From SME Server
Jump to navigationJump to search

PythonIcon.png Skill level: Advanced
The instructions on this page may require deviations from standard procedures. A good understanding of linux and Koozali SME Server is recommended.

Is this article helpful to you?
Please consider donating or volunteering
Thank you!


Maintainers(s) Affa3: Arnaud, stephdl (and please see above note.)
Copyright (C) 2004-2012 by Michael Weinberger


Alpha 10:
Contrib 9:
The latest version of smeserver-affa is available in the SME repository, click on the version number(s) for more information.


The main purpose of this affa package is to make a SME Server a dedicated backup box in a few minutes. Affa backs up as many SME servers as you like or any other servers which have sshd running and rsync installed. Once it is configured, Affa runs reliably unattended and sends warning messages in case of any errors.

All backup archives are full backups, as Affa makes use of the hardlink technique. Therefore a new full backup only needs disk space for the differences plus the filesystem overhead for the hardlinks and directories (which is typically 2-3%).

Affa is based on the rsync program and supports the rsync --compress option. This allows you to run backups over the internet or VPN. A typical setup is one or more Affa backup servers placed in different locations, which backup the production server(s) over the VPN.

A special feature is the rise option, which allows you to rise the backup server to your production server from a backup archive in case of a dead loss of your production server. The rise is executed within a extremely short time, even with huge amount of data. The rise feature uses hardlinks and therefore does not use up additional disk space.

The rise feature can also be used to upgrade from a lower version to a higher version of SME. This also work from SME 8.1 to 9. For this to work you need 2 separate installations of SME. One is the actual running server and the other is an affa backup server with a newer version of SME Server where you execute the rise command. See more here: Moving SME to new Hardware

A simple way to have a server running with maximum up time is to duplicate the hardware. E.g. to have 2 separate hardware boxes or virtual servers on separate hardware. One is the actual server and the other one is a backup system that in a very short time in case of problems can be upgraded to the actual server with the rise command. This also allow you to upgrade the SME Server software with minimum down time!

Affa is a command line tool for system administrators and is intentionally designed without a GUI. Therefore it can be efficiently managed on the console and over slow internet connections.

Note: This document also refers to the Affa Version 2 which is not maintained anymore. Information on Affa 3 will be gradually added here once it has been more tested.

Affa features at a glance

  • Affa is secure: All transfers and connections are made by using the ssh protocol with public/private key authentication
  • Periodically runs unattended full backups. Only differences to the previous full backup are transferred over the network
  • Using rsync with optional bandwidth limit and compression allows backups over the internet
  • Uses hardlink technique, i.e. physical disk space only needed for the differences between two full backups
  • Keeps a configurable number of scheduled, daily, weekly, monthly and yearly full backup archives
  • The archives are browseable. No need to unpack them first.
  • Archives can be mapped to a Samba share.
  • Backup jobs are started by the cron daemon
  • Interrupted jobs continuing with already transfered data when restarted
  • Backups the default e-smith directories and files, when property SMEServer is set to yes
  • Additional directories and files can be included
  • Directories and files can be excluded from the backup
  • Non-SME server linux systems can be backed up by setting the SMEServer property to no and using an include list
  • In ESXi mode, running virtual machines can be backed up. See Backup of ESXi Virtual Machines using Affa
  • Configurable nice level for rsync processes on the backup and source server
  • Optional run of custom programs before and after a job run (e.g. running tape backup)
  • Checks the disk space left after a job run with warning levels strict, normal or risky
  • Extensive checking of failure conditions
  • Sends failure messages to a configurable list of email addresses
  • Sends a warning message, if the backup server runs out of disk space
  • Installs an optional watchdog on the source server in case the backupserver fails (SME Server only)
  • Watchdog sends warning, if an expected backup did not run (SME Server only)
  • Watchdog sends a daily reminder message, if the error continues unchecked (SME Server only)
  • Option to display current status of all jobs showing times of last and next run, size and disk usage
  • Status can be mailed on a daily, weekly or monthly schedule
  • Option to display all existing archives of a job shown date, number of files, size and bytes tranferred from the source
  • Option to send (and revoke) the public key to the source server (SME Server and ESXi only)
  • Option to rise the backup server to a production server from a backup (SME Server only)
  • The rise feature does not physically move data and therefore is extremly fast and needs (almost) no extra disk space
  • Rise option can be run remotely as the NIC driver configuration of the backup server are preserved
  • Compares installed RPMs on source with backup server. Sends warning message, if not in sync
  • Undo rise option to restore the backup server
  • Configurable via a e-smith style db, with one record for each job and a default record for all jobs
  • Logs to /var/log/affa/JOB.log and /var/log/affa/affa.log with optional debug switch for higher verbosity
  • Log files are rotated weekly, with 5 logs kept
Warning.png Warning:
SME v7/8/9 are no longer supported.

Installation of Affa 3

The following instructions assume that a fresh Affa 3 installation is made on a fresh dedicated SME server which serves as a dedicated backup server ('affabox‘). No jobs are migrated from Affa 2. For using an existing SME server, just skip the instructions how to setup a fresh SME box.

Setup a dedicated machine with SME 8.1 or SME 9.0 from CDROM. Use the following example settings:

  1. Domain name: athome.xx (use your existing domain name)
  2. Host name: affabox (must not match to existing host)
  3. IP address: (must not match to existing IP address)
  4. Subnet:
  5. role: Server only
  6. Gateway: (use your existing gateway)
  7. DHCP: DEactivate
  8. DNS server: (use your existing DNS server)
  9. activate changes: yes
  10. The fresh server will then restart with the above settings.

To update your server, login as user root on local console. Use 'top' command to display running jobs. Wait until 'yum' stopped running.

yum clean all
yum update

The fresh server will be updated.

signal-event post-upgrade
signal-event reboot

The server will reboot. Afterwards it is ready for installation of Affa 3.

Login as root on local or remote SSH console of affabox.

For SME10

There is a new development version 3.3.1

Warning.png Warning:
This will almost certainly break both your old backups and your new server. ONLY use this in a test scenario until it is fully released

Warning.png Warning:
Do not use ANY Koozali SME v9 affa versions with v10. It will probably break both your backups and your new server

Note that we KNOW that this is Alpha code under heavy development (though nearing beta at the time of this update) and almost certainly will not work correctly and still needs a lot of testing and fixes. See the Bugs below.

We have done very minimal testing on migration from v9 to v10.


Under normal circumstances you would execute this to install :

yum install smeserver-extrarepositories-epel smeserver-extrarepositories-openfusion
signal-event yum-modify
config set UnsavedChanges no
yum install --enablerepo=smecontribs,epel,openfusion smeserver-affa

Please post bug reports in the bug tracker

The server again needs to be updated.

signal-event post-upgrade
signal-event reboot

The server will reboot. Afterwards Affa 3 is ready for use.

Now manually create directory for archive files (to be fixed):

mkdir /var/affa

Warning.png Warning:
Note that on SME9, a new SSH AutoBlock feature interferes with the Affa requirement for regular SSH logins! You might have to disable this feature to allow Affa to operate correctly.

Creation of backup jobs

Job configurations of Affa 3 are no longer stored in an e-smith style database. They are stored in configuration files located in /etc/affa . Create your configuration files in this directory, e.g.

touch /etc/affa/backup-smeserver.conf

where backup-smeserver is your jobname, and add content as described below.

Quick start example

You have a SME production server with hostname 'smeserver‘ and IP
You have a second SME box as your backup server with hostname 'affabox‘ and IP

Login to your 'affabox' as root and edit /etc/affa/backup-smeserver.conf . Using e.g. editor nano, create the following example job configuration file for jobname backup-smeserver :

Description=Backup of smeserver.athome.xx

If you have a Letsencrypt certificate on the production server, then you should also include


Then save your job configuration file.

Now check that your configuration is OK:

affa --configcheck

This should throw out no errors. Now create / send key files to your productive smeserver:

affa --send-key backup-smeserver

The following output should appear on the console:

Job sme-backup: Generating RSA keys...
Successfully created RSA key pair.
root@'s password: <Enter password of and hit ENTER>
Public key sent to

Now run your job manually, both for test purposes, as well as to create RSA keys:

affa --run backup-smeserver

The following output should appear on the console:

The authenticity of host 'backup-smeserver (' can't be established.
RSA key fingerprint is 3b..........65.
Are you sure you want ton continue connecting (yes/no)? <yes>

Now your backup job should run for the first time. Depending on the volume of the files, this takes quite a while.

Once the job is done, check that the archive is available:

affa –-list-archives

or run the job a second time:

affa --run backup-smeserver

Note that you won't be asked for the password again. Note as well, that this second run of the job backup-smeserver should run considerably faster, because not all files are copied again: just the difference to the previous run is backed up, the rest is stored as hard links to the existing file copies.

Check that the second archive is available too:

affa --list-archives

From here you are able to work as with Affa 2. Modify your config file as required and described below. Automate the backup using the command affa --make-cronjobs, see below.

Job configuration properties

Note 1: The default values shown in this table are the Affa program defaults and not to be confused with the preset values in the job configuration files, e.g. backup-smeserver.conf.

Note 2: The complete documentation of the settings of Affa3 is available in [1] . It should be taken in consideration!

Note 3: Affa 3 for SME is a fork of Affa 3 for CentOS. It adds a few options back again which have been removed. Generally the configuration properties as described here do apply. Arnaud added the following functions for SME:

  • The parameter “SMEServer”
  • The parameter and function “Watchdog”
  • The parameter and function “RPMCheck”
  • The functions “--rise” and “--undo-rise”

Consequently the list below should reproduce the list for the unforked Affa 3 version, plus adding the properties above. For details refer to Arnaud's website.

Property Value Multivalue Default Description
remoteHostName FQHN or IP no <none> FQHN or IP of the source host (mandatory)
remoteUser account no root The user account to be used with all ssh logins. May be case sensitive, e.g. Administrator account on Windows
Description text string no <none> Any text that describes the job
TimeSchedule HHMM yes 2230 The time the job will run. Use exactly 4 digits, no colon, no point. Important: Using the proper format HHMM is essential. Badly formatted TimeSchedule will cause strange Perl errors. Multiple times can be achieved by repeating the line, each line with a different time.
status enabled or disabled no enabled When set to disabled, no cron entries will made. You can still run a job manually.
Include full path yes <none> File(s) or directory(s) to be included in the backup.
Exclude full path yes <none> File(s) or directory(s) to be excluded from the backup.
RootDir full path no /var/affa where to store the backup archives. Do not use /home/e-smith or /root as these are included in the backup and therefore the rise option will not work! Recommended: /var/affa
scheduledKeep integer >= 1 no 1 how many of the scheduled backups to be kept
dailyKeep integer >= 0 no 7 how many of the daily backups to be kept
weeklyKeep integer >= 0 no 4 how many of the weekly backups to be kept
monthlyKeep integer >= 0 no 12 how many of the monthly backups to be kept
yearlyKeep integer >= 0 no 2 how many of the yearly backups to be kept
EmailAddress yes root comma separated list of mail addresses, where the messages should be sent to
Note: By default Affa only sends messages on errors, never on success (see property chattyOnSuccess).
RetryAttempts integer >= 0 no 4 When set to a value>0, Affa re-run a failed job RetryAttempts times with a delay of RetryAfter seconds.
RetryAfter seconds >= 0 no 900 when set to a value>0, wait RetryAfter seconds before re-running the job after an error. Only applicable with RetryAttempts>0
RetryNotification yes or no no no when set to no, Affa does not send an error message when a job has failed and RetryAttempts is configured. An error message is only sent when the last attempt has failed.
NRPEtrigger hours no 24 NRPE reports a job as critical when the last successful run is older then NRPEtrigger hours. To exclude the job from monitoring set to a value < 0.
SambaShare yes or no no no Access to the job archives via CIFS protocol.
SambaValidUser local account yes affa User who has permission to access the job archives via the CIFS protocol.
path relative to /etc/affa/scripts/ yes <none> Scripts to be executed before a job run. The job name and type (scheduled, daily etc.) are passed as arguments to the program. The preJobCommandRemote scripts are copied to the remote server and then executed there. The scripts are executed in alphabetical order. Use a numerical prefix if you need a specific order e.g. 01scriptB, 02remotescriptA, 03scriptA. Use the included and as a starting point for own scripts.
Note: If you use the parameter “SMEServer=yes”, you may delete the “Included” relative to SME default directories in the jobconfig.ini: they will be backuped automatically.
path relative to /etc/affa/scripts/ yes <none> Scripts to be executed after a job run. The job name, the type (scheduled, daily etc.) and the exit code of the job run are passed as arguments to the program. The postJobCommandRemote scripts are copied to the remote server and then executed there. The scripts are executed in alphabetical order. Use a numerical prefix if you need a specific order e.g. 01scriptB, 02remotescriptA, 03scriptA. Use the included and as a starting point for own scripts.
Note: If you use the parameter “SMEServer=yes”, you may delete the “Included” relative to SME default directories in the jobconfig.ini: they will be backuped automatically.
dedup yes or no no no The purpose of the deduplication is to remove duplicate files to save backup space. When set to 'yes' file deduplication is run after the synchronization has been completed. It looks for files that have identical content, user, group and permissions and replace duplicates by hardlinks. Deduplication scans the just completed archive and the previous one, that usually is

scheduled.0 and daily.0 or scheduled.0 and scheduled.1. Consider this scenario: A user has renamed directories or files. Rsync sees those as new ones and copies them. Deduplication finds the identical copies in the previous archive and replace them by hardlinks. To use deduplication the Freedup program needs to be installed. Affa actually runs freedup -upg scheduled.0 <previous_archive>.

dedupKill yes or no no no When set to 'no' the job ignores affa --kill or affa --killall when deduplication is running. This is useful in the context of killAt which is typically used to stop bandwith utilisation.
DiskSpaceWarn strict or normal or risky or none no strict Checks disk space remainin on the backup device and issue a warning via email.
sshPort service port no 22 When sshd on the remote host listens on a non-standard port, set the port here.
ConnectionCheckTimeout seconds no 120 Affa checks the ssh connection before the rsync process is started and exits with an error after the configured time if the host did not respond.
BandwidthLimit integer>=0 kilobytes per second no 0 Limits the data transfer rate. A value of zero specifies no limit.
rsyncTimeout seconds no 900 Rsync exits after the configured time if no data was transferred. This avoids infinitely hanging in case of a network error.
rsyncCompress yes or no no yes Compress the transferred data. May be useful with slow internet connections. Increases CPU load on remote and backup host.
rsync--inplace yes or no no yes Set to no if the rsync version on the remote hist does not support this option.
rsync--modify-window integer >= 0 no 0 When comparing two timestamps, rsync treats the timestamps as being equal if they differ by no more than the modify-window value. This is normally 0 for an exact match. A value >= 0 is useful if you can't get the clocks of the remote host and the Affa server in sync.
rsyncOptions string no <none> Additional option string to be passed to rsync
localNice -19...+19 no 0 run rsync local process niced.
remoteNice -19...+19 no 0 run rsync process on source niced.
killAt HHMM no <none> The time at which a job will be killed if it was still running. You can use it for example to kill jobs that are running over the internet early in the morning so that your users have the full bandwidth available during office hours.
resumeKilledAt HHMM no <none> The time at which a killed job will be resumed. This allows you to start earlier in the evening than the scheduled time.
chattyOnSuccess integer >= 0 no 0 When set to a value>0, Affa sends a message on a successfully completed job run and decrements the chattyOnSuccess value. When the value has reached zero, Affa falls back to the default and only sends messages on errors.
full path no <none> Device and mountpoint of backup device (e.g. USB disk). Device is automounted before a job starts and unmounted after job completion. With both properties empty no automount is done.
AutomountOptions string no <none> An option string passed to the mount command.
AutoUnmount yes or no no yes When set to 'no' the automounted device stay mounted after the Affa run.
Debug yes or no no no Set to yes to increase log verbosity.
remoteRsyncBinary full path no /usr/bin/rsync If the rsync program on the remote server is located in non-standard location, set it here
remoteNiceBinary full path no /bin/nice If the nice program on the remote server is located in non-standard location, set it here.
localRsyncBinary full path no /usr/bin/rsync If the local rsync program is located in non-standard location, set it here
localNiceBinary full path no /bin/nice If the nice local program is located in non-standard location, set it here.
RemoteAuthorizedKeysFile path relative to remote user's home or full path no .ssh/authorized_keys2 If the remote host stores the authorized keys file in a non-standard location, set it here.
rsyncdMode yes or no no no Set to yes to connect to the rsync daemon on the remote host (instead of running rsync over ssh).
rsyncdModule string no AFFA The rsyncd module name (only applicable with rsyncdMode=yes).
rsyncdUser string no affa The username for authentication to the rsync daemon (only applicable with rsyncdMode=yes).
rsyncdPassword string no <none> The password for authentication to the rsync daemon (only applicable with rsyncdMode=yes).
globalStatus enabled or disabled or jobs no jobs Set to enabled or disabled to overide the status settings in all job sections. When set to value jobs, the status settings in the job sections are effictive.
Note: This property is allowed only in the [GlobalAffaConfig] section.
sendStatus daily or weekly or monthly or never no weekly Defines how frequently the status will be sent.
Note: This property is allowed only in the [GlobalAffaConfig] section.
SMEServer yes or no no no When set to yes, the default e-smith directories are automatically included and the property RPMCheck=yes can be used.
Note: this property is not implemented in Affa 3 for CentOS. It has been specifically added to the fork for SME.
RPMCheck yes or no no no Only applicable to jobs that backup a SME server. Compares the packages installation of the source host with this affa backup host. Sends a message with diff list if not in sync. This check is useful, if you want have the option to rise the backup server to a production server from a backup.
Note: this property is not implemented in Affa 3 for CentOS. It has been specifically added to the fork for SME.
Watchdog yes or no no yes Only applicable to jobs that backups a SME server. When a job is started, affa installs a watchdog script on the source in /etc/cron.d/, which sends a warning message, if the next scheduled job (taken from the TimeSchedule property + 10 minutes) did not run. This guarantees, that you will be notfied even in case of a affa server outage. The watchdog script send a daily reminder message, if the error continues. The next run job replaces the watchdog script with a new trigger time.
Note: this property is not implemented in Affa 3 for CentOS. It has been specifically added to the fork for SME.

Default configuration properties

For all 'name=value' properties defaults can be set in the [GlobalAffaConfig] section which are used in all job configuration when the corresponding property is omitted. For example, when these properties are set in [GlobalAffaConfig] section, they can be omitted in the specific job configurations. This is useful, when you set up many similar jobs.

Example: You want to set the property 'localNice' to 19 for all jobs. Then add the following section in one of your configuration files:


Consequently you don't need to set this property for your individual jobs. Properties set in the job record override the defaults.

Properties for global Settings

The following special properties are only applicable to the [GlobalAffaConfig] section:

  • sendStatus
  • globalStatus

All jobs can be disabled for execution through 'cronjobs" with setting 'globalStatus' to 'disabled'.


To re-enable run, either delete the above configuration line, or set to:


which is the default value.

You can also set the value to


which enforces the the job execution through 'cronjobs' and overrides the specified values in the section for the specific job.

Usage and command line options

Important.png Note:
Options can be abbreviated to uniqueness, e.g. --mak is equal to --make-cronjobs

affa --run JOB

Starts a job run. Usually done by the cronjob.

affa --make-cronjobs

Configures the cronjobs as scheduled in the jobs records. Run this command to make changes of time related properties effective i.e. TimeSchedule, killAt and resumeKilledAt properties. By default this command will by executed by 'cronjobs' every 15 minutes. So you don't need to run this command manually, you can also just wait max. 15 minutes before your updates job configurations become effective.

affa --configcheck

Checks the syntax and values in all configuration files found in /etc/affa/. Run this command after modifying the configuration. Lower/Upper case errors in property names are corrected automatically.

affa --send-key [JOB JOB ...]

This first generates the RSA key on the Affa Server, if not already done. Then the public key is send to the hosts 'remoteHostName' as defined in section of each job JOB and generates the job specific ssh known host entry.

Important.png Note:
When initially doing this step, you will need to temporarily enable "Allow secure shell access using standard passwords" on the source server.

affa --check-connections [JOB JOB ...]

Checks the ssh login and if applicable rsyncd auth for all jobs given as arguments. Without any arguments all jobs are checked.

affa --full-restore [--preserve-newer=no] [--delete=yes] JOB [ARCHIVE]

Does a full restore remote server of all backed up files and directories from the backup ARCHIVE. If ARCHIVE is not given, the archive 'scheduled.0' is used as the default. The full restore reconstructs the server as it was at the time of the backup. After the restore the source host reboots.

With option --preserve-newer=yes files on the remote server with modification time newer than on the backup are not overwritten.

With option --delete=yes all files on the remote server, which are not in the backup, are deleted.

If the parameter “SMEServer=yes” is set, the signal-event pre-restore and signal-event post-upgrade will be used automatically.

affa --list-archives JOB

Displays a table of all present archives of job JOB with date, number of files, size and and bytes received. While column buTime shows the actual rsync time, the column ddTime shows the length of the subsequent deduplication run. ddYld is the gained deduplication yield in bytes.

Affa version 3.1.0-0 on
| Job: fshare-ak                                                               |
| Description: Fileserver AK                                                   |
| Directory: /var/affa/fshare-ak/                                              |
| Hostname:                                                       |
| Email:                                                 |
| Run | Completion date      | buTime | ddTime | ddYld | Files | Size  | Recvd |
| Y 0 | Wed 2010-06-23 20:26 | 11m53s |      - |     - |  412k |  143G |  470M |
| M11 | Sun 2010-08-29 20:22 |  7m50s |      - |     - |  417k |  153G |   14M |
| M10 | Sun 2010-09-26 20:23 | 8m57s  |      - |     - |  430k |  156G |   14M |
| M 9 | Sun 2010-10-31 20:25 | 10m05s |      - |     - |  448k |  161G |   15M |
| M 8 | Sun 2010-11-28 20:31 | 16m42s |      - |     - |  463k |  167G |   17M |
| M 7 | Sun 2011-01-02 20:48 | 33m07s |      - |     - |  486k |  173G |  1.1G |
| M 6 | Sun 2011-01-30 20:33 | 18m54s |      - |     - |  493k |  176G |   24M |
| M 5 | Sun 2011-02-27 20:28 | 13m31s |      - |     - |  490k |  176G |   19M |
| M 4 | Sun 2011-03-27 20:28 | 13m08s |      - |     - |  491k |  175G |   17M |
| M 3 | Sun 2011-05-01 20:30 | 15m41s |      - |     - |  493k |  179G |   18M |
| M 2 | Sun 2011-05-29 20:28 | 13m40s |      - |     - |  494k |  182G |   31M |
| M 1 | Sun 2011-06-19 20:26 | 11m20s |      - |     - |  493k |  183G |   17M |
| M 0 | Sun 2011-07-03 20:28 | 13m46s |      - |     - |  496k |  183G |   18M |
| W 3 | Sun 2011-07-10 20:28 | 13m38s |      - |     - |  479k |  180G |   16M |
| W 2 | Mon 2011-07-11 20:35 | 20m27s |      - |     - |  479k |  180G |  301M |
| W 1 | Fri 2011-07-15 20:30 | 15m53s |      - |     - |  480k |  180G |   62M |
| W 0 | Sun 2011-07-24 19:39 |  9m25s |      - |     - |  482k |  181G |   16M |
| D 6 | Thu 2011-07-28 19:42 | 12m22s |      - |     - |  483k |  182G |  176M |
| D 5 | Fri 2011-07-29 23:29 | 11m10s |  5h33m |   45G |  483k |  182G |   16M |
| D 4 | Sat 2011-07-30 19:53 | 23m26s |  2h30m |  8.3G |  483k |  182G |   17M |
| D 3 | Sun 2011-07-31 20:07 | 37m31s |  4m47s |  8.3G |  483k |  182G |   17M |
| D 2 | Mon 2011-08-01 20:44 |  1h14m |  7h50m |  8.5G |  484k |  182G |  630M |
| D 1 | Tue 2011-08-02 20:02 | 32m28s | 12h20m |  8.3G |  484k |  182G |   74M |
| D 0 | Wed 2011-08-03 19:58 | 28m46s | 11h01m |  8.5G |  484k |  182G |  214M |
| S 0 | Thu 2011-08-04 20:00 | 30m28s | 11h52m |  8.5G |  484k |  182G |  203M |
affa --list-archives [--csv] JOB

With --csv, the output is in machine readable colon separated format.

affa --status [--csv]

Displays a table of all configured jobs with enable status, time of last and next run, size of the most recent archive, exectution time of the last run and the number of scheduled (S), daily (D), weekly (W), monthly (M) and yearly (Y) archives. If last time shows 'failed', the job did not run in the last 24h. For disabled jobs 'Last' always shows 'failed' after 24 h. To see the date and time of the last run of those jobs use the --list-archives option. Column 'Next' shows the time when the next run will be started, if 'Enabled' is 'yes'. Column 'ddYld' shows the deduplication yield in bytes.

Affa version 3.1.0-0 on
| Job          | ENA |  Last |   Time |  Next |  Size | ddYld | N of S,D,W,M,Y |
| ads-ak       | yes | 20:20 |  0m29s | 20:20 |   28M |     - |  1, 7, 2, 0, 0 |
| ak-user-ma.. | yes | 22:31 |  1m51s | 22:30 |  170M |     - |  3, 7, 4, 6, 0 |
| azubi-1      | yes | 03:39 |  3h24m | 00:15 |   44G |  1.2G |  1, 7, 4,10, 1 |
| dcpant       | yes | 03:06 |  1m03s | 03:05 |  1.3G |     - |  1, 7, 4,10, 1 |
| eshare       | yes | 21:06 | 21m01s | 20:45 |   24G |  178M |  1, 7, 4,10, 1 |
| etherpad     | yes | 20:40 |  0m13s | 20:40 |  3.1M |     - |  1, 7, 4, 2, 0 |
| fazubi       | yes | 08:16 | 11h16m | 21:00 |  132G |  3.3G |  1, 7, 4,10, 1 |
| fschare-rh   | yes | de-duplicating (pid 9719)              |  1, 7, 4,10, 1 |
| fsh-02       | yes | 07:53 | 12h23m | 19:30 |  182G |  8.5G |  1, 7, 4,10, 1 |
| fshare-ak2   | yes | 00:30 |  0m26s | 00:30 |  415M |  544k |  1, 7, 4,10, 1 |
| helpdesk     | yes | 21:27 |  2m16s | 21:25 |  895M |  138k |  1, 7, 4,10, 0 |
| it-share     | yes | running rsync (pid 9744)               |  1, 7, 1, 0, 0 |
| lightroom    | yes | waiting (pid 9528)                     |  1, 7, 4, 2, 0 |
| localhost    | yes | 02:15 |  0m08s | 02:15 |  395k |     - |  1, 7, 2, 0, 0 |
| mediawiki    | yes | 16:10 |  0m06s | 19:40 |  1.1G |  6.7M |  3, 7, 4,10, 1 |
| mshare       | yes | 00:48 |  1h33m | 23:15 |   18G |  2.4G |  1, 7, 4,10, 1 |
| wshare       | yes | 00:34 |  1h49m | 22:45 |   23G |  484M |  1, 7, 4,10, 1 |
2 disabled jobs not listed. Use --all to display.

With --csv, the output is in machine readable colon separated format.

affa --show-config-pathes [--csv] [JOB JOB ...]

Prints the full pathes of the file where the section of job JOB is defined.

affa --show-default-config

Prints a list of all allowed property names with their default values. These values are used, when omitted in the job sections.

affa --show-schedule [--all]

Prints a 'graphical' timetable for all enabled jobs. The resolution is 30 minutes. An 'S' character marks the scheduled start times. The duration of the job runs are marked with '=' characters.

Affa version 3.1.0-0 on
              TIME 12:00    16:00    20:00    0:00     4:00     8:00
            fsh-02 -------- -------S ~~~~~~~~ ~~~~~~~~ ~~~~~~~~ --------
            ads-ak -------- -------- S------- -------- -------- --------
          etherpad -------- -------- -S------ -------- -------- --------
            eshare -------- -------- -S~----- -------- -------- --------
            fazubi -------- -------- --S~~~~~ ~~~~~~~~ ~~~~~~~~ ~-------
          helpdesk -------- -------- --S----- -------- -------- --------
          it-share -------- -------- ---S=~~~ ~~~~~~~~ ~~~~~~~- -------- busy
            wshare -------- -------- -----S~~ ~~------ -------- --------
ak-user-management -------- -------- -----S-- -------- -------- --------
        fschare-rh -------- -------- -----S=- -------- --~----- -------- busy
            mshare -------- -------- ------S= ~~------ -------- --------
           azubi-1 -------- -------- -------- S=~~~~~~ -------- --------
        fshare-ak2 -------- -------- -------- -S------ -------- --------
         mediawiki -------- -------- -------- ---S---- -------- ----S---
         localhost -------- -------- -------- ----S--- -------- --------
            dcpant -------- -------- -------- ------S- -------- --------
         lightroom -------- -------- -------- -------- -------- S------- busy
Symbols: S=scheduled K=kill R=resume '='=rsync '~'=dedup
2 disabled jobs not listed. Use --all to display.
affa --log-tail [JOB]

Displays the tail of the logfile of job JOB with live update. This command is identical to tail -n 50 -f /path/to/logfile. Without the JOB argument the global logfile is shown.

affa --send-status

Sends the status table, the disk-usage and the archive list of all jobs to the email addresses configured in the [GlobalAffaConfig] section.

affa --disk-usage

Shows the current disk usage of all root dir filesystems

Affa version 3.1.0-0 on
| Use% |   Used |  Avail | Root Dir                                           |
|  71% |  938GB |  365GB | /var/affa                                          |
|  20% |  194GB |  759GB | /mnt/affadev                                       |

With --csv, the output is printed in a machine readable colon separated format.

affa --cleanup JOB

After you have lowered a keep value, e.g. scheduledKeep, then archives with a higher indices will no longer be shifted and will exist for ever. This option finds these archives and deletes them after confirmation. When running the cleanup command on a locally attached USB drive, manually mount the USB drive BEFORE running this command. When finished, manually unmount the USB drive. This requirement may apply to some of the other commands listed here.

affa --rename-job JOB NEWNAME

Renames the job JOB to NEWNAME including the section name and archive directories.

affa --move-archive JOB NEWROOTDIR

Moves the archive directory of job JOB to the rootdir NEWROOTDIR and adjusts the value of property RootDir. NEWROOTDIR must be a full path starting with a slash. As moving across filesystems (e.g. from an external USB drive to the local disk) is not possible, Affa uses a copy command in this case and deletes the source directory after that. Depending on the archive size, copying across filesystems can take a long time.

affa --delete-job [--revoke-key] JOB

Irreversibly deletes all archives, of job JOB and set the status property to disable. With --revoke-key option, the public key on the remote server will be deleted.

affa --revoke-key JOB

Deletes the public RSA key on the remote server.

affa --kill JOB

Terminates the running job JOB and all its child processes (rsync processes).

affa --killall

Terminates all running jobs.

affa --mailtest JOB

Sends a test email to the email addresses configured in the JOB section. Use this to verify, that your mail system is working.

Important.png Note:
By default Affa only sends messages on errors, never on success (see property chattyOnSuccess).

affa --nrpe [JOB JOB ...]

Checks for jobs that were not run the last NRPRtrigger hours and reports them as failed. A Nagios/ICINGA NRPE complient message is printed and exit status returned.

affa --version

Displays the Affa version number and checks Samba and Freedup installation.

affa --warranty

Displays the disclaimer of liability.

affa --license

Displays the license Affa is released under.

affa --help

Displays a short help.

affa --debug

Enables verbose logging. Overrides job and global configurations.

SME Specific usage and command line options

These commands are implemented in a fork of Affa 3 project for use with SME. They are not part of the Affa 3 for CentOS project.

SME server 'rise' function

Important.png Note:
Before trying a rise please check /var/affa/YourServer/rpms-missing.txt

You may find there are missing rpms/contribs that template fragments depends on.

It may be necessary to install some of these rpms before a rise to prevent issues.

The SME server version of Affa has a unique feature restore feature enabling you to turn your Affa backup server into a replica of your server. It is an extremely fast way to restore your server.

affa --rise [--all] JOB [ARCHIVE]

Runs a full restore on the Affa server (!!!) of all standard files and directories from the backup ARCHIVE of job JOB. In other words: After completion, the Affa box reboots as a clone of the source server. Ensure, that the source server has been powered off before you reboot the Affa box, otherwise the network interface will not come up. This is important, when you run --rise remotely. The --rise feature only works with SME servers and should only be used on dedicated backup servers.

With option --all, all files and directories of the archive as defined by the include[] properties are restored. Files or directories with the same name on the Affa server will be overwritten and cannot be restored by a --undo-rise. This should not be an issue on a dedicated Affa server which does not hold any other data. After a possible --undo-rise those additional restored data must be removed manually.

Please note, that the rise process backs up the the Affa server itself before doing the restore from the archive. This backup is used by a possible --undo-rise run to restore the Affa server. Only the standard files and directories are backed up. Data in non-standard locations (like /opt) are untouched and will still exist after the rise run if they don't get overwritten by data of the backup (=if the backup contains /opt too e.g.).

Important.png Note:
The above command is not implemented in Affa 3 for CentOS. It has been specifically added to the fork for SME.

affa --undo-rise

This feature reverts a risen Affa box to a backup server. After a reboot, all configured jobs based on standard files and directories will work again.

Important.png Note:
The above command is not implemented in Affa 3 for CentOS. It has been specifically added to the fork for SME.

Example setups

Dedicated backup server

  • Setup a dedicated server and install Affa 3 for SME.
  • Setup a job for every server you want to backup.
  • Send the public keys to every server.
affa --send-key JOBNAME
  • Check whether password-less logins are working.
affa --check-connections
  • Check whether the scheduled jobs are evenly distributed over the day.
affa --show-schedule
  • Create the cron jobs.
affa --make-cronjobs
  • Check the status after 24 hours.
affa --status

Backup of single ibays

Suppose you want to backup the ibays 'staff1' and 'staff2' on your production server with WAN IP to an Affa server in a different location over the internet every night at 2:30am.

  • Log into the Affa server.
  • Create a dedicated job file /etc/affa/ibay-staff.conf . Edit it and set
Description=Backup of ibay-staff on
  • Save the configuration
  • send the public key to the production server
affa --send-key ibay-staff
  • check next morning
affa --list-archives ibay-staff
affa --status
ls /var/affa/ibay-staff

Two production servers backup each other

You have two sites connnected via a VPN and a SME Server running on each site. In that case you don't need a dedicated Affa backup server. Both production servers can additionally act as Affa backup servers backing up the server of the other site. Simply install Affa and configure a job that backs up the other one. You can use all Affa features except of the rise feature.

When using the rise feature, the server becomes any of the backed up systems, which is less useful in this scenario as it would give you a running copy of the server of the other site while the server of this site is down.

To get redundancy and a faster restore you can configure a local backup to an external USB or NAS device.

Please pay attention, that you do not backup the archives back to the other site. Set the RootDir property to a path which is not included in the SME default backup list. When leaving the RootDir property to its default '/var/affa', this is guaranteed.

Use Affa to backup to a NFS-mounted NAS or a local attached USB drive

Important.png Note:
This chapter still needs to be checked whether it works with Affa 3 for SME.

You want to backup your SME production server with hostname 'prodbox‘ and IP on a mounted filesystem instead of setting up a dedicated Affa box.

Setup NAS

You have a FreeNAS box with IP up and running with NFS service enabled. The disk is mounted to /mnt/affashare. You have authorized the IP address of your prodbox server to access share /mnt/affashare.

  • log into the 'prodbox' and install the NFS packages
/usr/bin/yum install --enablerepo=smecontribs smeserver-nfs
  • now enable and start the portmapper service
config setprop portmap status enabled
service portmap start
  • mount the NFS share
mkdir -p /mnt/affadevice
mount /mnt/affadevice

Alternatively setup a USB drive
  • log into the 'prodbox'
  • connect a USB hard disk to the USB Bus. Now you must determine what device the kernel has assigned to the drive. View the /var/log/message and search for Initializing USB Mass Storage driver. A few lines below you'll find the name of the device. In this example it is sdh. Replace /dev/sdh by your device in following instructions.
  • use the fdisk program to create a linux partition. Verify that this is really the attached USB drive before you continue!
fdisk /dev/sdh

You'll most likely find an existing vfat dos partition, which you have to delete first. In the following we assume, that you have created a single partition /dev/sdh1.

  • now format the drive with an ext3 filesystem
mkfs.ext3 /dev/sdh1
  • make the mount point
mkdir -p /mnt/affadevice
  • add the following line to the /etc/fstab
/dev/sdh1 /mnt/affadevice ext3 defaults
  • mount the drive
mount /mnt/affadevice
  • crosscheck your work using the df command
Setup Affa

You want to run backups at 11:30 h, 15:30 h and 19:30 h and you want to keep the last 3 scheduled backups, 7 daily, 5 weekly, 12 monthly and 1 yearly backups.

  • log into the 'prodbox' and install the Affa packages as described above.
  • create/edit the conf file of the job: nano /etc/affa/prodbox.conf

and set:


Review the other properties and change them to your needs.

  • run the job manually
affa --run prodbox

With this kind of setup you cannot use the affa rise feature, as it requires the backup archive to be located on the same fileystem as the server installation. The rise option uses hardlinks, which are not working across filesystems.


Having the backup archives in the same filesystem is always a risk, which can be minimized by using the automount feature. Then the external filesystem is only mounted during a job run.

In the NAS example set


and skip the step 2.

In the USB drive example set


and skip the steps 5 to 8.

The mount point will be automatically created, if it does not exist.
To access the archive directory, you need to mount it manually.

Copying a AFFA USB hard drive archive to a new disk

Affa uses copious amounts of hard links to compress and preserve disk space for its backups. If you are in the situation where you want to copy such a disk archive to a new (bigger) disk, you need to ensure that the hard links are copied correctly or the destination copy may became significantly bigger than the total of the source archive.

One way to copy across file systems (i.e. two different USB disks) and preserve the hard links is as follows:

  • mount both USB drives but with different mount points. e.g. /media/backup1 & /media/backup2 and then:
mkdir /media/backup2/archive 
cd /media/backup1/archive
tar cpf - . | ( cd /media/backup2/archive && tar xpf - )

where archive is the name of the AFFA job you want to move to the other disk.


Restore single files or directories

Example 1: It's Tuesday January 6th 2009, when user 'briedlin' asks you to restore the messages of his mailbox 'orders' he has accidentally deleted on Monday.

  • You first must check what backup archives are available. The jobname of this server backup is 'primmail'. To get a listing of all archives run
affa --list-archives primmail

(see the example listing in chapter Affa#Usage_and_command_line_options

  • Choose the daily.0 archive, which was created Monday night. Now restore the mailbox 'orders' using the rsync command.
  • Now run the rsync command (note the trailing slash!) on the Affa backup server:
export RDIR=/home/e-smith/files/users/briedlin/Maildir/.orders/  # this variable is used to shorten the next command line
rsync -av /var/affa/primmail/daily.0/$RDIR$RDIR

If the servers are configured to use a different ssh port eg 2222, then instead do:

export RDIR=/home/e-smith/files/users/briedlin/Maildir/.orders/  # this variable is used to shorten the next command line
rsync -av -e 'ssh -i /root/.ssh/id_rsa_affa -p 2222' /var/affa/primmail/daily.0/$RDIR$RDIR

Example 2: A user has deleted the file orderform.pdf from ibay 'docs' on the server and asks you to restore it.

  • You have searched and found the latest version of this file in weekly archive of job 'prodserv'.
  • To copy it back to the server run on the Affa server
export RFILE=/home/e-smith/files/ibays/docs/files/orderform.pdf  # this variable is used to shorten the next command line
rsync -av /var/affa/prodserv/weekly.1/$RFILE$RFILE
  • If the servers are configured to use a different ssh port eg 2222, then instead do:
export RFILE=/home/e-smith/files/ibays/docs/files/orderform.pdf  # this variable is used to shorten the next command line
rsync -av -e 'ssh -i /root/.ssh/id_rsa_affa -p 2222' /var/affa/prodserv/weekly.1/$RFILE$RFILE

Example 3: Restoring a file with special characters in the path

Do not use shell variables to shorten the command. It would complicate things more than it would help. Quote the source and destination path. In the destination path escape blank characters with a double backslash and brackets with a single backslash. On the Affa backup server do:

rsync -av "/var/affa/fileshare/daily.4/home/e-smith/files/ibays/mechfiles/files/Valve Control (Design Gr)/VALVE LIST FOR ISSUED.xls" "\\ Control\\ \(Design\\ Gr\)/VALVE\\ LIST\\ FOR\\ ISSUED.xls"

If the servers are configured to use a different ssh port eg 2222, then instead do:

rsync -av -e 'ssh -i /root/.ssh/id_rsa_affa -p 2222' "/var/affa/fileshare/daily.4/home/e-smith/files/ibays/mechfiles/files/Valve Control (Design Gr)/VALVE LIST FOR ISSUED.xls" "\\ Control\\ \(Design\\ Gr\)/VALVE\\ LIST\\ FOR\\ ISSUED.xls"

Full restore

Important.png Note:
Please for notes about rpms-missing.txt above:


affa --full-restore [optional settings] JOB [ARCHIVE]

This rsyncs the data (files and directories )from the backup ARCHIVE back to the 'remoteHostname' defined in the configuration of the job JOB. If ARCHIVE is not given, the archive 'scheduled.0' is used as the default. The --full-restore reconstructs the server as it was at the time of the backup and takes following IMPORTANT optional settings in consideration:

With option [--preserve-newer=yes] files on the remote server with modification time newer than on the backup are not overwritten.

With option [--delete=yes] all files on the remote server, which are not in the backup, are deleted.

If the parameter “SMEServer=yes” is set, the signal-event pre-restore and signal-event post-upgrade will be used automatically.

After the restore is done, the restored server reboots automatically.

Example: You have backuped your production server 'prodsrv' as job 'prodbox'. To restore only lost data from the latest backup run

affa --full-restore prodbox

To restore exactly from the older archive daily.3 run

affa --full-restore [--preserve-newer=no] [--delete=yes]  prodbox daily.3

Warning.png Warning:
A full restore with --preserve-newer no and --delete yes reconstructs the server as it was at the time of the backup. That means, that all files created or server configuration changes made after the backup will be lost!

Moving a SME server installation to new hardware using the Affa rise feature or the backup-restore functions

Please see this Howto: Moving SME to new Hardware

Restore from USB drive on new server

This tip comes from

Q) I have complete backups using affa stored on a usb hard drive connected to our affa backup server. I need to restore an earlier monthly backup of our job "mailbackup" to a test server rather than back to the original system. If I did it from the backup server I see the instructions of how to rise that server to the current backup on that server but I want to restore a point about a month ago before some strange things happened. And I want to do it on a machine that is not our backup server or our production server. I tried to figure out how but am lost in the options. My goal is to do some testing.

A) On your testserver setup a job "mailbackup" identical to that on your backup server (a copy of the conf file is available into the folder of the archive, as an hidden file .mailbackup-setup.ini) but set property RootDir to /var/affa and property status to disabled. Connect the USB drive and copy the archive of the job mailbackup to /var/affa. Then run affa --rise mailbackup ARCHIVE


What files and directories are included by default?

With SMEServer=no nothing at all.

With SMEServer=yes the SME default backup list is the output of following command:

perl -e 'use esmith::Backup;$b=new esmith::Backup;print join("\n",$b->restore_list)."\n"' 

and as additional folder:


Important.png Note:
The following example needs to be reworked for Affa 3.

Can I exclusively backup image files from a specific directory? Yes. Assuming you want to backup all gif and jpg files from directory /home/e-smith/files/ibays/pictures/files use this configuration

db affa setprop JOBNAME SMEServer no           # do not include SME Server default directories and files
db affa setprop JOBNAME Include[0] '/home/e-smith/files/ibays/pictures/files' # start searching here
db affa setprop JOBNAME Include[1] '*/'        # search the tree recursively
db affa setprop JOBNAME Include[2] '*.gif'     # copy files that match these patterns
db affa setprop JOBNAME Include[3] '*.jpg'  
db affa setprop JOBNAME Exclude[0] '*'         # exclude all others
db affa setprop JOBNAME rsyncOptions '-m'      # do not create empty folders

How can I move a backup archive from my production server in a remote branch office to the Affa server using an USB disk? I want to avoid to download the first full backup over the internet. Install Affa on your production server, connect and mount the USB disk. Setup a job to write the archive to the USB disk. After the job run, connect the USB disk to your Affa server and setup the job in the same way. Use the setup script from the archive directory. You only need to change the remoteHostName property. Now use the --move-archive option to move the archive to the local disk. If not already done, send the public key to your production server. You're done.

How do I backup two SME servers behind a firewall? First you need to configure port forwardings for the ssh service on your firewall. Use a non-standard port, e.g 2200->22, for your second server. Setup a job on your Affa server for each of your production servers. Set up the job property sshPort=2200 for second server.


This removes the installed Affa package, all configuration data and all backup archives.

Run the following commands for each job:

affa --revoke-key JOBNAME
affa --delete-job JOBNAME

Verify that all jobs have been deleted:

affa --status

Remove the Affa package:

yum remove smeserver-affa-

Cleaning up:

rm -f /etc/cron.d/affa
rm -rf /etc/affa
rm -rf /var/affa /var/log/affa


Affa stops working after prodserver was rebuild.


  • Affa stopped working or is broken.
  • When affaserver tries to make an ssh connection to prodserver it fails.
  • In the /var/log/affa/jobname.log file you will this message: "SSH connection to prodserver ip failed. Did you send the public key".
  • Resending the keys does not solve the problem.


  • This is probably due to the key for prodserver that is stored on the affaserver no longer matches the key of prodserver.
  • You need remove the existing key so that the next ssh connection re-gets the keys from prodserver.
  • From the command line edit the /root/.ssh/known_hosts file and remove the line for prodserver. It starts with: prodserver ip ssh-rsa

Bug report

Affa is listed in the bugtracker contribs section. Please report all bugs, new feature requests and documentation issues there.

IDProductVersionStatusSummary (14 tasks)
11845SME Contribs10.0RESOLVEDRestoring a file - Rsync can't connect via ssh to production server
11784SME Contribs10.0UNCONFIRMEDConf files placed in /etc/affa cannot contain spaces and/or sub folders with spaces
11503SME Contribs10betaVERIFIEDaffa --rise gives message 'atalk' is not a valid service name
11502SME Contribs10betaVERIFIEDaffa --version reports error samba service not installed
11501SME Contribs10betaVERIFIEDaffa --license file not found
11500SME Contribs10betaVERIFIEDaffa --warranty file not found
11024SME Contribs10alphaIN_PROGRESSFirst Import to SME10 tree [smeserver-affa]
10783SME Contribs10.0VERIFIEDaffa generates but does not use a dsa ssh key (this keytype is obsolete).
10762SME Contribs9.2UNCONFIRMEDUse of uninitialized value $s in substitution
10467SME Contribs10.0CONFIRMEDNFR : before a rise check if installed rpm on original server are also present on new server
10373SME Contribs10.0CONFIRMEDMissing newline at end of config prevents conjob
10299SME Contribs10.0CONFIRMEDqpsmtpd user missing from /etc/shadow
9551SME Contribs9.1UNCONFIRMEDAffa installation fails on starting job on one 9.1 server, works on another
9186SME Contribs10.0CONFIRMEDAFFA hardlinks using NFS storage for archives

Additional information

For details of performance, changelog etc see Affa:Additional information