backup question: rsnapshot + amazon S3

Using s3 over fuse is notoriously unstable, I believe http://code.google.com/p/s3backer/ is generally a more preferred approach, (or you could be like me and use simple put requests using curl).

To cut down on the amount of traffic sent over the wire, I usually go with duplicity. It maintains the current state of the remote store locally, so it doesn't have to list and download files to check them. I also upload a couple specific files individually (such as mysqldumps) so they're restorable without duplicity.

I think rsnapshot relies on hard links, which may not work with an S3 fuse filesystem; you might have to resort to s3backer for things to work right.

I'm using cloudcity with rackspace cloud files.

It's stable and cheaper than s3.

You only need python-cloudfiles and duplicity packages.

@hoopycat:

I think rsnapshot relies on hard links, which may not work with an S3 fuse filesystem; you might have to resort to s3backer for things to work right.

Thanks everyone.

Today I found that Rsnapshot DOESN'T work on amazon S3 so I will try duplicity

thanks again

Duplicity