[
https://issues.apache.org/jira/browse/HADOOP-14661?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17235493#comment-17235493
]
Steve Loughran commented on HADOOP-14661:
-----------------------------------------
(I see looking @ your AWS issue comments that our request factory plan wouldn't
have worked. And we would have had to find a requester pays bucket with > 1
page of entries to find the list problem. nice.
FWIW the ability to overwrite the factory was mainly to let us plug in an
inconsistent s3 client. you've found a new use.
Be aware of a problem you may hit with spark and hive's isolated classloader
https://github.com/apache/hadoop/blob/trunk/hadoop-tools/hadoop-aws/src/site/markdown/tools/hadoop-aws/troubleshooting_s3a.md#class-does-not-implement-awscredentialsprovider
> S3A to support Requester Pays Buckets
> -------------------------------------
>
> Key: HADOOP-14661
> URL: https://issues.apache.org/jira/browse/HADOOP-14661
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: common, util
> Affects Versions: 3.0.0-alpha3
> Reporter: Mandus Momberg
> Assignee: Mandus Momberg
> Priority: Minor
> Attachments: HADOOP-14661.patch
>
> Original Estimate: 2h
> Remaining Estimate: 2h
>
> Amazon S3 has the ability to charge the requester for the cost of accessing
> S3. This is called Requester Pays Buckets.
> In order to access these buckets, each request needs to be signed with a
> specific header.
> http://docs.aws.amazon.com/AmazonS3/latest/dev/RequesterPaysBuckets.html
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
