flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "dongxiao.yang (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-8270) TaskManagers do not use correct local path for shipped Keytab files in Yarn deployment modes
Date Mon, 18 Dec 2017 03:35:00 GMT

    [ https://issues.apache.org/jira/browse/FLINK-8270?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16294489#comment-16294489
] 

dongxiao.yang commented on FLINK-8270:
--------------------------------------

[~tzulitai]  
Hi Gordon ,I'm the reporter of the mail http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Flink-1-4-0-keytab-is-unreadable-td17292.html
. May I do some works to help solve this issue?

> TaskManagers do not use correct local path for shipped Keytab files in Yarn deployment
modes
> --------------------------------------------------------------------------------------------
>
>                 Key: FLINK-8270
>                 URL: https://issues.apache.org/jira/browse/FLINK-8270
>             Project: Flink
>          Issue Type: Bug
>          Components: Security
>    Affects Versions: 1.4.0
>            Reporter: Tzu-Li (Gordon) Tai
>            Priority: Blocker
>             Fix For: 1.5.0, 1.4.1
>
>
> Reported on ML: http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Flink-1-4-0-keytab-is-unreadable-td17292.html
> This is a "recurrence" of FLINK-5580. The TMs in Yarn deployment modes are again not
using the correct local paths for shipped Keytab files.
> The cause was accidental due to this change: https://github.com/apache/flink/commit/7f1c23317453859ce3b136b6e13f698d3fee34a1#diff-a81afdf5ce0872836ac6fadb603d483e.
> Things to consider:
> 1) The above accidental breaking change was actually targeting a minor refactor on the
"integration test scenario" code block in {{YarnTaskManagerRunner}}. It would be best if we
can remove that test case code block from the main code.
> 2) Unit test coverage is apparently not enough. As this incidence shows, any slight changes
can cause this issue to easily resurface again.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message