[
https://issues.apache.org/jira/browse/YARN-1881?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13948040#comment-13948040
]
Karthik Kambatla commented on YARN-1881:
----------------------------------------
[~rym] - thanks for reaching out. We typically use the user-lists
([email protected]) to discuss issues encountered by users. Once we arrive
at a specific cause, we use JIRA to keep track of the progress.
> configuration for performance tuning
> ------------------------------------
>
> Key: YARN-1881
> URL: https://issues.apache.org/jira/browse/YARN-1881
> Project: Hadoop YARN
> Issue Type: Test
> Components: nodemanager
> Affects Versions: 2.2.0
> Reporter: rrim
>
> Hi,
> my MR jobs run very slow on hadoop, some block in reduce tasks
> -input data is small, number of splits is 2
> -memory is available ('free -m' shows a usage under 4GB per node), each node
> has 32GB
> -yarnchild(s) are consuming 100% CPU (tests with "top")
> and I don't find ways to improve performance
> best
--
This message was sent by Atlassian JIRA
(v6.2#6252)
