[
https://issues.apache.org/jira/browse/HADOOP-13686?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15564974#comment-15564974
]
Weiwei Yang commented on HADOOP-13686:
--------------------------------------
Hello [~xyao]
Thanks for your comments, it's helpful. I have addressed #1, #2 and #3 in v2
patch. But I am a little hesitated to do #4 and #5. I don't think HDFS-10922
will use AuditableTrashPolicy/AuditableCheckpoints, they are helper classes to
verify trash intervals in {{testTrashRestarts}}, I can't see how to reuse it in
HDFS trash tests. Regarding to #5, I used static AuditableCheckpoints and
static vars, because I need to share checkpoint states between multiple
instances of trash policies while simulating restart, I used atom integer to
avoid thread safety problem.
Once this one is committed, I will add test case in HDFS-10922 to call
{{TestTrash#testMoveEmptyDirToTrash}} and {{TestTrash#testTrashPermission}}, so
that to reuse these tests in HDFS. I already tried locally and it worked fine.
Please let me know how v2 looks, appreciate your help.
> Adding additional unit test for Trash (I)
> -----------------------------------------
>
> Key: HADOOP-13686
> URL: https://issues.apache.org/jira/browse/HADOOP-13686
> Project: Hadoop Common
> Issue Type: Test
> Reporter: Xiaoyu Yao
> Assignee: Weiwei Yang
> Attachments: HADOOP-13686.01.patch, HADOOP-13686.02.patch
>
>
> This ticket is opened to track adding the forllowing unit test in
> hadoop-common.
> #test users can delete their own trash directory
> #test users can delete an empty directory and the directory is moved to trash
> #test fs.trash.interval with invalid values such as 0 or negative
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
