If you’re like most people, backing up your important data is an afterthought at best - at least until the day comes that you lose all your pictures, email, important documents, and that doctorate thesis you’ve been working on for the past few years. Even then some of these same victims of data loss become lax in their backup practices over time putting themselves at risk for another data loss event. What is it they say about doing the same thing twice and expecting a different outcome?

When you run several servers and you have a computer count that is in the double digits, backing up all the data on those servers and workstations is a serious matter. Several years ago, I bit the bullet and started searching for a great solution. After evaluating many backup applications that just didn’t work out, I hit the mother lode and found one of my favorite applications, BackupPC developed by Craig Barratt. I have been successfully using BackupPC for a while now, and a year or so ago I turned the BackupPC server into its own embedded network appliance running off of compact flash (and storing the data pool to a 300GB IDE hard drive).

About BackupPC
BackupPC is an open source application written in Perl. BackupPC runs as a server, connecting to clients, and extracting and saving backup data on disk. BackupPC has a very nice web-based user interface and an extensive set of configuration options that provide flexibility to specify the exact backup policy you desire. BackupPC is compatible with Windows, Linux, Mac and pretty much any operating system that supports SMB protocol (i.e. Windows file sharing) or rsync

BackupPC Media Management
One of my favorite features of BackupPC is its media management. BackupPC makes efficient use of the backup media by implementing backup data pooling and compression. Duplicate files encountered in a given backup set, across sets, across workstations, or any combination of these are managed such that only one real copy is kept in the backup data pool and the rest are “links” to this master copy. For example suppose Alice and Bob are two Mechanical Engineers working on a large SolidWorks design. Suppose they both have the latest revision of this SolidWorks design checked out onto their local workstations (you do use version control don’t you?). Suppose BackupPC makes a backup of Alice’s computer first. When it gets around to backing up Bob’s computer, it realizes it already has a copy of the Solidworks file and creates a hard link to the file backed up from Alice’s computer, saving a network transfer and backup disk space. BackupPC also can be configured to compress the backup data pool stored on the backup disk, and it only has to run the compression algorithm once per unique file thanks to the data pooling algorithm BackupPC employs. BackupPC can also be configured to create permanent backups of the backup data pool to removable media such as DVD or tape for offsite storage.

Versioned Backups
BackupPC’s efficient use of the backup medium allows it to provide versioned backups of your servers and workstations without the worry of quickly exceeding the capacity of the backup disk. BackupPC easily manages full and incremental backup sets, automatically merging incremental sets into their corresponding full backup set. Restoring a file is simply a task of logging into the web-based user interface, selecting the full or incremental backup desired, and navigating the file tree and selecting the files you want to restore.

My BackupPC Media Stats:
To give an idea of the amount of data and number of versions you can keep around with BackupPC, I included a few screenshots from the user interface. I have a 300GB drive as my backup medium. First, the uncompressed statistics:

Now the backup data pool statistics on compression and disk usage:

So I’ve only used around 86GB (37%) of my backup disk, even though I backup seven hosts and have around 280GB of raw backup data. Pretty impressive, in my opinion.

BackupPC User Interface
BackupPC has a great web-based user interface that allows both administrator and end-user access allowing the appropriate privileges for each role. Like I mentioned above, selecting backup sets and files is a simple task. Here are some screenshots indicative of the Backup Summary and Backup Browsing available on the BackupPC website.

Other Backup Options
To be complete I thought I’d mention a few other backup options that seem worthwhile and briefly discussion pros and cons of each.

RAID vs. Traditional Backup
RAID is nice if you can afford it, and especially if you can afford something running RAID 5 or a variant thereof (Check out Infrant’s ReadyNAS if you’re interested). The only other reason you should consider traditional backup methods is because RAID is primarily a solution for hardware fault tolerance. It won’t save you when you’ve accidentally deleted the last available copy of that document or source tree you’ve been working on.

Real-time Backup with Carbonite
I read about the Carbonite real-time backup service in the Boulder Daily Camera a few months ago. Their service looks very compelling as they’re offering real-time backup of unlimited data for $49.95 per computer per year. The only main drawback with the Carbonite service is that it requires installing a backup client. At the time of this writing that client is only available for Windows XP. A Mac version appears to be in beta, no Vista support yet, and I imagine no Linux either, but that will probably change as they develop their product for other operating systems. I would almost list the price as a drawback, but for $5 per month per computer per year, well seems like a steal to me. Wonder how long they can continue that pricing model?

Real-time Photo Backup with ProtectMyPhotos
Interested in just protecting your photos? Check out ProtectMyPhotos which seems to be a service similar to Carbonite but specific to backing up your photos and media. It’s also a Boulder company - you can read some coverage on this company here.

EMC Retrospect
EMC Retrospect seems to be the leader in commercial software backup options. With versions available for Windows and Mac, as well as server versions and the ability to backup pretty much any operating system, Retrospect boasts a number of features that I’ll let you peruse at the link provided. Retrospect’s main drawback is its hefty price tag. From a single computer version starting at $119 to their Small Business Server standard edition at $479, and their multi-server edition at $1409, BackupPC at a price tag of $0 just might be the way to go.

This entry was posted on Saturday, January 27th, 2007 at 10:57 pm and is filed under Best of Tech, Disruptive Tech, Net Appliances, Use and Recommend. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.