You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-issues@hadoop.apache.org by "Wilfred Spiegelenburg (Jira)" <ji...@apache.org> on 2019/12/23 23:50:00 UTC
[jira] [Comment Edited] (YARN-10053) Placement rules do not use
correct group service init
[ https://issues.apache.org/jira/browse/YARN-10053?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17002545#comment-17002545 ]
Wilfred Spiegelenburg edited comment on YARN-10053 at 12/23/19 11:49 PM:
-------------------------------------------------------------------------
The junit test failures are not related to my change. The tests pass locally. One failures is caused by a port bind exception, they seem environment related or are already known (YARN-9740, YARN-9338).
Uploading new patches to fix checkstyle issues in 3.1 and 3.2 branches.
was (Author: wilfreds):
The junit test failures are not related to my change. The tests pass locally. One failures is caused by a port bind exception, they seem environment related or are already known (YARN-9740, YARN-9338).
in 3.1 and 3.2 branches
Uploading new patches to fix checkstyle issues
> Placement rules do not use correct group service init
> -----------------------------------------------------
>
> Key: YARN-10053
> URL: https://issues.apache.org/jira/browse/YARN-10053
> Project: Hadoop YARN
> Issue Type: Bug
> Components: yarn
> Affects Versions: 3.1.3
> Reporter: Wilfred Spiegelenburg
> Assignee: Wilfred Spiegelenburg
> Priority: Major
> Attachments: YARN-10053-branch-3.1.001.patch, YARN-10053-branch-3.1.002.patch, YARN-10053-branch-3.1.003.patch, YARN-10053-branch-3.2.001.patch, YARN-10053.001.patch, YARN-10053.002.patch
>
>
> The placement rules, CS and FS, all create a new group service instead of using the shared group mapping service. This means that the cache for the placement rules is not same as for the ACL and other parts of the RM service.
> It also could cause an issue with the configuration that is passed in to create the cache: the scheduler config might not have the same values as the service config and could thus cause issues. This second issue just seems to affect the CS not the FS.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org