NAME -- Automated interface to create `dump/restore' or `dar' backups.

SYNOPSIS [--test] [--verbose] [--usage|-?] [--help]
              [--date=<string>] [--name-prefix=<string>]
              [--dump-program=<dump-prog>] [--[no]dar]
              [--gzip | -z] [--bzip2 | -y] [ --compression[=[algo:]level] ]
              [--dest-dir=<destination-dir>] [--dump-dir=<dest-dir>]
              [--partition=<block-special-device> | <partition> ]
              [--level=<digit> | <level>]


This script creates and verifies a backup dump using the `dump' and `restore' programs, or the `dar' program, both of which are linked below.

The product of this procedure is a set of dump files on disk somewhere that has been verified against the backed-up partition. More than one file may be required if the dump is to be written to offline media; in that case, use the --volsize option to limit the maximum file size. If not supplied, a suitable series of file names is chosen based on the partition mount point, current date, and backup level. Optionally, the file can be moved to somewhere else in the destination file system after it has been verified; this makes it easy to use the script to write the resulting dump file(s) to a CD. The whole process is readily automatable via cron jobs.

[Writing to tape probably doesn't work; I have no tape on my system, so I don't know how to rewind it. -- rgr, 21-Oct-02.]

Each dump file name looks something like home-20021021-l9.dump or home-20051102-l0a.dump, or home-20051102-l0.17.dar for DAR, and consists of the following five components:

  1. A prefix tag (e.g. home). This is normally the last component of the directory where the partition is mounted, but can be specified via the --name-prefix option. The tag is arbitrary and may consist of multiple words, as in usr-local; its purpose is solely to group all of the backups make for a given partition.

  2. The date the backup was made. This is normally the current date in "YYYYMMDD" format, e.g. '20021021', but can be overridden via the --date option.

  3. The dump level as specified by --level, a digit from 0 to 9, with a lowercase "L" prefix, as in l9.

  4. An optional volume suffix. DAR creates them by appending .#.dar to the stem, where "#" is a number starting from one; all DAR backups always have an explicit volume suffix. If a dump backup requires two or more volumes, then alphabetic volume designators are assigned from 'a', as in home-20051102-l0a.dump, home-20051102-l0b.dump, home-20051102-l0c.dump, etc. This is only necessary for large backups that are later copied to physical media. [Note that mkisofs imposes a limit of 2147483646 bytes (= 2^31-2) on files burned to DVD-ROM; see the "DVDs created with too large files" thread at for details. -- rgr, 4-Nov-05.]

  5. A file suffix (extension), which is ".dump" for dump backups, and (not surprisingly) ".dar" for DAR backups.

If the --file-name option is specified, then it overrides the first three components. There is no way to override the volume suffix or file suffix.



If specified, no commands will be executed. Instead, the commands will just be echoed to the standard error stream.


If specified, extra information messages are printed during the backup. Since the output of dump and restore are included unedited, the default output is pretty verbose even without this.


If specified, gives the name of the dump file excluding the directory. The default looks something like home-20021021-l9.dump, and depends on (a) the last component of the directory where the partition is normally mounted, e.g. 'home', (b) the current date, e.g. '20021021', and (c) the dump level, e.g. '9'. Suffixes of 'a', 'b', etc., will be added after the dump level if needed for a multivolume dump.

If some of the default values are not acceptable, you can either specify a specific file name, or use the --name-prefix or --date parameters to override how the default name is constructed. For example, if you had '/usr/local' and '/seq/local' partitions and needed to make the resulting file names distinct, you could say "--name-prefix=usr-local" for the first, and "--name-prefix=seq-local" for the second.


Overrides the date value in the default dump file name; see the description of the --file-name option.


Overrides the partition abbreviation (the last file name component of the mount point) in the default dump file name; see the description of the --file-name option.


Specifies the names of the dump and restore programs to use. The defaults are '/sbin/dump' and '/sbin/restore', respectively, followed by whatever it can find on $PATH.

If you specify a --dump-program that ends in "dar", will assume the use of --dar, and use '/usr/bin/dar' as the default --dump-program. If --dar is specified or implied, then --restore-program is ignored.


Use the DAR (Disk Archiver) program to create the dump. If you specify a --dump-program that ends in "dar", will assume the use of --dar. In order for incrementals to work, you must have a previous catalog or dump set in the same destination directory. See dar for details.

If you specify a full (level 0) DAR dump, will automatically create a catalog of it using the base name of the dump plus "-cat", e.g. home-20080521-l0-cat.1.dar for a home-20080521-l0.*.dar full dump set. This is so that we can create L1 dumps of everything since the full dump without having to keep all of the full dump around, which DAR would otherwise require.


Specifies the bzip2 or gzip compression level; the default is no compression. The optional integer values are for the compression level, from 1 to 9; if omitted, a value of 9 (maximum compresssion) is used. Note that these options are only available for DAR.


The value of "what" can be "ctime" (the default) to specify that the inode times are to be modified in order to preserve the "atime" at the expense of altering the "ctime", or "atime" to specify that the inode times are to be left alone at the cost of keeping the modified "atime". See the "dar" man page for a more detailed explanation of why this tradeoff is necessary.


Specifies the level of compression desired; the default is none. The value can specify an algorithm from the set gzip, bzip2, or lzo, a compression level of 1 to 9, or both separated by a colon. If no value is supplied, then gzip is the default algorithm, and 9 (the maximum) is the default level. This option is only available for DAR.

Note that this is a synonym for the -y and -z options. Regardless of how the compression is specified, it is always passed on to the dar command using its --compression option, so this will not work with older versions of DAR.


If specified, names a directory to which we should move the dump file after it has been verified successfully. This directory must exist and be writable by the user running The default is "." (the current directory).


Specifies a temporary directory to which to write the dump file. If the backup fails, any dump in progress is left in this directory to aid debugging, without interfering with any job that is supposed to pick up completed dumps. The --dump-dir defaults to tmp underneath --dest-dir, and must be on the same partition as --dest-dir; it is created if it does not exist.


The name of a block-special device file for the ext2 or ext3 partition that is to be backed up, e.g. /dev/hda12. There is no default; this option must be specified. A positional partition argument is also supported for backward compatibility.


A digit for the backup level. Level 0 is a full backup, level 9 is the least inclusive incremental backup. For more details, see dump(8), or my "System backups" page ( The level defaults to 9. A positional level argument is also supported for backward compatibility.


Specifies the size of the largest dump file that the backup medium can hold, in 1 kilobyte blocks. If the backup medium fills up before this limit is reached, then dump will pause and wait for you to "change volumes" (by renaming the current dump file) before continuing, which will mess up the 'restore' phase of the operation. The default depends on the --cd option.

If the dump requires multiple volumes, will instruct dump to write a series of files into the same directory (so the directory must have enough space for all of them). This is ideal for unattended creation of multiple backup files to be copied onto multiple physical volumes at a later time. Just be sure to specify a --volsize that fits on the ultimate storage medium.

Note that dump can do physical end-of-media (tape) and disk-full detection on its own. This is only useful if (a) you are writing directly to the end medium, and (b) you are around to change media when they fill up.


[need some. -- rgr, 7-Jan-03.]


If this script used a backup.conf file, it could get per-site defaults, plus instructions for doing a number of backups at once, This would greatly simplify backup crontab entries; only one would be needed. should refuse to proceed if the size of the dumps it produces are expected to be larger than the free space remaining on the disk. If you can't finish, there's no point getting started. Unfortunately, computing the likely size without actually doing the dump is impossible if compression is used.

The --bzip2 and --gzip options should be supported for dump/restore as well.

If you find any more, please let me know.


Dump/Restore at SourceForge (
DAR home page
System backups ( (


Copyright (C) 2000-2011 by Bob Rogers <>. This script is free software; you may redistribute it and/or modify it under the same terms as Perl itself.




Bob Rogers <>