Creating a safe copy of a few directories is governed by laws different from those that apply to backing up complete data centers. The tools need to be especially easy to use, without complicated configuration files or deeply nested menus. The usual space-saving techniques, such as incremental backups, should be available. Pre-and post-backup scripts should be able to stop applications like databases during the backup. Several backup versions need to be archived, and encryption and compression would not be bad.

The backup should be capable of restoring to different hardware; but, for the sake of simplicity, it would be preferable not to store the data in a proprietary format, which would necessitate reinstalling the operating system and backup software in the case of disaster. LVM or RAID configurations should not be an obstacle. File permissions need to remain intact, and the backup should not trip over links or named pipes. Finally, you should also be able to check easily whether everything went as desired.

Treat

One open source tool that satisfies all these requirements is Areca Backup [1]. Beyond the features already listed, Areca has a few more treats to offer. For example, you can include or exclude source files in or from the backup and filter by various criteria. Constructs with AND, OR, and NOT are allowed. Backups can be simulated, so that you can estimate in advance what would reach the backup disk, and to what extent, with the given settings.

Areca can back up to and recover from network drives via (S)FTP. In addition to incremental and differential backups, delta backups store only the changed portions of files. To save even more space, you can compress the backup and merge various archives into a single archive. Files can be restored to their state on a certain date. All critical processes are treated as transactions and can thus be restarted at defined starting points after an interruption. All user actions can be recapped any time, thanks to a history function. On request, reports are generated and mailed.

First Backup

The install essentially just means unpacking. After launching the Java application via the areca.sh shell script (Figure 1), the first step is to configure a backup target. The term "target" is perhaps somewhat misleading, in that it not only defines where to store the backup, but also what to save (the source) and how. Targets can be grouped on request: for example, to group backups logically with related content from various sources (directories). If you want to use this feature, first create a group by right clicking in the upper left panel of the main window and selecting New target from the context menu.

Figure 1: The empty main window after the first start. New targets are defined by right clicking in the upper half of the left panel.

In the dialog for setting up a target (Figure 2) you first need to decide on the directory in which to store the backup. Areca cannot handle tapes. The user specifies the target name and decides whether to create a conventional file-based backup (Standard), back up only the changes within files (Delta), or create a single large archive file, which is updated on each backup (Image). Additionally, you can compress all the individual files of a backup or compress the whole backup as one file. In the latter case, however, interrupted backups can not be resumed.

Figure 2: The dialog box for setting up a backup target in Areca.

The next step is to choose the directories to be backed up (Sources) by simply naming the source folders in the dialog box or dragging and dropping. Next, you set up data compression then, in Advanced, specify whether to include subfolders or follow symbolic links.

Filters lets you define expressions that exclude certain files from the backup. The basis can be file extensions, the date, or file size; you can exclude locked files or special types (e.g., pipes, sockets, and so on), or you can enter regular expressions as exclusion criteria that are applied to file or folder names or both. Filters can be combined, so you can adjust the results to pretty much what you want to back up.

Two items allow you to configure pre- and post-scripts that, for example, stop applications before a backup and then restart them. You can then also specify a set of data to use after resuming an interrupted process. Finally, you can enter a description for each backup.

After you have selected the destination in the right panel, you can launch the backup by pressing the icon with the Plus in the navigation bar at the top. Alternatively, select Run in the menu, or simulate a backup first; the results (Figure 3) show the number of files in the backup, the total data volume, and an accounting of files backed up.

Figure 3: The results of simulating a differential backup in Areca.

Finally, in the backup dialog, you can choose whether to perform an incremental, differential, or full backup (Figure 4), whether the backup should be verified on completion, and whether to add information about the backed up files (manifest).

Figure 4: Specifying the backup.

A log records every step of the backup and allows the admin to ensure retroactively that the backup really has landed on the hard disk – or at least see what kind of error occurred.

And Back Again

"The proof of the backup is in the restore" is a much-cited truism. One initial advantage of Areca is that it is basically superfluous for the restore process of full backups because the files are stored – and possibly compressed – as 1:1 copies in the backup directory; therefore, you could restore files simply by copying them back to your working disk. This is especially helpful after a failure that has taken down the Areca installation. In this case, you do not need to reinstall the backup software first to access the backup.

If you do not want to restore complete archives, but individual files, you can search for the files you need, or you can use regular expressions in the search – either in the current or in all archives. Areca will show you the details of the matches, providing such information as the volume of data that would be restored.

The restore process again offers many options (Figure 5). You can always skip existing files or those files for which the backed up version is older. You can just as easily overwrite, or let Areca prompt you for a decision, and choose whether to restore deleted files. Finally, you can check all the files after the restore.

Figure 5: Various restore options give the user freedom of choice in data recovery.