About six months ago, my trusty Sony VAIO laptop died – a victim of the infamous DC power jack problem. I ended up building a new desktop (AMD X2 5000, 2 GB of RAM, two hard drives in RAID configuration, lots of fans). In the process, I also took the plunge and switched from Windows XP to Ubuntu Linux. So far, I have been very happy using Linux and sort of suprised how well the desktop compares to Windows (hopefully a later post will summarize my experiences).
But one thing that I missed the both is online backup software that I used to use – Carbonite and Mozy. I have been using Carbonite for over a year and it has paid for itself several times. I was planning to switch to Mozy Pro at the time so I can get their “30 days of changes” feature but my laptop died first. Being a busy kind of guy, I never got around to looking for something on Linux until now. The funny thing is that even though I am spread out among several servers and cloud environments with multiple projects in different online systems, there is still plenty of local data and projects which I rather not move to the web (a subject for another post).
I had several requirements for backups:
- Unlimited storage with my own control over it – ability to backup to Amazon S3
- Encrypted backups with my own encryption key, not accessible to the provider – I don’t want my storage to be subpoenable or accessible to hackers if my Amazon credentials get public
- Keeping 30 days worth of changes like Mozy does – you never know when you need that old file
- Automatic backup at regular intervals
- GUI interface preferred
- Ability to select which folders/files to backup
- Incremental backup
- Bandwidth throttling like Mozy and Carbonite
After looking around, the best open source solution I found was a combination of cron jobs, tar, some sort of an S3 FUSE or shell interface, and lots of hacking. The problem is that I do not have the time to do the hacking which would probably take a few hours. So I ended up trying out JungleDisk which does all of the above but will cost $20 after the 30 day trial. As I wrote in my previous post, it is an easy to use tool that solves a specific problem, one that I would be happy to pay a low fee for rather than do it myself.
As I write this, JungleDisk is copying about 3 GBs of data to S3. I will try to post a followup in a few weeks.