[
https://issues.apache.org/jira/browse/HADOOP-14675?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16108874#comment-16108874
]
Steve Loughran commented on HADOOP-14675:
-----------------------------------------
the path exists due a fundamental expectation in filesystems: paths are lists
of path elements separated by "/", all path elements being one or more
characters long. This is how "real" filesystems work, and so how we try to make
Hadoop's s3 client behave.
See:
https://hadoop.apache.org/docs/r2.8.0/hadoop-project-dist/hadoop-common/filesystem/introduction.html
It's never going to be fixed; you'll have to accept that you can't use paths
like that. Sorry,
> Path object disallows access to S3 objects with // in their names
> -----------------------------------------------------------------
>
> Key: HADOOP-14675
> URL: https://issues.apache.org/jira/browse/HADOOP-14675
> Project: Hadoop Common
> Issue Type: Bug
> Components: fs/s3
> Affects Versions: 2.7.1
> Environment: Plain vanila Hadoop 2.7.
> Reporter: Andi
>
> Seems that there's explicit code to disallow FS paths with //.
> In our environment this means that direct access using url's like
> s3a://bucket/data//part1 are inaccessible.
> Why does this code exist? Can it be removed?
> https://github.com/apache/hadoop/blob/6fefb8f4a486b1d8071a7071e9e96f0c6299752d/hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/Path.java#L289-L289
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
