Ottawa PC Users' Group, Inc.
Linux - Part 9
by Alan German
I now have a
fully-functional, production machine running Ubuntu Linux
and so, in this latest in the series of articles looking
at various aspects of Linux, it's time to turn our
attention to maintenance issues and, in particular, to
backup systems. Those who have read some of my previous
reviews of Windows-based utilities will know of my
fondness for file synchronization programs. So, my
initial foray into backup mechanisms for Linux has
followed this path. We will take a look at one
open-source file synchronization utility that will
provide us with a simple backup system, and learn a
little about script files in the process.
I have a spare 256 MB compact flash memory card from an
old digital camera kicking around. And, I am currently
working on a project that involves receiving lots of word
processing files by E-mail, compiling these into a single
document, sending this out for review, and receiving
feedback from multiple reviewers. It's important to
maintain a backup of all of the files associated with
this project in order that nothing gets lost in the
shuffle. So, my initial goal is to mirror my working
document directory onto the compact flash memory card.
This needs to be done in a simple fashion that will allow
me to easily and quickly make backups of the file
structure as it is modified.
Linux has a great little utility rsync that
will do precisely what I require. To quote from the
program's documentation (man rsync): Rsync copies
files either to or from a remote host, or locally on the
current host. Rather than using the powerful
communication capabilities of the program, we will merely
transfer files between two disks on the local computer,
in which case rsync serves as an enhanced copy command.
Rsync is a command-line program, so we will need to run
it in a terminal window (Applications Accessories
Terminal). The primary format of the command used
options source destination
There is an
almost mind-numbing array of optional commands (see
rsync's manual for full details), but we will use just
three of them: a (archive), v (verbose) and delete. Our
source directory will be specified as /mnt/windows_data/carsp/cmrsc_18/
while the flash memory card is seen by the filesystem as /media/disk.
Thus, the command string to be entered in terminal's
-av --delete /mnt/windows_data/carsp/cmrsc_18/
retains file attributes, such as ownership and
permissions, when the files are transferred. Switching on
verbose mode means that a complete list of the names of
the files that are transferred as a result of the command
is displayed. The delete option (note the double leading
dashes required here) causes any files that are present
on the destination drive but not on the source to be
erased. The result is that any new or modified files are
transferred from the source to the destination, any
existing files on both systems remain unchanged, and any
old files on the destination that have been deleted from
the source are deleted on the destination. Thus, running
this command produces a mirror of the files
currently located in the source directory to appear on
the destination drive.
There are a couple of important tricks to
note in the specifications for the source and destination
drives. Firstly, there is a trailing slash on the cmrsc_18
sub-folder specified as being the source. This tells
rsync to copy all of the files (and sub-folders) from
this folder on the source to the specified destination.
But, most importantly, it says: don't create a cmrsc_18
folder on the destination. If we were to omit the
trailing slash, a folder within a folder (i.e. .../cmrsc_18/cmrsc_18)
would be created on the destination drive, which is not
what we want. The other trick with the command is that,
on the first time around, the cmrsc_18 folder is
created on the destination drive if it doesn't already
exist. With subsequent implementations of the command,
the files (and folders) in the destination folder are
modified to precisely match those in the source folder.
If the above seems to be a complex explanation of the
nuances of the command structure, try a few simple tests.
Use two temporary directories, temp (source) and temp2
(destination), with just a couple of small text or image
files in temp, and see what happens when you include the
trailing slash, or leave it out.. Try also adding,
editing and deleting files in the source. Run the command
and check that all your changes are indeed mirrored in
the destination folder.
Given the directory path that I wish to be specified, the
command line is rather long. And, perhaps I don't want to
be bothered re-typing such a long command every time I
wish to run the backup process. We should recall that
previously-used commands are stored in a buffer, and can
be accessed by scrolling through the command history
(using the up-arrow). So, that's one way to reuse the
command line. Another is to store the command in a
special kind of file and run the command by typing the
name of the file. Does anyone remember batch files in
DOS? Welcome to the Linux world of shell scripting!
In particular, we are going to create a script file (.sh)
that we will run in bash (the Bourne Again Shell). While
a simple script will accomplish our desired task, it's
worth noting that bash scripting is a very powerful
programming technique. Because of this, many books and
on-line tutorials are available to show you how to make
effective use of the system.
But, for our present purposes, we need only limited
knowledge of bash scripts. In fact, our script consists
of just three lines of code which we type into gedit
(Applications Accessories Text editor) and
save as the file mirror_cmrsc18.sh in our home
echo "Mirror cmrsc_18 to compact flash memory
rsync -av --delete /mnt/windows_data/carsp/cmrsc_18/
line indicates that the file is a shell script and
identifies the location of bash as being in the /bin
folder. (You can check this location on your system by
issuing the command which bash.)
The second line (quite wordy by Unix programming
standards) serves to document the purpose of
the script by displaying a message on the display screen
prior to activation of rsync.
The third line is merely the command line that we would
have typed manually in a terminal window to run our
selected backup process.
Note that the file name chosen for the script
(mirror_cmrsc18.sh) is also rather long. Perhaps a
shorter name would be more desirable. It would certainly
be easier to type. But, in this case, it does tell us
(well, it tells me!) what the script will do.
Now we have our script, we need to know how to run it.
Firstly, in a terminal window, we navigate to the home
folder and issue the command chmod +x mirror_cmrsc18
to give ourselves permission to execute the file.
Alternatively, pull up the Nautilus file manager
(Places), right click on the file mirror_cmrsc18.sh,
select Properties Permissions, and check the box
marked Execute Allow executing file as a program.
Note that if you try to run the script file without
taking this step, bash will return the error message:
Now, we simply run the script file with the command ./mirror_cmrsc18.sh.
Note the use of a dot and a slash (./) in front of the
script file's name. The dot indicates that we wish to use
the current directory, and the slash is a separator
between the directory name and the filename. The
requirement for this seemingly obscure format is a
security feature in Linux. The current directory is not
automatically on the path and so we must specify that the
script file is to be run from the current directory by
prefacing the file name with ./. Note also
that it is necessary to include the .sh file extension in
But, that's it. The rsync program runs, transfers files
between the two disks, and lists the files processed. Our
files are safe. We have backup!
Now, let's make our bash script just a little more
sophisticated by checking that the memory card is
available before we try to run rsync. A few extra
commands inserted into the script file will do the trick:
echo "Mirror cmrsc_18 to compact flash memory
# Check that the memory card is available
if [ -d /media/disk/cmrsc_18 ]; then
rsync -av --delete
echo "Insert memory card and try
line commences with the # symbol that defines this line
as a comment. The workhorse statements are the
if-then-else sequence in the last five lines.
These either run rsync, or provide a warning message and
exit gracefully. The command
-d /media/disk/cmrsc-18 checks if the directory is
present on the removable drive. If it is, we can proceed
to back up our files. If not, we need to insert the flash
So, if you yearn for the days of DOS batch files, bash
scripts in Linux can bring back a whole world of
enjoyment (and/or frustration!) As noted previously, bash
scripting is a very powerful tool. Check out the wide
variety of available commands, and the multitude of
programming techniques, that can be used for all sorts of
different purposes. Lots of on-line assistance is
available to you, including pre-built scripts that you
can readily customize for your specific applications.
Andrew Tridgell and Paul Mackerras
Introduction to Bash Scripting
Click here to view the
full OPCUG website with frames.
Copyright and Usage
Ottawa Personal Computer Users' Group (OPCUG), Inc.
3 Thatcher Street, Ottawa, ON K2G 1S6
opinions expressed in these reviews do not necessarily
represent the views of the OPCUG or its members.
comments or suggestions to the .