sandflee commented on YARN-5395:
--------------------------------
we could separate delayQueue to amDelayQueue and nmDelayQueue, then the delay
of app wouldn't delay the nm heartbeat.
> All node are Lost in sls test for big cluster
> ---------------------------------------------
>
> Key: YARN-5395
> URL: https://issues.apache.org/jira/browse/YARN-5395
> Project: Hadoop YARN
> Issue Type: Improvement
> Reporter: sandflee
> Assignee: sandflee
>
> we are using sls to test sheduler performance, with 10000 nodes and 1000
> queues, after a few minutes all node are LOST, jstack shows executor pool
> are all acquired by am simulator.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
