FS BACKUP TYPE
POSTGRES BACKUP TYPE
MYSQL BACKUP TYPE
LDAP BACKUP TYPE
beam.conf − configuration file for BEAM
The beam.conf file defines what data are to be included in the backup, what backup methods to use and where to store the resulting archives. The backup(1) and restore(1) utilities expect to find it in the /usr/local/etc directory.
The file is essentially a shell script that defines variables controlling backup and restore procedures. Names of the variables used by the backup scripts begin with backup_.
Sets file mode creation mask. This mask will apply to all files created during the backup: archives, snapshots and temporary files. The default mask is 077.
Sets the directory where temporary files will be stored. The default is /tmp.
If set, redirect standard error and standard output to that file. The file will be created if it does not exist. Otherwise, the output will be appended to it.
By default, this file will be automatically rotated. See openlog_hook below for a discussion of this.
Configuration text for logrotate(8), used to rotate the logfile. The default corresponds to:
backup_logrotate_conf = "
Note, that no block statements are allowed within backup_logrotate_conf, because its entire contents will be placed within an appropriate block statement.
Use this command to access remote archives. Unless set, the backup utilities will first look for an executable file named /usr/local/bin/speedssh. If not found, they will fall back to using /usr/bin/ssh. It is the responsibility of the administrator to provide for authentication on the remote system, e.g. by using shared-key authentication.
Sets the interval, in weeks, during which old backups and snapshots are retained. Empty value or zero means never delete the backups.
archives in a s3 bucket you will need the
Additional options for s3backer(1).
Use it, e.g., to pass the −−vhost option if your bucket is located outside of the US region.
Sets the name of the s3 bucket to keep the archives in.
Defines mount point for the "raw" bucket.
Default value is /mnt/s3backer.
Defines mount point for the actual s3-backed filesystem.
Default value is /mnt/s3.
variables control additional options and arguments passed to
Use this variable to pass additional options to tar. Do not place tar operation switches (such as -c, -t, etc.) here! These will be added automatically by appropriate scripts, depending on the operation being performed.
By default this variable is empty (no additional options passed to tar).
This variable supplies suffix for created archive names. The default is tar . The dot is inserted between the file name and its suffix.
This variable is useful if you request compression and want archive file names to reflect it. For example:
Directory where archive files are located. It’s OK to specify a remote directory here, e.g.
This variable must be defined and whatever directory it points to must already exist.
If it is not defined and backup_bucket_name is set, indicating that the S3 storage is used for backups, backup_archive_dir will be set automatically to the same value as backup_mp_s3.
Directory for incremental snapshot files. These files are be named after their archive counterparts, by appending the .db suffix.
Set this variable to any non-empty value to request additional verbosity. The effect of this setting depends on backup methods being used. For example, the fs method adds a −v option to each invocation of tar(1).
item identifies information that needs to be backed up.
Syntactically, items follow the same rules as shell variable
names. Backup items are introduced using the following
A whitespace-separated list of backup items.
For each backup item name the configuration file must provide a set of variables describing where to obtain this information and how to store it to the archive. These variables follow the same naming convention:
where name is the item name and var is a valid shell identifier.
The name_type variable must always be defined. It identifies the backup method to be applied to that item. The following backup methods are implemented:
A file system backup.
Backup of PostgreSQL databases.
Backup of MySQL databases.
Backup of LDAP databases.
This is the
basic backup type. It creates incremental backups of a
specified file system. For each item of this type, the
following variables must be defined:
The absolute path to the file system being backed up.
A whitespace-separated list of directories and/or files in the item_dir directory.
variables are optional:
A whitespace-separated list of file names (or glob(3) patterns) to exclude from archivation. This is translated to a list of −−exclude options to tar(1). Make sure to quote globbing patterns as you would have done in shell, to prevent them from being expanded too early. For example:
A list of files to read exclusion patterns from. It is translated into a list of −−exclude−from options.
A list of additional options to pass to tar. Use this with caution.
This configuration requests incremental backups of everything found in the directories /etc and /var/spool.
postgres type creates a dump of a PostgreSQL database
and archives it using tar(1). The following variables
must be defined for items of this type:
The database name.
mysql backup type creates a dump of a MySQL database
and archives it with tar(1). The database to dump and
access credentials are specified using the following
The database name. If this variable is empty, all databases will be dumped.
A full pathname of the MySQL defaults file which contains credentials for accessing this database. This file must have at least the mysqldump and mysql sections, the former being used when dumping the database and the latter when restoring it.
A LDAP database
is introduced with the ldap backup type. The
following keywords specify what to backup:
Specifies the pathname of a directory that contains the database files. This is used at restore time to clean up old database contents prior to restoring it.
Number of the database to dump, if your LDAP configuration contains several databases. If absent, all databases will be dumped. This keyword cannot be used in conjunction with item_database_suffix (see below).
Use the specified suffix to determine which database to operate upon.
Use dn, scope and filter from the supplied argument URI to handle only matching entries.
Restore the database with the privileges of this user.
A list of additional options for slapcat(8C). Use with caution.
A list of additional options to pass to slapadd(8C) when restoring from the backup.
Apart from usual logging, backup reports can be sent via email to selected recipients. To enable this feature, the variable backup_notify_email must be defined to a comma-separated list of recipient email addresses.
A number of
variables is provided to customize the email headers and
Sets a list of emails to receive backup report. Multiple emails must be separated with commas. If not set, no notification will be sent.
Defines the sender email address. Backup reports will appear to be sent from this address. The default value is root@hostname, when hostname is the name of the host on which the backup is run.
If set, this variable must contain a single email address, without personal part or comments, e.g.
Defines personal part of the sender email. The personal part will be enclosed in double quotes and prepended to the value of backup_sender_email to form a valid RFC-2822 From header. For example, if you have:
backup_sender_personal="Automatic backup report"
then the resulting report will contain:
From: "Automatic backup report" <firstname.lastname@example.org>
Defines additional headers for the report message. By default, the following headers are generated: From, To, Subject, X−Beam−Items, X−Beam−Round, and X−Beam−Level. The three X− headers contain the backed up items (as set in the backup_items variable), backup round and backup level numbers, correspondingly.
To supply multiple headers, delimit them with single newline characters.
Sets the introductory text to be displayed before the actual report.
Sets the signature text, which will be output after the report body.
Defines the mailer program and its arguments. The default is
/usr/sbin/sendmail −oi −t −F $backup_sender_email
variables, called hooks, allow you to supply arbitrary
commands to be run at particular stages of the backup or
Commands listed in this variable are run before opening the logfile. The most common use for this hook is to rotate an oversized logfile prior to opening it.
The default configuration file initializes this hook with beam_logrotate, a shell function that uses logrotate(8) to manage the log file. Configuration for logrotate(8) can be supplied in the backup_logrotate_conf variable. By default, the log file is rotated weekly and 4 weeks worth of backlogs are preserved.
This variable contains a whitespace-separated list of commands to be invoked before starting backup (or restore). These commands are started in order of their appearance in the list and are called without arguments. By default, this variable is empty.
This variable contains a whitespace-separated list of commands to be invoked after the backup (or restore) has finished. These commands are started in order of their appearance in the list and are called without arguments. By default, this variable is empty.
A simple configuration file for backing up the contents of /etc, /home, and the PostreSQL database savane:
# Request bzip2
# Modify file name suffix accordingly
# Write data to /dev/rmt1 on a remote machine:
# Keep snapshots in this directory:
# Two backup
items are defined:
dbdump item dumps the database:
system item backs up the file system:
# Notify root
about the results.
backup_sender_personal="Automatic daily backup"
backup_report_intro="Today’s daily backup produced the following
Beam Automatic Backup"
Report bugs to <email@example.com>.
Manpage server at Ulysses.gnu.org.ua.
Powered by mansrv 1.0