[
https://issues.apache.org/jira/browse/YARN-6641?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16024818#comment-16024818
]
Jason Lowe commented on YARN-6641:
----------------------------------
Thanks for the patch! Patch looks good overall.
At this point the only callers of the LocalResourcesTrackerImpl constructor
that omits the directory handler are tests, and I think it would be better to
simply remove this constructor and update the few places in the tests to
explicitly pass null. That way future maintainers won't be lulled into
thinking it's OK to call the constructor without a handler, since it clearly
needs a dir handler to properly deal with resources that get orphaned on bad
disks.
> Non-public resource localization on a bad disk causes subsequent containers
> failure
> -----------------------------------------------------------------------------------
>
> Key: YARN-6641
> URL: https://issues.apache.org/jira/browse/YARN-6641
> Project: Hadoop YARN
> Issue Type: Bug
> Affects Versions: 2.8.0
> Reporter: Kuhu Shukla
> Assignee: Kuhu Shukla
> Attachments: YARN-6641.001.patch, YARN-6641.002.patch
>
>
> YARN-3591 added the {{checkLocalResource}} method to {{isResourcePresent()}}
> call to allow checking an already localized resource against the list of
> good/full directories.
> Since LocalResourcesTrackerImpl instantiations for app level resources and
> private resources do not use the new constructor, such resources that are on
> bad disk will never be checked against good dirs.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
