[
https://issues.apache.org/jira/browse/YARN-8672?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16684242#comment-16684242
]
Eric Yang commented on YARN-8672:
---------------------------------
[~csingh] The token file is used in two purposes. One copy of token file is
used to localize resource, and another copy is used to run the task. The copy
used to run task doesn't change frequently. The race condition happens for
localize resource where the token file is create/deleted after each
localization. This JIRA must focus on changing filename for the copy that is
used for localize resource to avoid token file overwrite during multiple
localization threads. Therefore, we might not need to change logic in
container-executor C code. Can you validate if we are agreeing on the same
problem?
> TestContainerManager#testLocalingResourceWhileContainerRunning occasionally
> times out
> -------------------------------------------------------------------------------------
>
> Key: YARN-8672
> URL: https://issues.apache.org/jira/browse/YARN-8672
> Project: Hadoop YARN
> Issue Type: Bug
> Components: nodemanager
> Affects Versions: 3.2.0
> Reporter: Jason Lowe
> Assignee: Chandni Singh
> Priority: Major
> Attachments: YARN-8672.001.patch, YARN-8672.002.patch,
> YARN-8672.003.patch, YARN-8672.004.patch, YARN-8672.005.patch
>
>
> Precommit builds have been failing in
> TestContainerManager#testLocalingResourceWhileContainerRunning. I have been
> able to reproduce the problem without any patch applied if I run the test
> enough times. It looks like something is removing container tokens from the
> nmPrivate area just as a new localizer starts.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
