Real World Computing
Restore to basic backup
On these pages in columns past, we've talked about various open-source backup systems, including Amanda and Bacula, which are both highly admirable, if rather complex, network backup utilities. But this month, I've been taking a look at a rather different - and very promising - newcomer to the scene, called simply Restore. Before going any further, I can't resist noting what a bloody awful name this is for such a promising backup program! When you're doing a web search to try and discover other people's experiences with a particular product (as every user with any real nous will do nowadays), or to resolve some problem, having a sensibly unique program name to search for makes life an awful lot easier, rather than sifting through the tens of thousands of results returned by searching on such a pathetically generic name...
Restore is a Linux-based product and is available either as a set of Debian/Ubuntu packages (if you've already got a Linux machine lying around), as a "Live CD" that lets you try the program or install it to your hard disk using a minimal version of Ubuntu Linux, or as a Virtual Machine Appliance that runs under VMware's virtualisation product VMware Workstation. This third option is a very nice solution for those who have Windows servers and don't want to dedicate an entire machine to Linux.
Installation is a breeze, as is running from the live server (I didn't test the virtual machine version, although comments on the website from other users suggest it works just as well). A full graphical user interface gets installed (or runs from the Live CD), so even regular Windows users needn't be scared off by the big bad command line. In point of fact, once it's performed the basic step of finding out what IP address your server is running at, you won't need to access the backup machine directly at all. That's because the entire system is accessed via a web browser from a client machine, and once you're connected you can add clients (your own machine or others on the network), view and modify schedules, and run immediate backups, all via the web interface.
Creating a new backup is simple enough. As an example, to back up part of a Windows filesystem, you'll first need to share that filesystem on the network in the usual way (of course, you'll want to restrict the share to a specific username and password, unless you want everyone to have access to it). Then from the main Restore screen, choose Filestores and click on the Add Target button to add a new filestore. You'll see that you can choose between FTP, SFTP, Windows File Share and MySQL - yes, you can back up your MySQL database servers using Restore, too.
If you chose Windows File Share, you'll be asked for the hostname, username and password to connect to, and a few seconds after clicking the Next button you'll be presented with a tree view of all the shared directories and files on the source machine. Select those directories you want to back up, hit Next, supply a name for the target on your backup server ("my documents backup", for example) and you're done. At this point, you're at the Snapshots window, which displays all the backups currently held on the system. Of course, at present it will be showing that no backups have been performed, so click on the icon at the top right of the screen that looks like a [>] Play button, and your first backup will be taken.
And that, at its simplest form, is all there is to Restore. While the backup is running, a console window will inform you of what's taking place, which hopefully will be that all your files are being added to the backup set and then copied. Backups on the systems we tested ran quickly and we had very few problems once we'd set up our firewalls and the like to allow traffic to actually pass through them.
