[jira] [Commented] (JCLOUDS-768) S3 creating container without location

["Andrew Gaul (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/JCLOUDS-768?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14196090#comment-14196090 ] 

Andrew Gaul commented on JCLOUDS-768:
-------------------------------------

[~ahajnal] Invoke {{ContextBuilder.newBuilder}} with {{"s3"}} instead of {{"aws-s3"}}.  The latter is specific to Amazon while the former works with S3-compatible blobstores.

> S3 creating container without location
> --------------------------------------
>
>                 Key: JCLOUDS-768
>                 URL: https://issues.apache.org/jira/browse/JCLOUDS-768
>             Project: jclouds
>          Issue Type: Improvement
>          Components: jclouds-blobstore
>    Affects Versions: 1.8.1
>         Environment: All
>            Reporter: Akos Hajnal
>
> Dear Developers,
> I would like to use jclouds to access an S3 compatible storage (Ceph Object Gateway), but not Amazon. Many functions work well, but when I try to create a container (bucket) with:
> blobStore.createContainerInLocation(null, "jclouds"); 
> jclouds tries to use a default Amazon location, even if I created blobStore with my own endpoint:
> 	BlobStoreContext context = ContextBuilder.newBuilder("aws-s3")
> 				 .endpoint("https://s3.lpds.sztaki.hu")
> 	             .credentials("id","pass")
> .buildView(BlobStoreContext.class);
> I get exception from Amazon's server that my id is not there (that is right):
> The AWS Access Key Id you provided does not exist in our records.
> (With AWS SDK I used to have similar problem at specifying no location. With the latest API createBucket(bucketName) works well on own S3 storages as well.)
> Regards,
> Akos Hajnal



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)