[ 
https://issues.apache.org/jira/browse/HADOOP-8545?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13529812#comment-13529812
 ] 

Tom White commented on HADOOP-8545:
-----------------------------------

I'm not sure whether the blockstore implementation is the right approach. The 
main drawback is that it's a Hadoop-specific format, which other tools can't 
read, and which will introduce a support burden. The S3 blockstore was 
introduced when there was a limit on the size of files, which has since been 
removed, so if I was doing it today I would probably not have a S3 blockstore 
implementation. I would encourage you to look at more general support for large 
files in Swift, e.g. 
http://docs.openstack.org/developer/swift/overview_large_objects.html.
                
> Filesystem Implementation for OpenStack Swift
> ---------------------------------------------
>
>                 Key: HADOOP-8545
>                 URL: https://issues.apache.org/jira/browse/HADOOP-8545
>             Project: Hadoop Common
>          Issue Type: New Feature
>          Components: fs
>    Affects Versions: 2.0.3-alpha, 1.1.2
>            Reporter: Tim Miller
>            Priority: Minor
>         Attachments: HADOOP-8545-1.patch, HADOOP-8545-javaclouds-2.patch, 
> HADOOP-8545.patch, HADOOP-8545.patch
>
>
> Add a filesystem implementation for OpenStack Swift object store, similar to 
> the one which exists today for S3.

--
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

Reply via email to