[
https://issues.apache.org/jira/browse/FLINK-2213?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14994471#comment-14994471
]
Johannes commented on FLINK-2213:
---------------------------------
I just ran into this problem on a cluster configured with the
{{LinuxContainerExecutor}}
Which is fairly easy to configure with CDH.
When there are more slots configured per Taskmanager and each do some heavy
computation, the whole container is terminated.
In this scenario the setup should be vcores = #slots, per container
The rational behind the {{LinuxContainerExecutor}} is to have more services
running on the cluster, not just yarn but also Hbase and Impala etc. and
provide some resource isolation.
> Configure number of vcores
> --------------------------
>
> Key: FLINK-2213
> URL: https://issues.apache.org/jira/browse/FLINK-2213
> Project: Flink
> Issue Type: Improvement
> Components: YARN Client
> Affects Versions: 0.10
> Reporter: Ufuk Celebi
>
> Currently, the number of vcores per YARN container is set to 1.
> It is desirable to allow configuring this value. As a simple heuristic it
> makes sense to at least set it to the number of slots per container.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
