[
https://issues.apache.org/jira/browse/FLINK-21981?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17310371#comment-17310371
]
Paul Lin commented on FLINK-21981:
----------------------------------
[~Bo Cui] Just to confirm, did you try exporting `HADOOP_CONF_DIR` before run
any flink commands? It should work for multi-tenant scenarios. Env variables
are more flexible and environment-specific, thus should take higher priority
than static configurations.
> Increase the priority of the parameter in flink-conf
> ----------------------------------------------------
>
> Key: FLINK-21981
> URL: https://issues.apache.org/jira/browse/FLINK-21981
> Project: Flink
> Issue Type: Bug
> Components: Client / Job Submission
> Reporter: Bo Cui
> Priority: Major
>
> in my cluster, env has HADOOP_CONF_DIR and YARN_CONF_DIR, but they are hadoop
> and yarn conf, not flink conf. and hadoop conf of flink(env.hadoop.conf.dir)
> is different from them. so we should use `env.hadoop.conf.dir` first
> https://github.com/apache/flink/blob/57e93c90a14a9f5316da821863dd2b335a8f86e0/flink-dist/src/main/flink-bin/bin/config.sh#L256
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
