[
https://issues.apache.org/jira/browse/FLINK-13839?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16915455#comment-16915455
]
Yang Wang commented on FLINK-13839:
-----------------------------------
Hi [~yanghua],
Thanks for your kind reminder. I think this ticket is not a duplicate of
FLINK-7836.
FLINK-7836 uses ApplicationSubmissionContext#setNodeLabelExpression() to set
the node label for application. All the containers of this application will be
started in the specified partition.
We want to split it into two separate config options so that we could make the
flink master container and taskmanager containers running on different
partitions. The ResourceRequest#setNodeLabelExpression() will be used. It is
very useful for cloud environment. Some nodes are stable and others are cheap
but not stable(preemptible instances in alibaba cloud, etc.). If we could start
the flink master on the stable nodes, the flink job will be more stable.
> Support to set yarn node label for flink jobmanager and taskmanager container
> -----------------------------------------------------------------------------
>
> Key: FLINK-13839
> URL: https://issues.apache.org/jira/browse/FLINK-13839
> Project: Flink
> Issue Type: New Feature
> Components: Deployment / YARN
> Reporter: Yang Wang
> Priority: Major
>
> Yarn node label feature is introduced from 2.6. It is a way to group nodes
> with similar characteristics and applications can specify where to run. In
> the production or cloud environment, we want to the jobmanager running on
> some more stable machines. The node label could help us to achieve that.
>
> However, the ResourceRequest.setNodeLabelExpression have not been supported
> in the current hadoop version dependency(2.4.1). So we need to bump the
> hadoop version to 2.6.5.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
