[
https://issues.apache.org/jira/browse/WHIRR-286?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13021187#comment-13021187
]
Andrei Savu commented on WHIRR-286:
-----------------------------------
> For Hadoop it would be great to generalize the NativeS3FileSystem to use any
> jclouds BlobStore
Cool project. Maybe next year at GSoC :)
> Add CLI commands so that Whirr can be used as a minimal blob store client
> -------------------------------------------------------------------------
>
> Key: WHIRR-286
> URL: https://issues.apache.org/jira/browse/WHIRR-286
> Project: Whirr
> Issue Type: Improvement
> Reporter: Andrei Savu
>
> I believe it would be helpful in many workflows if you could create and
> delete blobs just by using the Whirr CLI (data used for processing). I
> suggest that we should add new commands to the CLI so that Whirr can be used
> as minimal generic blob store client that's also aware of the location of the
> nodes (a typical mistake is to upload files to a different location).
> Workflow (common?):
> {noformat}
> $ whirr create-container --config.. --name data
> $ whirr upload-file --config.. --file local.tar.gz --container data --blob
> data.tar.gz
> $ whirr launch-cluster --config..
> # start a map reduce job and collect the results
> $ whirr download-file --config.. --file output.tar.gz --container data --blob
> output.tar.gz
> $ whirr destroy-cluster --config..
> $ whirr destroy-container --config.. --name data
> {noformat}
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
