I really like the affordability of Amazon’s S3 service. I don’t need to back up a lot of data, but I do need it available anywhere I have an internet connection. I also need it to be automated. S3 lets me have those features.
I back up my personal e-mail server, my home desktop, and my MythTV box to Amazon’s S3 service. I run weekly full backups of select directories on my desktop, and then incrementals throughout the week. My MythTV box backs up its database on its own, so I just tarball the output and dump it to a folder on my desktop machine, which gets put on S3 when the backup job runs. My mail server runs full backups of select directories every night.
I used some bits and pieces from Paul Stamatiou’s article on S3 backups. You can read that article here.
I added in some bits and pieces for using an include/exclude file, and modified it to run incrementals on my desktop. I also got to use mktemp to do some of my bidding, which was very handy.
All told, my Amazon S3 bill for last month was $1.48 for ~8GB of transfer in, and ~3.7 storage GB/mo. I pair down my stored files weekly to meet some retention guidelines, so that contributed some to my low bill.
But all in all, I’m VERY impressed with S3. I definitely got my $1.48 worth! 