[
https://issues.apache.org/jira/browse/YARN-8348?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16487886#comment-16487886
]
Íñigo Goiri commented on YARN-8348:
-----------------------------------
Technically the null check in AfterClass shouldn't be needed as a failure in
BeforeClass should trigger the error everywhere else.
In any case, is good to not have a NPE if the BeforeClass fails.
So in the output we went from a double NoClassDefFound+NPE to just
NoClassDefFound.
I think this is an improvement but we need to figure out the reason for the
NoClassDefFound (probably a separate JIRA).
The real fix here would be the one in TestHBaseTimelineStorageDomain which
leaves the mini cluster open.
[^YARN-8348.v1.patch] LGTM.
Let's wait for Yetus.
> Incorrect and missing AfterClass in HBase-tests
> -----------------------------------------------
>
> Key: YARN-8348
> URL: https://issues.apache.org/jira/browse/YARN-8348
> Project: Hadoop YARN
> Issue Type: Bug
> Reporter: Giovanni Matteo Fumarola
> Assignee: Giovanni Matteo Fumarola
> Priority: Major
> Attachments: YARN-8348.v1.patch
>
>
> HBase tests are failing in
> [linux|https://builds.apache.org/view/H-L/view/Hadoop/job/hadoop-qbt-trunk-java8-linux-x86/789/testReport/]
> for 2 reasons:
> * incorrect afterClass;
> * not defined KeyProviderTokenIssuer.
> While in windows are failing for the previous 2 reasons plus * missing
> afterClass.
> This Jira tracks the effort to fix part of HBase-tests and reduces the failed
> tests in Linux.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
