[
https://issues.apache.org/jira/browse/FLINK-16005?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17040797#comment-17040797
]
Yang Wang commented on FLINK-16005:
-----------------------------------
Using the hadoop/yarn configuration directly makes sense to me. I just have one
concern. For yarn configuration, it is easy to override. However, for hadoop
configuration(keys in core-site.xml and hdfs-site.xml) is not very easy. Since
in some situation, we do not directly new the {{Configuration}}. For example,
we have the following code in {{YarnClusterDescriptor}}.
{code:java}
FileSystem fs = FileSystem.get(yarnConfiguration)
{code}
If we want to make the hadoop config keys take effect, i think we need to new
a {{Configuration}} and override it, not using the {{YarnConfiguration}}.
> Propagate yarn.application.classpath from client to TaskManager Classpath
> -------------------------------------------------------------------------
>
> Key: FLINK-16005
> URL: https://issues.apache.org/jira/browse/FLINK-16005
> Project: Flink
> Issue Type: Improvement
> Components: Deployment / YARN
> Reporter: Zhenqiu Huang
> Priority: Major
>
> When Flink users want to override the hadoop yarn container classpath, they
> should just specify the yarn.application.classpath in yarn-site.xml from cli
> side. But currently, the classpath setting can only be used in flink
> application master, the classpath of TM is still determined by the setting in
> yarn host.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
