Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/4598#discussion_r139679505
--- Diff:
flink-yarn-tests/src/test/java/org/apache/flink/yarn/YARNSessionCapacitySchedulerITCase.java
---
@@ -153,68 +153,62 @@ public void testTaskManagerFailure() {
// ------------------------ Test if JobManager web interface is
accessible -------
- YarnClient yc = null;
- try {
- yc = YarnClient.createYarnClient();
- yc.init(YARN_CONFIGURATION);
- yc.start();
+ final YarnClient yc = YarnClient.createYarnClient();
+ yc.init(YARN_CONFIGURATION);
+ yc.start();
- List<ApplicationReport> apps =
yc.getApplications(EnumSet.of(YarnApplicationState.RUNNING));
--- End diff --
True, but I came across while testing the Dispatcher REST endpoint in the
YARN case. I think there is some value in letting the test fail with the proper
stack trace instead of only failing the test with the failure message.
---
