[
https://issues.apache.org/jira/browse/HADOOP-15140?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16336780#comment-16336780
]
Chris Douglas commented on HADOOP-15140:
----------------------------------------
I can't claim any grand insight, unfortunately. The semantics of an empty path
in the URI aren't crisp, since {{FileSystem}} instances maintain working/home
directories as context to resolve relative paths. In the past, that's gotten in
the way of defining clearer semantics, since FS overrides don't compose
compatibly. It's one of the reasons {{FileContext}} tried to move all the path
manipulation up a level, so the FS implementations only deal with absolute URIs
(IIRC, [~sanjay.radia] may have a clearer memory of this).
bq. To clarify the failure only occurs when passing a URI as an argument to the
Path constructor. Otherwise this is handled properly.
Can you expand on this?
bq. this surfaces in code which has a path p like s3a://bucket/file.txt and you
can getFileStatus(p.getParent()). The key of that parent is being mapped to "/"
and then s3guard is failing.
If I understand you, the path component of the URI should be resolving to "/",
which s3guard should resolve (even as a special case). Is it getting dropped?
> S3guard mistakes root URI without / as non-absolute path
> --------------------------------------------------------
>
> Key: HADOOP-15140
> URL: https://issues.apache.org/jira/browse/HADOOP-15140
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: fs/s3
> Affects Versions: 3.0.0
> Reporter: Steve Loughran
> Assignee: Abraham Fine
> Priority: Major
>
> If you call {{getFileStatus("s3a://bucket")}} then S3Guard will throw an
> exception in putMetadata, as it mistakes the empty path for "non-absolute
> path"
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: common-issues-h...@hadoop.apache.org
