[
https://issues.apache.org/jira/browse/HADOOP-9248?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13562859#comment-13562859
]
Steve Loughran commented on HADOOP-9248:
----------------------------------------
# there's always the option of extracting the s3 endpoint from the fqdn in the
url, eg. s3.atmos.myorg.org
# what we're doing in the HADOOP-8545 OpenStack storage support is adding not
just endpoint configuration, but supporting different authentication
credentials for each one.
you define a an openstack service by name, eg "mycluster", then specify the
containers by name.cluster in the URI -which triggers a new lookup of
fs.swift.service.mycluster.* credentials.
I'd propose something like that for S3, on the basis that once you support
endpoints other than AWS, you start thinking about supporting >1 endpoint...
> Allow configuration of Amazon S3 Endpoint
> -----------------------------------------
>
> Key: HADOOP-9248
> URL: https://issues.apache.org/jira/browse/HADOOP-9248
> Project: Hadoop Common
> Issue Type: New Feature
> Components: fs/s3
> Environment: All environments connecting to S3
> Reporter: Timur Perelmutov
>
> http://wiki.apache.org/hadoop/AmazonS3 page describes configuration of Hadoop
> with S3 as storage. Other systems like EMC Atmos now implement S3 Interface,
> but in order to be able to connect to them, the endpoint needs to be
> configurable. Please add a configuration parameter that would be propagated
> to underlying jets3t library as s3service.s3-endpoint param.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
