amanda - The Open Source Backup Platform
This manual page gives an overview of the Amanda commands and
configuration files for quick reference.
Here are all the Amanda commands. Each one has its own manual page. See
them for all the gory details.
There are four user-editable files that control the behavior of Amanda.
The first two are amanda.conf(5) and amanda-client.conf(5), the main
configuration files for the server and client, respectively. They
contain parameters to customize Amanda for the site.
Next is the disklist(5) file, which lists hosts and disk partitions to
Last is the seldom-edited tapelist(5) file, which lists tapes that are
currently active. These files are described in more detail in the
All configuration files are stored in individual configuration
directories, usually under /etc/amanda/. A site will often have more
than one configuration. For example, it might have a normal
configuration for everyday backups and an archive configuration for
infrequent full archival backups. The configuration files would be
stored under directories /etc/amanda/normal/ and /etc/amanda/archive/,
respectively. Part of the job of an Amanda administrator is to create,
populate and maintain these directories.
Most Amanda applications take a "config" parameter; this is generally
the (unqualified) name of the configuration directory, e.g., normal. If
the parameter is . (dot), the current directory is used. This feature
is present for backward compatibility, but is not commonly used.
All log and database files generated by Amanda go in corresponding
directories somewhere. The exact location is controlled by entries in
amanda.conf(5). A typical location would be under /var/adm/amanda. For
the above example, the files might go in /var/adm/amanda/normal/ and
As log files are no longer needed (no longer contain relevant
information), Amanda cycles them out in various ways, depending on the
type of file.
Detailed information about amdump runs are stored in dump logs -- files
named amdump.NN where NN is a sequence number, with 1 being the most
recent file. Amdump rotates these files each run, keeping roughly the
last tapecycle (see below) worth of them.
The file used by amreport to generate the mail summary is the trace
log. This file constitutes the "catalog" describing the data on the
tapes written in a run. It is named log.YYYYMMDDHHMMSS.NN where
YYYYMMDDHHMMSS is the datestamp of the start of the amdump or amflush
run and NN is a sequence number started at 0. At the end of each amdump
run, log files for runs whose tapes have been reused are renamed into a
subdirectory of the main log directory (see the logdir parameter below)
named oldlog. It is up to the Amanda administrator to remove them from
this directory when desired.
Index (backup image catalogue) files older than the full dump matching
the oldest backup image for a given client and disk are removed by
amdump at the end of each run.
For Samba access, Amanda needs a file on the Samba server (which may or
may not also be the tape server) named /etc/amandapass with share
names, (clear text) passwords and (optional) domain names, in that
order, one per line, whitespace separated. By default, the user used to
connect to the PC is the same for all PC´s and is compiled into Amanda.
It may be changed on a host by host basis by listing it first in the
password field followed by a percent sign and then the password. For
With clear text passwords, this file should obviously be tightly
protected. It only needs to be readable by the Amanda-user on the Samba
HOST & DISK EXPRESSION
All host and disk arguments to programs are special expressions. The
command applies to all DLEs that match the arguments. This section
describes the matcher.
The matcher matches by word, each word is a glob expression, words are
separated by the separator ´.´ for host and ´/´ for disk. You can
anchor the expression at left with a ´^´. You can anchor the expression
at right with a ´$´. The matcher is case insensitive for host but is
case sensitive for disk. A match succeeds if all words in your
expression match contiguous words in the host or disk.
If the disk is a UNC ("\\windows\share") then all ´\´ are converted to
´/´ before the match. Using ´\´ is complicated because of the extra
quoting required by the shell and amanda. It´s easier to use ´/´
because it require less quoting ("//windows/share")
word separator for a host
word separator for a disk
word separator for a UNC disk
anchor at left
anchor at right
match exactly one character except the separator
match zero or more characters except the separator
match zero or more characters including the separator
match a single character, namely any of the characters enclosed by
match a single character, namely any characters that is not
enclosed by the brackets.
The shell interpret some of these characters, they must be escaped by a
backslash ´\´ and/or the expression must be enclosed in simple or
Will match hosta, foo.hosta.org, and hoSTA.dOMAIna.ORG but not
Will match host but not hosta.
Will match hosta and hostb, but not host.
Will match hoina but not ho.aina.org.
Will match hoina and ho.aina.org.
Will match hosta but not foo.hosta.org.
Will match /dev/sda1 and /dev/sda12.
Will match the disk opt but not the host opt.
(note dots:) .opt.
Will match the host opt but not the disk opt.
Will match the disk / but no other disk.
Will match the disks /usr and /usr/local.
Will match the disks /usr but not /usr/local.
Will match the disks \\windows1\share and \\windows2\share.
Will match the disks \\windows\share1 and \\windows\share2.
Will match the disk \\windows\share.
A datestamp expression is a range expression where we only match the
prefix. Leading ^ is removed. Trailing $ forces an exact match.
match all dates beginning with 20001212, 20001213 or 20001214
same as previous
match all dates between 20001212 and 20001224
match all dates that start with 2000121 (20001210-20001219)
match all dates that start with 2 (20000101-29991231)
match all dates between 20000101-20101231
match only 200010
A dump specification selects one or more dumps. It has the form
[host][:disk][@datestamp], where each component is a pattern as
described above. If a component is missing, it is treated as a
wildcard. The characters ´:´, ´@´, and ´\´ may be escaped within any
component by preceding them with a ´\´.
all dumps of client17
All dumps on with datestamps matching 20080615
All dumps of /var/www on host webserver
The dump of webserver with datestamp 200806150317
All dumps of /var/www on any host
Most commands allow the override of specific configuration options on
the command line, using the -o option. This option has the form
-oname=value. An optional space is allowed after the -o. Each
configuration option should be specified in a separate command-line
For global options, name is simply the name of the option, e.g.,
For options in a named section of the configuration, name has the form
SECTION:section_name:name, where SECTION is one of TAPETYPE, DUMPTYPE,
HOLDINGDISK, or INTERFACE, and section_name is the name of the
tapetype, dumptype, holdingdisk, or interface. Examples:
amdump -o TAPETYPE:HP-DAT:length=2000m
amdump -o DUMPTYPE:no-compress:compress="server fast"
amdump -o HOLDINGDISK:hd1:use="-100 mb"
amdump -o INTERFACE:local:use="2000 kbps"
When overriding device properties, one must carefully quote the command
line to simulate the syntax of real configuration files. The following
example should serve as a guide:
amdump -o ´device-property="PROPERTY_MAX_VOLUME_USAGE" "100000"´
Note that configuration overrides are not effective for tape changers,
which supply a tapedev based on their own configuration. In order to
override tapedev, you must also disable any changer:
amdump -otapedev=/dev/nst1 -otpchanger=´´
James da Silva <firstname.lastname@example.org>
Stefan G. Weichinger <email@example.com>