[Sdruby] Backing up an S3 bucket

[Nick Zadrozny]

On 7/30/07, Matt Aimonetti <mattaimonetti at gmail.com> wrote:
>
> I don't know if the files would actually be copied to your server and then
> saved back on a new bucket, hopefully not.


Actually I think so. Last I knew, renames can't even be done in place. I use
JungleDisk to backup stuff to S3 and they have a post on their blog about
how they use EC2 to support fast copies and renames:

http://blog.jungledisk.com/2007/02/22/fast-file-rename-and-copy/

It's more marketing than technical information, but has a few hints on
implementation I guess. I don't keep all that close an eye on the nuts and
bolts of S3, so, who knows, fast server side renames and copies might be
possible at this point. Either way, no magical rake tasks from me, but maybe
this will help your research.

-- 
Nick Zadrozny • beyondthepath.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://comox.textdrive.com/pipermail/sdruby/attachments/20070730/fc93961b/attachment-0001.htm