**NAME**

  tbku - Table-driven backup script


**SYNOPSIS**

  ``tbku allsets | [fileset] ...``


**DESCRIPTION**

  ``tbku`` is a utility script for producing "tarball" backups of
  some- or all of your file systems.  It is useful both for producing
  incremental backups and for systemwide images or "snapshots".  The
  script can be run either from the command line or, more typically,
  as a ``cron`` job to automate system backup tasks.

  ``tbku`` was originally developed as a backup tool for FreeBSD
  systems.  Since then, it has been updated to also work with SUSE
  Linux.  ``tbku`` should work with little- or no modification on any
  other Unix-like system.


**HOW tbku WORKS**


  This program produces tarballs of backup "sets".  The tarball is
  written to $BKUDIR.  By default, this is '/bku' but you can override
  this with the $TBKUDIR environment variable to point to any
  directory you wish.  If the directory does not exist, 'tbku' will
  create it for you when it runs.

A backup "set" consists of a related set of files and directories
that are all to be backed up together into a single tarball.  The
file containing all the names of all the files and directories in a
given backup set is called the "fileset".  'tbku' expects to find
filesets in $FILESETDIR.  By default, this is '/root/tbku', but you
can override this by setting the $TBKUSETS environment variable to
point to the directory of your choice. 

Notice that if this directory does not exist OR if the fileset you
name does not actually exist, 'tbku' will appear to run, but will
produce an empty backup set with an error in it's log indicating
that it could not find the named fileset.  A backup that fails for
this reason will also be flagged by having the "*" character in its
backup, log, and timestamp filenames.

The naming of filesets is important.  Fileset names must end in
".backup name".  If the name begins with $COMMON, then this program
understands that set to be done "automatically".  That is, if
nothing is passed on the command line, the program does an
"automatic" backup - i.e., It processess ALL the filesets found in
the backup directory whose names begin with $COMMON.

The program can also be started with command line arguments.  It
understands these to be the name of the backups desired.  For
instance:
           tbku foo bar baz

Will look for filesets ending in

".foo", ".bar", ".baz".  

In this case, the program does not care what the beginning of the
fileset name is.  This can potentially be a problem if you have two
filesets whose names end identically.  Say you have 'manual.bku.foo'
and "auto.bku.foo" in your $FILESETDIR directory and you issue the
command "tbku foo".  BOTH of these filesets will be processed since
they both have the backup name "foo".  The last one to be processes
will be the one whose backup tarball ends in $BKUDIR.  The simple
way to avoid this is to never name two filesets with the same ending
backup name.

Do NOT use the special setname ".allsets".  This name is reserved.
When passed on the command line ("tbku allsets"), it does all of the
backup sets it finds regardless of whether they are set for
automatic or not.

If the setname is "tape", then output will be written to the device
or file specified in the $TAPEDEV variable instead of a file in the
backup directory.

**CUSTOMIZING tbku**

  ``tbku`` is written to be "smart" enough to figure out where your
  system keeps needed tools like ``tar`` or ``sed``.  The only
  requirement here is that ``tbku`` be run in an environment that can
  find the ``which`` command somewhere in its ``$PATH`` - ``tbku``
  uses ``which`` to figure out just where everything it needs lives on
  your filesystem.  If ``tbku`` cannot figure out where your system
  keeps things, it will use the FreeBSD default values.

  For most FreeBSD and Linux users, this should work without any
  customization beyond setting environment variables to override
  default behavior (described below).  In rare circumstances, you may
  need further customization.  All the things you're likely to ever
  want to change appear first in the actual ``tbku`` script, and are
  briefly documented there.


**ENVIRONMENT VARIABLES & DEFAULTS**

  You can override the various ``tbku`` defaults by setting a
  corresponding environment variable.
  

  =================  ===============================  ======================
  **Env. Variable**  **Default Value**                **Meaning**
  -----------------  -------------------------------  ----------------------
  TBKUDIR            /bku                             Where to write backups
  TBKUNAME           $MACHINE-$OSTYPE-$OSREV-$HWTYPE  Tarball base name
  TBKUSETS           $HOME/tbku                       Filesets found here
  TBKUTAPE           /dev/sa0                         Tape device (or file)
  =================  ===============================  ======================


  Examples::

    export TBKUDIR=/mnt/backups    # Backups written to /mnt/backups

    export TBKUNAME=JoeBackup      # Backups named: JoeBackup-<filesetname>

    export TBKUSETS=/tbku          # Looks for filesets in /tbku

    export TBKUTAPE-/tmp/faketape  # Tape backups actually written to *file*


**UPDATES & SUPPORT**

  To get the latest version of 'tbku', go to:
  
    http://www.tundraware.com/Software/tbku

  For questions, comments, or other feedback, send email to:

    tbku@tundraware.com


**AUTHOR**

  Tim Daneliuk, TundraWare Inc.


**COPYRIGHT & LICENSING**

  ``tbku`` is Copyright (c) 2004-2008, TundraWare Inc., Des Plaines, IL, USA

  There is no fee for using ``tbku`` either personally or commercially
  *so long as the terms of the tbku license are met*.  Please read the
  ``tbku-license.txt`` file for a full explanation of the licensing
  terms.

  

**DOCUMENT INFORMATION**

  This document was produced using the very useful
  ``reStructuredText`` tools in the ``docutils`` package.  For more
  information, see:

    http://docutils.sourceforge.net/rst.html

``$Id: tbku.txt,v 1.105 2008/03/19 16:31:36 tundra Exp $``