S3 Device

Amazon S3 is a simple storage service (S3, get it?) provided by Amazon.com. Before you get started with the S3 device, you should familiarize yourself with the service itself, and sign up for an account. Once you've signed up, you'll get two different keys -- an access key and a secret key.

Configuring

Configuring the S3 plugin is relatively easy; put the following in your amanda.conf:

tapedev "s3:bucket/prefix-"
device_property "S3_SECRET_KEY" "your-secret-key"
device_property "S3_ACCESS_KEY" "your-access-key"

The bucket and prefix in the device name determine where volumes are stored. Each volume consists of a lot of objects on S3, but all objects will be stored in the named bucket and start with the named prefix. You can therefore store multiple volumes in the same bucket by using different prefixes. Since bucket names are global across all Amazon customers, buckets named 'test', 'backups', and 'DailySet1' are probably already taken. We recommend that you name your bucket with your access key and some helpful suffix, e.g., '009809AB0981-daily'. Inside a bucket, then namespace is all yours, so the prefix can be as simple as 'slotNN-'.

It doesn't matter what tapetype you use; tapetypes will be used for estimation only and not for limiting storage use (since Amazon S3 provides essentially unlimited storage).

If you want to use multiple buckets, configure the chg-multi changer.

See Also

For more detailed instructions, see How To:Backup to Amazon S3.