Using tar with RAB

Usually, RAB will create backups by copying files either to a removable backup medium or to some CD. Nevertheless, RAB may also be configured to save some directory trees as compressed tarballs. This feature is intended to be used as a special exception for a few specific directories. If you are primarily interested in backups using tar(1) RAB probably is not the right tool for your needs. In fact, tar(1) support in RAB is somewhat limited. For example, RAB will not try to split large tarballs to multiple CD's, and may use a large amount of space on your hard drive to store temporary files.

Each directory tree going to be used with tar(1) has to be configured using a tarball configuration command as described in the next section. However, RAB will not call the tar(1) command directly, but will use a wrapper shell script tar.sh. This shell script is expected to be an executable file in the library directory used when installing RAB, i.e. /usr/local/lib/rab. The current version of RAB pretty much expects to find GNU tar. Otherwise, the RAB build process will install a generic script, which works but is probably far from being optimal. In these cases, you are free to modify the supplied tar.sh script. Of course, the script may also be replaced by a customized version tailored to your special needs.

Tar.sh expects to be called using the following synopsis.

tar.sh destdir tarfile basedir filelist

The first two of these arguments destdir and tarfile are the directory the tarball is going to be created in, and the name of this tarball, respectively. The directory basedir is the toplevel directory of the tarball tar.sh is expected to create, and the final argument filelist is a list of files to be packed into this tarball. The entries in this list are separated by null characters, and each entry is a path relative to basedir.

When using GNU tar, the default tar.sh script will write compressed tarballs using either gzip(1) or bzip2(1) if your version of tar(1) supports the option -j.