Server Backup Strategy for Amazon EC2

Your approach sounds very good - but I can think of a possible way to improve it.

To reduce the impact of data loss since the last backup, and EBS volume failure (unlikely, but still possible) you could store your data on a separate EBS volume than your system files, and back up the data volume more frequently than the system volume.

With your current strategy, you'll lose any data that was created between the time of the last backup and the time your instance failed. With the new approach, the data volume will be getting written to right up until the instance failure, so you can just reattach it to your new instance once it's up and running.

Here's what I did. I made LVM snapshots and then used dd to copy them to the external drive and that worked.

As explained in a comment, I considered rsync but it doesn't meet my needs. First, it won't back up a complete running server. And, restoring from that kind of backup doesn't really satisfy my requirements either. So the snapshots were the solution.

BTW, now I'm using btrfs, so snapshots are even easier.