[
https://issues.apache.org/jira/browse/HADOOP-15140?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16346095#comment-16346095
]
Steve Loughran commented on HADOOP-15140:
-----------------------------------------
* Well, you've got the evidence that there's nothing wrong with s3guard or s3a,
just that it was there where I found it.
* and, with HDFS setting the policy, we are consistent.
* Most likely to fail here is bits of spark or hive which pass around URIs; you
may get jobs which fail if the test is the root of the bucket. Unfortunate,
but, well, not so common a use that we get lots of complaints.
Local FS is the odd one as it varies from wherever you run the tests (and the
JVM), so we can't use it as a reference for anything, and trying to make it
consistent with HDFS in terms of behaviour or error reporting not something you
can ever complete.
I think we may want to consider this "at least consistent with HDFS", retitle
it S3A (as it is more than just s3guard), and mark as a WONTFIX, or at least
pull in the test which works consistently.
> 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: [email protected]
For additional commands, e-mail: [email protected]
