[
https://issues.apache.org/jira/browse/FLINK-7125?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16103358#comment-16103358
]
ASF GitHub Bot commented on FLINK-7125:
---------------------------------------
Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/4280#discussion_r129877902
--- Diff:
flink-yarn-tests/src/test/java/org/apache/flink/yarn/YarnTestBase.java ---
@@ -549,7 +559,12 @@ protected void runWithArgs(String[] args, String
terminateAfterString, String[]
final int startTimeoutSeconds = 180;
final long deadline = System.currentTimeMillis() +
(startTimeoutSeconds * 1000);
- Runner runner = new Runner(args, type, expectedReturnValue);
+ Runner runner = new Runner(
+ args,
+ flinkConfiguration,
+ CliFrontend.getConfigurationDirectoryFromEnv(),
--- End diff --
Since the configuration directory is only required for the session case in
`Runner`, the same argument applies as above. I agree that the
`FlinkYarnSessionCli` should have a configurable configuration directory which
can be controlled by a command line option.
> Remove configuration loading from AbstractYarnClusterDescriptor
> ---------------------------------------------------------------
>
> Key: FLINK-7125
> URL: https://issues.apache.org/jira/browse/FLINK-7125
> Project: Flink
> Issue Type: Improvement
> Components: YARN
> Reporter: Till Rohrmann
> Assignee: Till Rohrmann
> Labels: flip-6
>
> We should remove the explicit configuration loading from the
> {{AbstractYarnClusterDescriptor}} and instead pass it to the class via the
> constructor. This passed in configuration should then be sent to the
> application master when being started. That way, we can easily configure the
> started Yarn cluster.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
