Ubuntu Manpage: Requirements

Provided by: backuppc_3.3.0-1ubuntu1.1_amd64

Installing BackupPC

Requirements
BackupPC requires:

• A linux, solaris, or unix based server with a substantial amount of free disk space
(see the next section for what that means). The CPU and disk performance on this
server will determine how many simultaneous backups you can run. You should be able to
run 4-8 simultaneous backups on a moderately configured server.

Several users have reported significantly better performance using reiserfs compared
to ext3 for the BackupPC data file system. It is also recommended you consider either
an LVM or RAID setup (either in HW or SW; eg: 3Ware RAID10 or RAID5) so that you can
expand the file system as necessary.

When BackupPC starts with an empty pool, all the backup data will be written to the
pool on disk. After more backups are done, a higher percentage of incoming files will
already be in the pool. BackupPC is able to avoid writing to disk new files that are
already in the pool. So over time disk writes will reduce significantly (by perhaps a
factor of 20 or more), since eventually 95% or more of incoming backup files are
typically in the pool. Disk reads from the pool are still needed to do file compares
to verify files are an exact match. So, with a mature pool, if a relatively fast
client generates data at say 1MB/sec, and you run 4 simultaneous backups, there will
be an average server disk load of about 4MB/sec reads and 0.2MB/sec writes (assuming
95% of the incoming files are in the pool). These rates will be perhaps 40% lower if
compression is on.

• Perl version 5.8.0 or later. If you don't have perl, please see
<http://www.cpan.org>.

• Perl modules Compress::Zlib, Archive::Zip and File::RsyncP. Try "perldoc
Compress::Zlib" and "perldoc Archive::Zip" to see if you have these modules. If not,
fetch them from <http://www.cpan.org> and see the instructions below for how to build
and install them.

The File::RsyncP module is available from <http://perlrsync.sourceforge.net> or CPAN.
You'll need to install the File::RsyncP module if you want to use Rsync as a transport
method.

• If you are using smb to backup WinXX machines you need smbclient and nmblookup from
the samba package. You will also need nmblookup if you are backing up linux/unix DHCP
machines. See <http://www.samba.org>. Samba versions 3.x are stable and now
recommended instead of 2.x.

See <http://www.samba.org> for source and binaries. It's pretty easy to fetch and
compile samba, and just grab smbclient and nmblookup, without doing the installation.
Alternatively, <http://www.samba.org> has binary distributions for most platforms.

• If you are using tar to backup linux/unix machines, those machines should have version
1.13.7 at a minimum, with version 1.13.20 or higher recommended. Use "tar --version"
to check your version. Various GNU mirrors have the newest versions of tar; see
<http://www.gnu.org/software/tar/>.

• If you are using rsync to backup linux/unix machines you should have version 2.6.3 or
higher on each client machine. See <http://rsync.samba.org>. Use "rsync --version" to
check your version.

For BackupPC to use Rsync you will also need to install the perl File::RsyncP module,
which is available from <http://perlrsync.sourceforge.net>. Version 0.68 or later is
required.

• The Apache web server, see <http://www.apache.org>, preferably built with mod_perl
support.

What type of storage space do I need?
BackupPC uses hardlinks to pool files common to different backups. Therefore BackupPC's
data store (__TOPDIR__) must point to a single file system that supports hardlinks. You
cannot split this file system with multiple mount points or using symbolic links to point
a sub-directory to a different file system (it is ok to use a single symbolic link at the
top-level directory (__TOPDIR__) to point the entire data store somewhere else). You can
of course use any kind of RAID system or logical volume manager that combines the capacity
of multiple disks into a single, larger, file system. Such approaches have the advantage
that the file system can be expanded without having to copy it.

Any standard linux or unix file system supports hardlinks. NFS mounted file systems work
too (provided the underlying file system supports hardlinks). But windows based FAT and
NTFS file systems will not work.

Starting with BackupPC 3.1.0, run-time checks are done at startup and at the start of each
backup to ensure that the file system can support hardlinks, since this is a common area
of configuration problems.

How much disk space do I need?
Here's one real example for an environment that is backing up 65 laptops with compression
off. Each full backup averages 3.2GB. Each incremental backup averages about 0.2GB.
Storing one full backup and two incremental backups per laptop is around 240GB of raw
data. But because of the pooling of identical files, only 87GB is used. This is without
compression.

Another example, with compression on: backing up 95 laptops, where each backup averages
3.6GB and each incremental averages about 0.3GB. Keeping three weekly full backups, and
six incrementals is around 1200GB of raw data. Because of pooling and compression, only
150GB is needed.

Here's a rule of thumb. Add up the disk usage of all the machines you want to backup
(210GB in the first example above). This is a rough minimum space estimate that should
allow a couple of full backups and at least half a dozen incremental backups per machine.
If compression is on you can reduce the storage requirements by maybe 30-40%. Add some
margin in case you add more machines or decide to keep more old backups.

Your actual mileage will depend upon the types of clients, operating systems and
applications you have. The more uniform the clients and applications the bigger the
benefit from pooling common files.

For example, the Eudora email tool stores each mail folder in a separate file, and
attachments are extracted as separate files. So in the sadly common case of a large
attachment emailed to many recipients, Eudora will extract the attachment into a new file.
When these machines are backed up, only one copy of the file will be stored on the server,
even though the file appears in many different full or incremental backups. In this sense
Eudora is a "friendly" application from the point of view of backup storage requirements.

An example at the other end of the spectrum is Outlook. Everything (email bodies,
attachments, calendar, contact lists) is stored in a single file, which often becomes
huge. Any change to this file requires a separate copy of the file to be saved during
backup. Outlook is even more troublesome, since it keeps this file locked all the time, so
it cannot be read by smbclient whenever Outlook is running. See the Limitations section
for more discussion of this problem.

In addition to total disk space, you should make sure you have plenty of inodes on your
BackupPC data partition. Some users have reported running out of inodes on their BackupPC
data partition. So even if you have plenty of disk space, BackupPC will report failures
when the inodes are exhausted. This is a particular problem with ext2/ext3 file systems
that have a fixed number of inodes when the file system is built. Use "df -i" to see your
inode usage.

Step 1: Getting BackupPC
Some linux distributions now include BackupPC. The Debian distribution, supported by
Ludovic Drolez, can be found at <http://packages.debian.org/backuppc> and is included in
the current stable Debian release. On Debian, BackupPC can be installed with the command:

apt-get install backuppc

In the future there might be packages for Gentoo and other linux flavors. If the packaged
version is older than the released version then you may want to install the latest version
as described below.

Otherwise, manually fetching and installing BackupPC is easy. Start by downloading the
latest version from <http://backuppc.sourceforge.net>. Hit the "Code" button, then select
the "backuppc" or "backuppc-beta" package and download the latest version.

Step 2: Installing the distribution
Note: most information in this step is only relevant if you build and install BackupPC
yourself. If you use a package provided by a distribution, the package management system
should take of installing any needed dependencies.

First off, there are five perl modules you should install. These are all optional, but
highly recommended:

Compress::Zlib
To enable compression, you will need to install Compress::Zlib from
<http://www.cpan.org>. You can run "perldoc Compress::Zlib" to see if this module is
installed.

Archive::Zip
To support restore via Zip archives you will need to install Archive::Zip, also from
<http://www.cpan.org>. You can run "perldoc Archive::Zip" to see if this module is
installed.

XML::RSS
To support the RSS feature you will need to install XML::RSS, also from
<http://www.cpan.org>. There is not need to install this module if you don't plan on
using RSS. You can run "perldoc XML::RSS" to see if this module is installed.

File::RsyncP
To use rsync and rsyncd with BackupPC you will need to install File::RsyncP. You can
run "perldoc File::RsyncP" to see if this module is installed. File::RsyncP is
available from <http://perlrsync.sourceforge.net>. Version 0.68 or later is required.

File::Listing, Net::FTP, Net::FTP::RetrHandle, Net::FTP::AutoReconnect
To use ftp with BackupPC you will need four libraries, but actually need to install
only File::Listing from <http://www.cpan.org>. You can run "perldoc File::Listing" to
see if this module is installed. Net::FTP is a standard module. Net::FTP::RetrHandle
and Net::FTP::AutoReconnect included in BackupPC distribution.

To build and install these packages you should use the cpan program. Alternatively, you
can fetch the tar.gz file from <http://www.cpan.org> and then run these commands:

tar zxvf Archive-Zip-1.26.tar.gz
cd Archive-Zip-1.26
perl Makefile.PL
make
make test
make install

The same sequence of commands can be used for each module.

Now let's move onto BackupPC itself. After fetching BackupPC-3.3.0.tar.gz, run these
commands as root:

tar zxf BackupPC-3.3.0.tar.gz
cd BackupPC-3.3.0
perl configure.pl

In the future this release might also have patches available on the SourceForge site.
These patch files are text files, with a name of the form

BackupPC-3.3.0plN.diff

where N is the patch level, eg: pl2 is patch-level 2. These patch files are cumulative:
you only need apply the last patch file, not all the earlier patch files. If a patch file
is available, eg: BackupPC-3.3.0pl2.diff, you should apply the patch after extracting the
tar file:

# fetch BackupPC-3.3.0.tar.gz
# fetch BackupPC-3.3.0pl2.diff
tar zxf BackupPC-3.3.0.tar.gz
cd BackupPC-3.3.0
patch -p0 < ../BackupPC-3.3.0pl2.diff
perl configure.pl

A patch file includes comments that describe that bug fixes and changes. Feel free to
review it before you apply the patch.

The configure.pl script also accepts command-line options if you wish to run it in a non-
interactive manner. It has self-contained documentation for all the command-line options,
which you can read with perldoc:

perldoc configure.pl

Starting with BackupPC 3.0.0, the configure.pl script by default complies with the file
system hierarchy (FHS) conventions. The major difference compared to earlier versions is
that by default configuration files will be stored in /etc/BackupPC rather than below the
data directory, __TOPDIR__/conf, and the log files will be stored in /var/log/BackupPC
rather than below the data directory, __TOPDIR__/log.

Note that distributions may choose to use different locations for BackupPC files than
these defaults.

If you are upgrading from an earlier version the configure.pl script will keep the
configuration files and log files in their original location.

When you run configure.pl you will be prompted for the full paths of various executables,
and you will be prompted for the following information.

BackupPC User
It is best if BackupPC runs as a special user, eg backuppc, that has limited
privileges. It is preferred that backuppc belongs to a system administrator group so
that sys admin members can browse BackupPC files, edit the configuration files and so
on. Although configurable, the default settings leave group read permission on pool
files, so make sure the BackupPC user's group is chosen restrictively.

On this installation, this is __BACKUPPCUSER__.

For security purposes you might choose to configure the BackupPC user with the shell
set to /bin/false. Since you might need to run some BackupPC programs as the BackupPC
user for testing purposes, you can use the -s option to su to explicitly run a shell,
eg:

su -s /bin/bash __BACKUPPCUSER__

Depending upon your configuration you might also need the -l option.

Data Directory
You need to decide where to put the data directory, below which all the BackupPC data
is stored. This needs to be a big file system.

On this installation, this is __TOPDIR__.

Install Directory
You should decide where the BackupPC scripts, libraries and documentation should be
installed, eg: /usr/local/BackupPC.

On this installation, this is __INSTALLDIR__.

CGI bin Directory
You should decide where the BackupPC CGI script resides. This will usually be below
Apache's cgi-bin directory.

It is also possible to use a different directory and use Apache's ``<Directory>''
directive to specifiy that location. See the Apache HTTP Server documentation for
additional information.

On this installation, this is __CGIDIR__.

Apache image Directory
A directory where BackupPC's images are stored so that Apache can serve them. You
should ensure this directory is readable by Apache and create a symlink to this
directory from the BackupPC CGI bin Directory.

Config and Log Directories
In this installation the configuration and log directories are located in the
following locations:

__CONFDIR__/config.pl main config file
__CONFDIR__/hosts hosts file
__CONFDIR__/pc/HOST.pl per-pc config file
__LOGDIR__/BackupPC log files, pid, status

The configure.pl script doesn't prompt for these locations but they can be set for new
installations using command-line options.

Step 3: Setting up config.pl
After running configure.pl, browse through the config file, __CONFDIR__/config.pl, and
make sure all the default settings are correct. In particular, you will need to decide
whether to use smb, tar,or rsync or ftp transport (or whether to set it on a per-PC basis)
and set the relevant parameters for that transport method. See the section "Step 5: Client
Setup" for more details.

Step 4: Setting up the hosts file
The file __CONFDIR__/hosts contains the list of clients to backup. BackupPC reads this
file in three cases:

• Upon startup.

• When BackupPC is sent a HUP (-1) signal. Assuming you installed the init.d script,
you can also do this with "/etc/init.d/backuppc reload".

• When the modification time of the hosts file changes. BackupPC checks the
modification time once during each regular wakeup.

Whenever you change the hosts file (to add or remove a host) you can either do a kill -HUP
BackupPC_pid or simply wait until the next regular wakeup period.

Each line in the hosts file contains three fields, separated by white space:

Host name
This is typically the host name or NetBios name of the client machine and should be in
lower case. The host name can contain spaces (escape with a backslash), but it is not
recommended.

Please read the section "How BackupPC Finds Hosts".

In certain cases you might want several distinct clients to refer to the same physical
machine. For example, you might have a database you want to backup, and you want to
bracket the backup of the database with shutdown/restart using $Conf{DumpPreUserCmd}
and $Conf{DumpPostUserCmd}. But you also want to backup the rest of the machine while
the database is still running. In the case you can specify two different clients in
the host file, using any mnemonic name (eg: myhost_mysql and myhost), and use
$Conf{ClientNameAlias} in myhost_mysql's config.pl to specify the real host name of
the machine.

DHCP flag
Starting with v2.0.0 the way hosts are discovered has changed and now in most cases
you should specify 0 for the DHCP flag, even if the host has a dynamically assigned IP
address. Please read the section "How BackupPC Finds Hosts" to understand whether you
need to set the DHCP flag.

You only need to set DHCP to 1 if your client machine doesn't respond to the NetBios
multicast request:

nmblookup myHost

but does respond to a request directed to its IP address:

nmblookup -A W.X.Y.Z

If you do set DHCP to 1 on any client you will need to specify the range of DHCP
addresses to search is specified in $Conf{DHCPAddressRanges}.

Note also that the $Conf{ClientNameAlias} feature does not work for clients with DHCP
set to 1.

User name
This should be the unix login/email name of the user who "owns" or uses this machine.
This is the user who will be sent email about this machine, and this user will have
permission to stop/start/browse/restore backups for this host. Leave this blank if no
specific person should receive email or be allowed to stop/start/browse/restore
backups for this host. Administrators will still have full permissions.

More users
Additional user names, separate by commas and with no white space, can be specified.
These users will also have full permission in the CGI interface to
stop/start/browse/restore backups for this host. These users will not be sent email
about this host.

The first non-comment line of the hosts file is special: it contains the names of the
columns and should not be edited.

Here's a simple example of a hosts file:

host dhcp user moreUsers
farside 0 craig jim,dave
larson 1 gary andy

Step 5: Client Setup
Four methods for getting backup data from a client are supported: smb, tar, rsync and ftp.
Smb or rsync are the preferred methods for WinXX clients and rsync or tar are the
preferred methods for linux/unix/MacOSX clients.

The transfer method is set using the $Conf{XferMethod} configuration setting. If you have
a mixed environment (ie: you will use smb for some clients and tar for others), you will
need to pick the most common choice for $Conf{XferMethod} for the main config.pl file, and
then override it in the per-PC config file for those hosts that will use the other method.
(Or you could run two completely separate instances of BackupPC, with different data
directories, one for WinXX and the other for linux/unix, but then common files between the
different machine types will duplicated.)

Here are some brief client setup notes:

WinXX
One setup for WinXX clients is to set $Conf{XferMethod} to "smb". Actually, rsyncd is
the better method for WinXX if you are prepared to run rsync/cygwin on your WinXX
client.

If you want to use rsyncd for WinXX clients you can find a pre-packaged zip file on
<http://backuppc.sourceforge.net>. The package is called cygwin-rsync. It contains
rsync.exe, template setup files and the minimal set of cygwin libraries for everything
to run. The README file contains instructions for running rsync as a service, so it
starts automatically everytime you boot your machine. If you use rsync to backup
WinXX machines, be sure to set $Conf{ClientCharset} correctly (eg: 'cp1252') so that
the WinXX file name encoding is correctly converted to utf8.

Otherwise, to use SMB, you can either create shares for the data you want to backup or
your can use the existing C$ share. To create a new share, open "My Computer", right
click on the drive (eg: C), and select "Sharing..." (or select "Properties" and select
the "Sharing" tab). In this dialog box you can enable sharing, select the share name
and permissions.

All Windows NT based OS (NT, 2000, XP Pro), are configured by default to share the
entire C drive as C$. This is a special share used for various administration
functions, one of which is to grant access to backup operators. All you need to do is
create a new domain user, specifically for backup. Then add the new backup user to the
built in "Backup Operators" group. You now have backup capability for any directory on
any computer in the domain in one easy step. This avoids using administrator accounts
and only grants permission to do exactly what you want for the given user, i.e.:
backup. Also, for additional security, you may wish to deny the ability for this user
to logon to computers in the default domain policy.

The relevant configuration settings are $Conf{SmbShareName}, $Conf{SmbShareUserName},
$Conf{SmbSharePasswd}, $Conf{SmbClientPath}, $Conf{SmbClientFullCmd},
$Conf{SmbClientIncrCmd} and $Conf{SmbClientRestoreCmd}.

BackupPC needs to know the smb share user name and password for a client machine that
uses smb. The user name is specified in $Conf{SmbShareUserName}. There are four ways
to tell BackupPC the smb share password:

• As an environment variable BPC_SMB_PASSWD set before BackupPC starts. If you
start BackupPC manually the BPC_SMB_PASSWD variable must be set manually first.
For backward compatibility for v1.5.0 and prior, the environment variable PASSWD
can be used if BPC_SMB_PASSWD is not set. Warning: on some systems it is possible
to see environment variables of running processes.

• Alternatively the BPC_SMB_PASSWD setting can be included in /etc/init.d/backuppc,
in which case you must make sure this file is not world (other) readable.

• As a configuration variable $Conf{SmbSharePasswd} in __CONFDIR__/config.pl. If
you put the password here you must make sure this file is not world (other)
readable.

• As a configuration variable $Conf{SmbSharePasswd} in the per-PC configuration file
(__CONFDIR__/pc/$host.pl or __TOPDIR__/pc/$host/config.pl in non-FHS versions of
BackupPC). You will have to use this option if the smb share password is
different for each host. If you put the password here you must make sure this file
is not world (other) readable.

Placement and protection of the smb share password is a possible security risk, so
please double-check the file and directory permissions. In a future version there
might be support for encryption of this password, but a private key will still have to
be stored in a protected place. Suggestions are welcome.

As an alternative to setting $Conf{XferMethod} to "smb" (using smbclient) for WinXX
clients, you can use an smb network filesystem (eg: ksmbfs or similar) on your
linux/unix server to mount the share, and then set $Conf{XferMethod} to "tar" (use tar
on the network mounted file system).

Also, to make sure that file names with special characters are correctly transferred
by smbclient you should make sure that the smb.conf file has (for samba 3.x):

[global]
unix charset = UTF8

UTF8 is the default setting, so if the parameter is missing then it is ok. With this
setting $Conf{ClientCharset} should be emtpy, since smbclient has already converted
the file names to utf8.

Linux/Unix
The preferred setup for linux/unix clients is to set $Conf{XferMethod} to "rsync",
"rsyncd" or "tar".

You can use either rsync, smb, or tar for linux/unix machines. Smb requires that the
Samba server (smbd) be run to provide the shares. Since the smb protocol can't
represent special files like symbolic links and fifos, tar and rsync are the better
transport methods for linux/unix machines. (In fact, by default samba makes symbolic
links look like the file or directory that they point to, so you could get an infinite
loop if a symbolic link points to the current or parent directory. If you really need
to use Samba shares for linux/unix backups you should turn off the "follow symlinks"
samba config setting. See the smb.conf manual page.)

The requirements for each Xfer Method are:

tar You must have GNU tar on the client machine. Use "tar --version" or "gtar
--version" to verify. The version should be at least 1.13.7, and 1.13.20 or
greater is recommended. Tar is run on the client machine via rsh or ssh.

The relevant configuration settings are $Conf{TarClientPath}, $Conf{TarShareName},
$Conf{TarClientCmd}, $Conf{TarFullArgs}, $Conf{TarIncrArgs}, and
$Conf{TarClientRestoreCmd}.

rsync
You should have at least rsync 2.6.3, and the latest version is recommended.
Rsync is run on the remote client via rsh or ssh.

The relevant configuration settings are $Conf{RsyncClientPath},
$Conf{RsyncClientCmd}, $Conf{RsyncClientRestoreCmd}, $Conf{RsyncShareName},
$Conf{RsyncArgs}, and $Conf{RsyncRestoreArgs}.

rsyncd
You should have at least rsync 2.6.3, and the latest version is recommended. In
this case the rsync daemon should be running on the client machine and BackupPC
connects directly to it.

The relevant configuration settings are $Conf{RsyncdClientPort},
$Conf{RsyncdUserName}, $Conf{RsyncdPasswd}, $Conf{RsyncdAuthRequired},
$Conf{RsyncShareName}, $Conf{RsyncArgs}, $Conf{RsyncArgsExtra}, and
$Conf{RsyncRestoreArgs}. $Conf{RsyncShareName} is the name of an rsync module (ie:
the thing in square brackets in rsyncd's conf file -- see rsyncd.conf), not a file
system path.

Be aware that rsyncd will remove the leading '/' from path names in symbolic links
if you specify "use chroot = no" in the rsynd.conf file. See the rsyncd.conf
manual page for more information.

ftp You need to be running an ftp server on the client machine. The relevant
configuration settings are $Conf{FtpShareName}, $Conf{FtpUserName},
$Conf{FtpPasswd}, $Conf{FtpBlockSize}, $Conf{FtpPort}, $Conf{FtpTimeout}, and
$Conf{FtpFollowSymlinks}.

You need to set $Conf{ClientCharset} to the client's charset so that file names are
correctly converted to utf8. Use "locale charmap" on the client to see its charset.

For linux/unix machines you should not backup "/proc". This directory contains a
variety of files that look like regular files but they are special files that don't
need to be backed up (eg: /proc/kcore is a regular file that contains physical
memory). See $Conf{BackupFilesExclude}. It is safe to back up /dev since it contains
mostly character-special and block-special files, which are correctly handed by
BackupPC (eg: backing up /dev/hda5 just saves the block-special file information, not
the contents of the disk).

Alternatively, rather than backup all the file systems as a single share ("/"), it is
easier to restore a single file system if you backup each file system separately. To
do this you should list each file system mount point in $Conf{TarShareName} or
$Conf{RsyncShareName}, and add the --one-file-system option to $Conf{TarClientCmd} or
$Conf{RsyncArgs}. In this case there is no need to exclude /proc explicitly since it
looks like a different file system.

Next you should decide whether to run tar over ssh, rsh or nfs. Ssh is the preferred
method. Rsh is not secure and therefore not recommended. Nfs will work, but you need
to make sure that the BackupPC user (running on the server) has sufficient permissions
to read all the files below the nfs mount.

Ssh allows BackupPC to run as a privileged user on the client (eg: root), since it
needs sufficient permissions to read all the backup files. Ssh is setup so that
BackupPC on the server (an otherwise low privileged user) can ssh as root on the
client, without being prompted for a password. There are two common versions of ssh:
v1 and v2. Here are some instructions for one way to setup ssh. (Check which version
of SSH you have by typing "ssh" or "man ssh".)

MacOSX
In general this should be similar to Linux/Unix machines. In versions 10.4 and later,
the native MacOSX tar works, and also supports resource forks. xtar is another
option, and rsync works too (although the MacOSX-supplied rsync has an extension for
extended attributes that is not compatible with standard rsync).

SSH Setup
SSH is a secure way to run tar or rsync on a backup client to extract the data. SSH
provides strong authentication and encryption of the network data.

Note that if you run rsyncd (rsync daemon), ssh is not used. In this case, rsyncd
provides its own authentication, but there is no encryption of network data. If you
want encryption of network data you can use ssh to create a tunnel, or use a program
like stunnel.

Setup instructions for ssh can be found at
<http://backuppc.sourceforge.net/faq/ssh.html> or on the Wiki.

Clients that use DHCP
If a client machine uses DHCP BackupPC needs some way to find the IP address given the
host name. One alternative is to set dhcp to 1 in the hosts file, and BackupPC will
search a pool of IP addresses looking for hosts. More efficiently, it is better to
set dhcp = 0 and provide a mechanism for BackupPC to find the IP address given the
host name.

For WinXX machines BackupPC uses the NetBios name server to determine the IP address
given the host name. For unix machines you can run nmbd (the NetBios name server)
from the Samba distribution so that the machine responds to a NetBios name request.
See the manual page and Samba documentation for more information.

Alternatively, you can set $Conf{NmbLookupFindHostCmd} to any command that returns the
IP address given the host name.

Please read the section "How BackupPC Finds Hosts" for more details.

Step 6: Running BackupPC
The installation contains an init.d backuppc script that can be copied to /etc/init.d so
that BackupPC can auto-start on boot. See init.d/README for further instructions.

BackupPC should be ready to start. If you installed the init.d script, then you should be
able to run BackupPC with:

/etc/init.d/backuppc start

(This script can also be invoked with "stop" to stop BackupPC and "reload" to tell
BackupPC to reload config.pl and the hosts file.)

Otherwise, just run

__INSTALLDIR__/bin/BackupPC -d

as user __BACKUPPCUSER__. The -d option tells BackupPC to run as a daemon (ie: it does an
additional fork).

Any immediate errors will be printed to stderr and BackupPC will quit. Otherwise, look in
__LOGDIR__/LOG and verify that BackupPC reports it has started and all is ok.

Step 7: Talking to BackupPC
You should verify that BackupPC is running by using BackupPC_serverMesg. This sends a
message to BackupPC via the unix (or TCP) socket and prints the response. Like all
BackupPC programs, BackupPC_serverMesg should be run as the BackupPC user
(__BACKUPPCUSER__), so you should

su __BACKUPPCUSER__

before running BackupPC_serverMesg. If the BackupPC user is configured with /bin/false as
the shell, you can use the -s option to su to explicitly run a shell, eg:

su -s /bin/bash __BACKUPPCUSER__

Depending upon your configuration you might also need the -l option.

You can request status information and start and stop backups using this interface. This
socket interface is mainly provided for the CGI interface (and some of the BackupPC sub-
programs use it too). But right now we just want to make sure BackupPC is happy. Each of
these commands should produce some status output:

__INSTALLDIR__/bin/BackupPC_serverMesg status info
__INSTALLDIR__/bin/BackupPC_serverMesg status jobs
__INSTALLDIR__/bin/BackupPC_serverMesg status hosts

The output should be some hashes printed with Data::Dumper. If it looks cryptic and
confusing, and doesn't look like an error message, then all is ok.

The jobs status should initially show just BackupPC_trashClean. The hosts status should
produce a list of every host you have listed in __CONFDIR__/hosts as part of a big cryptic
output line.

You can also request that all hosts be queued:

__INSTALLDIR__/bin/BackupPC_serverMesg backup all

At this point you should make sure the CGI interface works since it will be much easier to
see what is going on. That's our next subject.

Step 8: Checking email delivery
The script BackupPC_sendEmail sends status and error emails to the administrator and
users. It is usually run each night by BackupPC_nightly.

To verify that it can run sendmail and deliver email correctly you should ask it to send a
test email to you:

su __BACKUPPCUSER__
__INSTALLDIR__/bin/BackupPC_sendEmail -u MYNAME@MYDOMAIN.COM

BackupPC_sendEmail also takes a -c option that checks if BackupPC is running, and it sends
an email to $Conf{EMailAdminUserName} if it is not. That can be used as a keep-alive
check by adding

__INSTALLDIR__/bin/BackupPC_sendEmail -c

to __BACKUPPCUSER__'s cron.

The -t option to BackupPC_sendEmail causes it to print the email message instead of
invoking sendmail to deliver the message.

Step 9: CGI interface
The CGI interface script, BackupPC_Admin, is a powerful and flexible way to see and
control what BackupPC is doing. It is written for an Apache server. If you don't have
Apache, see <http://www.apache.org>.

There are two options for setting up the CGI interface: standard mode and using mod_perl.
Mod_perl provides much higher performance (around 15x) and is the best choice if your
Apache was built with mod_perl support. To see if your apache was built with mod_perl run
this command:

httpd -l | egrep mod_perl

If this prints mod_perl.c then your Apache supports mod_perl.

Note: on some distributions (like Debian) the command is not ``httpd'', but ``apache'' or
``apache2''. Those distributions will generally also use ``apache'' for the Apache user
account and configuration files.

Using mod_perl with BackupPC_Admin requires a dedicated Apache to be run as the BackupPC
user (__BACKUPPCUSER__). This is because BackupPC_Admin needs permission to access
various files in BackupPC's data directories. In contrast, the standard installation
(without mod_perl) solves this problem by having BackupPC_Admin installed as setuid to the
BackupPC user, so that BackupPC_Admin runs as the BackupPC user.

Here are some specifics for each setup:

Standard Setup
The CGI interface should have been installed by the configure.pl script in
__CGIDIR__/BackupPC_Admin. BackupPC_Admin should have been installed as setuid to the
BackupPC user (__BACKUPPCUSER__), in addition to user and group execute permission.

You should be very careful about permissions on BackupPC_Admin and the directory
__CGIDIR__: it is important that normal users cannot directly execute or change
BackupPC_Admin, otherwise they can access backup files for any PC. You might need to
change the group ownership of BackupPC_Admin to a group that Apache belongs to so that
Apache can execute it (don't add "other" execute permission!). The permissions should
look like this:

ls -l __CGIDIR__/BackupPC_Admin
-swxr-x--- 1 __BACKUPPCUSER__ web 82406 Jun 17 22:58 __CGIDIR__/BackupPC_Admin

The setuid script won't work unless perl on your machine was installed with setuid
emulation. This is likely the problem if you get an error saying such as "Wrong user:
my userid is 25, instead of 150", meaning the script is running as the httpd user, not
the BackupPC user. This is because setuid scripts are disabled by the kernel in most
flavors of unix and linux.

To see if your perl has setuid emulation, see if there is a program called sperl5.8.0
(or sperl5.8.2 etc, based on your perl version) in the place where perl is installed.
If you can't find this program, then you have two options: rebuild and reinstall perl
with the setuid emulation turned on (answer "y" to the question "Do you want to do
setuid/setgid emulation?" when you run perl's configure script), or switch to the
mod_perl alternative for the CGI script (which doesn't need setuid to work).

Mod_perl Setup
The advantage of the mod_perl setup is that no setuid script is needed, and there is a
huge performance advantage. Not only does all the perl code need to be parsed just
once, the config.pl and hosts files, plus the connection to the BackupPC server are
cached between requests. The typical speedup is around 15 times.

To use mod_perl you need to run Apache as user __BACKUPPCUSER__. If you need to run
multiple Apache's for different services then you need to create multiple top-level
Apache directories, each with their own config file. You can make copies of
/etc/init.d/httpd and use the -d option to httpd to point each http to a different
top-level directory. Or you can use the -f option to explicitly point to the config
file. Multiple Apache's will run on different Ports (eg: 80 is standard, 8080 is a
typical alternative port accessed via http://yourhost.com:8080).

Inside BackupPC's Apache http.conf file you should check the settings for ServerRoot,
DocumentRoot, User, Group, and Port. See
<http://httpd.apache.org/docs/server-wide.html> for more details.

For mod_perl, BackupPC_Admin should not have setuid permission, so you should turn it
off:

chmod u-s __CGIDIR__/BackupPC_Admin

To tell Apache to use mod_perl to execute BackupPC_Admin, add this to Apache's 1.x
httpd.conf file:

<IfModule mod_perl.c>
PerlModule Apache::Registry
PerlTaintCheck On
<Location /cgi-bin/BackupPC/BackupPC_Admin> # <--- change path as needed
SetHandler perl-script
PerlHandler Apache::Registry
Options ExecCGI
PerlSendHeader On
</Location>
</IfModule>

Apache 2.0.44 with Perl 5.8.0 on RedHat 7.1, Don Silvia reports that this works (with
tweaks from Michael Tuzi):

LoadModule perl_module modules/mod_perl.so
PerlModule Apache2

<Directory /path/to/cgi/>
SetHandler perl-script
PerlResponseHandler ModPerl::Registry
PerlOptions +ParseHeaders
Options +ExecCGI
Order deny,allow
Deny from all
Allow from 192.168.0
AuthName "Backup Admin"
AuthType Basic
AuthUserFile /path/to/user_file
Require valid-user
</Directory>

There are other optimizations and options with mod_perl. For example, you can tell
mod_perl to preload various perl modules, which saves memory compared to loading
separate copies in every Apache process after they are forked. See Stas's definitive
mod_perl guide at <http://perl.apache.org/guide>.

BackupPC_Admin requires that users are authenticated by Apache. Specifically, it expects
that Apache sets the REMOTE_USER environment variable when it runs. There are several
ways to do this. One way is to create a .htaccess file in the cgi-bin directory that
looks like:

AuthGroupFile /etc/httpd/conf/group # <--- change path as needed
AuthUserFile /etc/http/conf/passwd # <--- change path as needed
AuthType basic
AuthName "access"
require valid-user

You will also need "AllowOverride Indexes AuthConfig" in the Apache httpd.conf file to
enable the .htaccess file. Alternatively, everything can go in the Apache httpd.conf file
inside a Location directive. The list of users and password file above can be extracted
from the NIS passwd file.

One alternative is to use LDAP. In Apache's http.conf add these lines:

LoadModule auth_ldap_module modules/auth_ldap.so
AddModule auth_ldap.c

# cgi-bin - auth via LDAP (for BackupPC)
<Location /cgi-binBackupPC/BackupPC_Admin> # <--- change path as needed
AuthType Basic
AuthName "BackupPC login"
# replace MYDOMAIN, PORT, ORG and CO as needed
AuthLDAPURL ldap://ldap.MYDOMAIN.com:PORT/o=ORG,c=CO?uid?sub?(objectClass=*)
require valid-user
</Location>

If you want to disable the user authentication you can set $Conf{CgiAdminUsers} to '*',
which allows any user to have full access to all hosts and backups. In this case the
REMOTE_USER environment variable does not have to be set by Apache.

Alternatively, you can force a particular user name by getting Apache to set REMOTE_USER,
eg, to hardcode the user to www you could add this to Apache's httpd.conf:

Finally, you should also edit the config.pl file and adjust, as necessary, the CGI-
specific settings. They're near the end of the config file. In particular, you should
specify which users or groups have administrator (privileged) access: see the config
settings $Conf{CgiAdminUserGroup} and $Conf{CgiAdminUsers}. Also, the configure.pl script
placed various images into $Conf{CgiImageDir} that BackupPC_Admin needs to serve up. You
should make sure that $Conf{CgiImageDirURL} is the correct URL for the image directory.

See the section "Fixing installation problems" for suggestions on debugging the Apache
authentication setup.

How BackupPC Finds Hosts
Starting with v2.0.0 the way hosts are discovered has changed. In most cases you should
specify 0 for the DHCP flag in the conf/hosts file, even if the host has a dynamically
assigned IP address.

BackupPC (starting with v2.0.0) looks up hosts with DHCP = 0 in this manner:

• First DNS is used to lookup the IP address given the client's name using perl's
gethostbyname() function. This should succeed for machines that have fixed IP
addresses that are known via DNS. You can manually see whether a given host have a
DNS entry according to perl's gethostbyname function with this command:

perl -e 'print(gethostbyname("myhost") ? "ok\n" : "not found\n");'

• If gethostbyname() fails, BackupPC then attempts a NetBios multicast to find the host.
Provided your client machine is configured properly, it should respond to this NetBios
multicast request. Specifically, BackupPC runs a command of this form:

nmblookup myhost

If this fails you will see output like:

querying myhost on 10.10.255.255
name_query failed to find name myhost

If it is successful you will see output like:

querying myhost on 10.10.255.255
10.10.1.73 myhost<00>

Depending on your netmask you might need to specify the -B option to nmblookup. For
example:

nmblookup -B 10.10.1.255 myhost

If necessary, experiment with the nmblookup command which will return the IP address
of the client given its name. Then update $Conf{NmbLookupFindHostCmd} with any
necessary options to nmblookup.

For hosts that have the DHCP flag set to 1, these machines are discovered as follows:

• A DHCP address pool ($Conf{DHCPAddressRanges}) needs to be specified. BackupPC will
check the NetBIOS name of each machine in the range using a command of the form:

nmblookup -A W.X.Y.Z

where W.X.Y.Z is each candidate address from $Conf{DHCPAddressRanges}. Any host that
has a valid NetBIOS name returned by this command (ie: matching an entry in the hosts
file) will be backed up. You can modify the specific nmblookup command if necessary
via $Conf{NmbLookupCmd}.

• You only need to use this DHCP feature if your client machine doesn't respond to the
NetBios multicast request:

nmblookup myHost

but does respond to a request directed to its IP address:

nmblookup -A W.X.Y.Z

Other installation topics
Removing a client
If there is a machine that no longer needs to be backed up (eg: a retired machine) you
have two choices. First, you can keep the backups accessible and browsable, but
disable all new backups. Alternatively, you can completely remove the client and all
its backups.

To disable backups for a client $Conf{BackupsDisable} can be set to two different
values in that client's per-PC config.pl file:

1. Don't do any regular backups on this machine. Manually requested backups (via the
CGI interface) will still occur.

2. Don't do any backups on this machine. Manually requested backups (via the CGI
interface) will be ignored.

This will still allow the client's old backups to be browsable and restorable.

To completely remove a client and all its backups, you should remove its entry in the
conf/hosts file, and then delete the __TOPDIR__/pc/$host directory. Whenever you
change the hosts file, you should send BackupPC a HUP (-1) signal so that it re-reads
the hosts file. If you don't do this, BackupPC will automatically re-read the hosts
file at the next regular wakeup.

Note that when you remove a client's backups you won't initially recover much disk
space. That's because the client's files are still in the pool. Overnight, when
BackupPC_nightly next runs, all the unused pool files will be deleted and this will
recover the disk space used by the client's backups.

Copying the pool
If the pool disk requirements grow you might need to copy the entire data directory to
a new (bigger) file system. Hopefully you are lucky enough to avoid this by having
the data directory on a RAID file system or LVM that allows the capacity to be grown
in place by adding disks.

The backup data directories contain large numbers of hardlinks. If you try to copy
the pool the target directory will occupy a lot more space if the hardlinks aren't re-
established.

The best way to copy a pool file system, if possible, is by copying the raw device at
the block level (eg: using dd). Application level programs that understand hardlinks
include the GNU cp program with the -a option and rsync -H. However, the large number
of hardlinks in the pool will make the memory usage large and the copy very slow.
Don't forget to stop BackupPC while the copy runs.

Starting in 3.0.0 a new script bin/BackupPC_tarPCCopy can be used to assist the copy
process. Given one or more pc paths (eg: TOPDIR/pc/HOST or TOPDIR/pc/HOST/nnn),
BackupPC_tarPCCopy creates a tar archive with all the hardlinks pointing to
../cpool/.... Any files not hardlinked (eg: backups, LOG etc) are included verbatim.

You will need to specify the -P option to tar when you extract the archive generated
by BackupPC_tarPCCopy since the hardlink targets are outside of the directory being
extracted.

To copy a complete store (ie: __TOPDIR__) using BackupPC_tarPCCopy you should:

• stop BackupPC so that the store is static.

• copy the cpool, conf and log directory trees using any technique (like cp, rsync
or tar) without the need to preserve hardlinks.

• copy the pc directory using BackupPC_tarPCCopy:

su __BACKUPPCUSER__
cd NEW_TOPDIR
mkdir pc
cd pc
__INSTALLDIR__/bin/BackupPC_tarPCCopy __TOPDIR__/pc | tar xvPf -

Fixing installation problems
Please see the Wiki at <http://backuppc.wiki.sourceforge.net> for debugging suggestions.
If you find a solution to your problem that could help other users please add it to the
Wiki!

Restore functions

BackupPC supports several different methods for restoring files. The most convenient
restore options are provided via the CGI interface. Alternatively, backup files can be
restored using manual commands.

CGI restore options
By selecting a host in the CGI interface, a list of all the backups for that machine will
be displayed. By selecting the backup number you can navigate the shares and directory
tree for that backup.

BackupPC's CGI interface automatically fills incremental backups with the corresponding
full backup, which means each backup has a filled appearance. Therefore, there is no need
to do multiple restores from the incremental and full backups: BackupPC does all the hard
work for you. You simply select the files and directories you want from the correct
backup vintage in one step.

You can download a single backup file at any time simply by selecting it. Your browser
should prompt you with the file name and ask you whether to open the file or save it to
disk.

Alternatively, you can select one or more files or directories in the currently selected
directory and select "Restore selected files". (If you need to restore selected files and
directories from several different parent directories you will need to do that in multiple
steps.)

If you select all the files in a directory, BackupPC will replace the list of files with
the parent directory. You will be presented with a screen that has three options:

Option 1: Direct Restore
With this option the selected files and directories are restored directly back onto
the host, by default in their original location. Any old files with the same name
will be overwritten, so use caution. You can optionally change the target host name,
target share name, and target path prefix for the restore, allowing you to restore the
files to a different location.

Once you select "Start Restore" you will be prompted one last time with a summary of
the exact source and target files and directories before you commit. When you give
the final go ahead the restore operation will be queued like a normal backup job,
meaning that it will be deferred if there is a backup currently running for that host.
When the restore job is run, smbclient, tar, rsync or rsyncd is used (depending upon
$Conf{XferMethod}) to actually restore the files. Sorry, there is currently no option
to cancel a restore that has been started. Currently ftp restores are not fully
implemented.

A record of the restore request, including the result and list of files and
directories, is kept. It can be browsed from the host's home page.
$Conf{RestoreInfoKeepCnt} specifies how many old restore status files to keep.

Note that for direct restore to work, the $Conf{XferMethod} must be able to write to
the client. For example, that means an SMB share for smbclient needs to be writable,
and the rsyncd module needs "read only" set to "false". This creates additional
security risks. If you only create read-only SMB shares (which is a good idea), then
the direct restore will fail. You can disable the direct restore option by setting
$Conf{SmbClientRestoreCmd}, $Conf{TarClientRestoreCmd} and $Conf{RsyncRestoreArgs} to
undef.

Option 2: Download Zip archive
With this option a zip file containing the selected files and directories is
downloaded. The zip file can then be unpacked or individual files extracted as
necessary on the host machine. The compression level can be specified. A value of 0
turns off compression.

When you select "Download Zip File" you should be prompted where to save the
restore.zip file.

BackupPC does not consider downloading a zip file as an actual restore operation, so
the details are not saved for later browsing as in the first case. However, a mention
that a zip file was downloaded by a particular user, and a list of the files, does
appear in BackupPC's log file.

Option 3: Download Tar archive
This is identical to the previous option, except a tar file is downloaded rather than
a zip file (and there is currently no compression option).

Command-line restore options
Apart from the CGI interface, BackupPC allows you to restore files and directories from
the command line. The following programs can be used:

BackupPC_zcat
For each file name argument it inflates (uncompresses) the file and writes it to
stdout. To use BackupPC_zcat you could give it the full file name, eg:

__INSTALLDIR__/bin/BackupPC_zcat __TOPDIR__/pc/host/5/fc/fcraig/fexample.txt > example.txt

It's your responsibility to make sure the file is really compressed: BackupPC_zcat
doesn't check which backup the requested file is from. BackupPC_zcat returns a non-
zero status if it fails to uncompress a file.

BackupPC_tarCreate
BackupPC_tarCreate creates a tar file for any files or directories in a particular
backup. Merging of incrementals is done automatically, so you don't need to worry
about whether certain files appear in the incremental or full backup.

The usage is:

BackupPC_tarCreate [options] files/directories...
Required options:
-h host host from which the tar archive is created
-n dumpNum dump number from which the tar archive is created
A negative number means relative to the end (eg -1
means the most recent dump, -2 2nd most recent etc).
-s shareName share name from which the tar archive is created

Other options:
-t print summary totals
-r pathRemove path prefix that will be replaced with pathAdd
-p pathAdd new path prefix
-b BLOCKS BLOCKS x 512 bytes per record (default 20; same as tar)
-w writeBufSz write buffer size (default 1048576 = 1MB)
-e charset charset for encoding file names (default: value of
$Conf{ClientCharset} when backup was done)
-l just print a file listing; don't generate an archive
-L just print a detailed file listing; don't generate an archive

The command-line files and directories are relative to the specified shareName. The
tar file is written to stdout.

The -h, -n and -s options specify which dump is used to generate the tar archive. The
-r and -p options can be used to relocate the paths in the tar archive so extracted
files can be placed in a location different from their original location.

BackupPC_zipCreate
BackupPC_zipCreate creates a zip file for any files or directories in a particular
backup. Merging of incrementals is done automatically, so you don't need to worry
about whether certain files appear in the incremental or full backup.

The usage is:

BackupPC_zipCreate [options] files/directories...
Required options:
-h host host from which the zip archive is created
-n dumpNum dump number from which the tar archive is created
A negative number means relative to the end (eg -1
means the most recent dump, -2 2nd most recent etc).
-s shareName share name from which the zip archive is created

Other options:
-t print summary totals
-r pathRemove path prefix that will be replaced with pathAdd
-p pathAdd new path prefix
-c level compression level (default is 0, no compression)
-e charset charset for encoding file names (default: utf8)

The command-line files and directories are relative to the specified shareName. The
zip file is written to stdout. The -h, -n and -s options specify which dump is used to
generate the zip archive. The -r and -p options can be used to relocate the paths in
the zip archive so extracted files can be placed in a location different from their
original location.

Each of these programs reside in __INSTALLDIR__/bin.

Archive functions

BackupPC supports archiving to removable media. For users that require offsite backups,
BackupPC can create archives that stream to tape devices, or create files of specified
sizes to fit onto cd or dvd media.

Each archive type is specified by a BackupPC host with its XferMethod set to 'archive'.
This allows for multiple configurations at sites where there might be a combination of
tape and cd/dvd backups being made.

BackupPC provides a menu that allows one or more hosts to be archived. The most recent
backup of each host is archived using BackupPC_tarCreate, and the output is optionally
compressed and split into fixed-sized files (eg: 650MB).

The archive for each host is done by default using
__INSTALLDIR__/bin/BackupPC_archiveHost. This script can be copied and customized as
needed.

Configuring an Archive Host
To create an Archive Host, add it to the hosts file just as any other host and call it a
name that best describes the type of archive, e.g. ArchiveDLT

To tell BackupPC that the Host is for Archives, create a config.pl file in the Archive
Hosts's pc directory, adding the following line:

$Conf{XferMethod} = 'archive';

To further customise the archive's parameters you can adding the changed parameters in the
host's config.pl file. The parameters are explained in the config.pl file. Parameters may
be fixed or the user can be allowed to change them (eg: output device).

The per-host archive command is $Conf{ArchiveClientCmd}. By default this invokes

__INSTALLDIR__/bin/BackupPC_archiveHost

which you can copy and customize as necessary.

Starting an Archive
In the web interface, click on the Archive Host you wish to use. You will see a list of
previous archives and a summary on each. By clicking the "Start Archive" button you are
presented with the list of hosts and the approximate backup size (note this is raw size,
not projected compressed size) Select the hosts you wish to archive and press the "Archive
Selected Hosts" button.

The next screen allows you to adjust the parameters for this archive run. Press the
"Start the Archive" to start archiving the selected hosts with the parameters displayed.

Starting an Archive from the command line
The script BackupPC_archiveStart can be used to start an archive from the command line (or
cron etc). The usage is:

BackupPC_archiveStart archiveHost userName hosts...

This creates an archive of the most recent backup of each of the specified hosts. The
first two arguments are the archive host and the user name making the request.

Other CGI Functions

Configuration and Host Editor
The CGI interface has a complete configuration and host editor. Only the administrator
can edit the main configuration settings and hosts. The edit links are in the left
navigation bar.

When changes are made to any parameter a "Save" button appears at the top of the page. If
you are editing a text box you will need to click outside of the text box to make the Save
button appear. If you don't select Save then the changes won't be saved.

The host-specific configuration can be edited from the host summary page using the link in
the left navigation bar. The administrator can edit any of the host-specific
configuration settings.

When editing the host-specific configuration, each parameter has an "override" setting
that denotes the value is host-specific, meaning that it overrides the setting in the main
configuration. If you unselect "override" then the setting is removed from the host-
specific configuration, and the main configuration file is displayed.

User's can edit their host-specific configuration if enabled via
$Conf{CgiUserConfigEditEnable}. The specific subset of configuration settings that a user
can edit is specified with $Conf{CgiUserConfigEdit}. It is recommended to make this list
short as possible (you probably don't want your users saving dozens of backups) and it is
essential that they can't edit any of the Cmd configuration settings, otherwise they can
specify an arbitrary command that will be executed as the BackupPC user.

RSS
BackupPC supports a very basic RSS feed. Provided you have the XML::RSS perl module
installed, a URL similar to this will provide RSS information:

http://localhost/cgi-bin/BackupPC/BackupPC_Admin?action=rss

This feature is experimental. The information included will probably change.

BackupPC Design

Some design issues
Pooling common files
To quickly see if a file is already in the pool, an MD5 digest of the file length and
contents is used as the file name in the pool. This can't guarantee a file is
identical: it just reduces the search to often a single file or handful of files. A
complete file comparison is always done to verify if two files are really the same.

Identical files on multiples backups are represented by hard links. Hardlinks are
used so that identical files all refer to the same physical file on the server's disk.
Also, hard links maintain reference counts so that BackupPC knows when to delete
unused files from the pool.

For the computer-science majors among you, you can think of the pooling system used by
BackupPC as just a chained hash table stored on a (big) file system.

The hashing function
There is a tradeoff between how much of file is used for the MD5 digest and the time
taken comparing all the files that have the same hash.

Using the file length and just the first 4096 bytes of the file for the MD5 digest
produces some repetitions. One example: with 900,000 unique files in the pool, this
hash gives about 7,000 repeated files, and in the worst case 500 files have the same
hash. That's not bad: we only have to do a single file compare 99.2% of the time.
But in the worst case we have to compare as many as 500 files checking for a match.

With a modest increase in CPU time, if we use the file length and the first 256K of
the file we now only have 500 repeated files and in the worst case around 20 files
have the same hash. Furthermore, if we instead use the first and last 128K of the file
(more specifically, the first and eighth 128K chunks for files larger than 1MB) we get
only 300 repeated files and in the worst case around 20 files have the same hash.

Based on this experimentation, this is the hash function used by BackupPC. It is
important that you don't change the hash function after files are already in the pool.
Otherwise your pool will grow to twice the size until all the old backups (and all the
old files with old hashes) eventually expire.

Compression
BackupPC supports compression. It uses the deflate and inflate methods in the
Compress::Zlib module, which is based on the zlib compression library (see
<http://www.gzip.org/zlib/>).

The $Conf{CompressLevel} setting specifies the compression level to use. Zero (0)
means no compression. Compression levels can be from 1 (least cpu time, slightly worse
compression) to 9 (most cpu time, slightly better compression). The recommended value
is 3. Changing it to 5, for example, will take maybe 20% more cpu time and will get
another 2-3% additional compression. Diminishing returns set in above 5. See the zlib
documentation for more information about compression levels.

BackupPC implements compression with minimal CPU load. Rather than compressing every
incoming backup file and then trying to match it against the pool, BackupPC computes
the MD5 digest based on the uncompressed file, and matches against the candidate pool
files by comparing each uncompressed pool file against the incoming backup file.
Since inflating a file takes roughly a factor of 10 less CPU time than deflating there
is a big saving in CPU time.

The combination of pooling common files and compression can yield a factor of 8 or
more overall saving in backup storage.

BackupPC operation
BackupPC reads the configuration information from __CONFDIR__/config.pl. It then runs and
manages all the backup activity. It maintains queues of pending backup requests, user
backup requests and administrative commands. Based on the configuration various requests
will be executed simultaneously.

As specified by $Conf{WakeupSchedule}, BackupPC wakes up periodically to queue backups on
all the PCs. This is a four step process:

1. For each host and DHCP address backup requests are queued on the background command
queue.

2. For each PC, BackupPC_dump is forked. Several of these may be run in parallel, based
on the configuration. First a ping is done to see if the machine is alive. If this is
a DHCP address, nmblookup is run to get the netbios name, which is used as the host
name. If DNS lookup fails, $Conf{NmbLookupFindHostCmd} is run to find the IP address
from the host name. The file __TOPDIR__/pc/$host/backups is read to decide whether a
full or incremental backup needs to be run. If no backup is scheduled, or the ping to
$host fails, then BackupPC_dump exits.

The backup is done using the specified XferMethod. Either samba's smbclient or tar
over ssh/rsh/nfs piped into BackupPC_tarExtract, or rsync over ssh/rsh is run, or
rsyncd is connected to, with the incoming data extracted to __TOPDIR__/pc/$host/new.
The XferMethod output is put into __TOPDIR__/pc/$host/XferLOG.

The letter in the XferLOG file shows the type of object, similar to the first letter
of the modes displayed by ls -l:

d -> directory
l -> symbolic link
b -> block special file
c -> character special file
p -> pipe file (fifo)
nothing -> regular file

The words mean:

create
new for this backup (ie: directory or file not in pool)

pool
found a match in the pool

same
file is identical to previous backup (contents were checksummed and verified
during full dump).

skip
file skipped in incremental because attributes are the same (only displayed if
$Conf{XferLogLevel} >= 2).

As BackupPC_tarExtract extracts the files from smbclient or tar, or as rsync or ftp
runs, it checks each file in the backup to see if it is identical to an existing file
from any previous backup of any PC. It does this without needed to write the file to
disk. If the file matches an existing file, a hardlink is created to the existing file
in the pool. If the file does not match any existing files, the file is written to
disk and the file name is saved in __TOPDIR__/pc/$host/NewFileList for later
processing by BackupPC_link. BackupPC_tarExtract and rsync can handle arbitrarily
large files and multiple candidate matching files without needing to write the file to
disk in the case of a match. This significantly reduces disk writes (and also reads,
since the pool file comparison is done disk to memory, rather than disk to disk).

Based on the configuration settings, BackupPC_dump checks each old backup to see if
any should be removed. Any expired backups are moved to __TOPDIR__/trash for later
removal by BackupPC_trashClean.

3. For each complete, good, backup, BackupPC_link is run. To avoid race conditions as
new files are linked into the pool area, only a single BackupPC_link program runs at a
time and the rest are queued.

BackupPC_link reads the NewFileList written by BackupPC_dump and inspects each new
file in the backup. It re-checks if there is a matching file in the pool (another
BackupPC_link could have added the file since BackupPC_dump checked). If so, the file
is removed and replaced by a hard link to the existing file. If the file is new, a
hard link to the file is made in the pool area, so that this file is available for
checking against each new file and new backup.

Then, if $Conf{IncrFill} is set (note that the default setting is off), for each
incremental backup, hard links are made in the new backup to all files that were not
extracted during the incremental backups. The means the incremental backup looks like
a complete image of the PC (with the exception that files that were removed on the PC
since the last full backup will still appear in the backup directory tree).

The CGI interface knows how to merge unfilled incremental backups will the most recent
prior filled (full) backup, giving the incremental backups a filled appearance. The
default for $Conf{IncrFill} is off, since there is no need to fill incremental
backups. This saves some level of disk activity, since lots of extra hardlinks are no
longer needed (and don't have to be deleted when the backup expires).

4. BackupPC_trashClean is always run in the background to remove any expired backups.
Every 5 minutes it wakes up and removes all the files in __TOPDIR__/trash.

Also, once each night, BackupPC_nightly is run to complete some additional
administrative tasks, such as cleaning the pool. This involves removing any files in
the pool that only have a single hard link (meaning no backups are using that file).
Again, to avoid race conditions, BackupPC_nightly is only run when there are no
BackupPC_link processes running. When BackupPC_nightly is run no new BackupPC_link
jobs are started. If BackupPC_nightly takes too long to run, the settings
$Conf{MaxBackupPCNightlyJobs} and $Conf{BackupPCNightlyPeriod} can be used to run
several BackupPC_nightly processes in parallel, and to split its job over several
nights.

BackupPC also listens for TCP connections on $Conf{ServerPort}, which is used by the CGI
script BackupPC_Admin for status reporting and user-initiated backup or backup cancel
requests.

Storage layout
BackupPC resides in several directories:

__INSTALLDIR__
Perl scripts comprising BackupPC reside in __INSTALLDIR__/bin, libraries are in
__INSTALLDIR__/lib and documentation is in __INSTALLDIR__/doc.

__CGIDIR__
The CGI script BackupPC_Admin resides in this cgi binary directory.

__CONFDIR__
All the configuration information resides below __CONFDIR__. This directory contains:

The directory __CONFDIR__ contains:

config.pl
Configuration file. See "Configuration File" below for more details.

hosts
Hosts file, which lists all the PCs to backup.

pc The directory __CONFDIR__/pc contains per-client configuration files that override
settings in the main configuration file. Each file is named
__CONFDIR__/pc/HOST.pl, where HOST is the host name.

In pre-FHS versions of BackupPC these files were located in
__TOPDIR__/pc/HOST/config.pl.

__LOGDIR__
The directory __LOGDIR__ (__TOPDIR__/log on pre-FHS versions of BackupPC) contains:

LOG Current (today's) log file output from BackupPC.

LOG.0 or LOG.0.z
Yesterday's log file output. Log files are aged daily and compressed (if
compression is enabled), and old LOG files are deleted.

BackupPC.pid
Contains BackupPC's process id.

status.pl
A summary of BackupPC's status written periodically by BackupPC so that certain
state information can be maintained if BackupPC is restarted. Should not be
edited.

UserEmailInfo.pl
A summary of what email was last sent to each user, and when the last email was
sent. Should not be edited.

__TOPDIR__
All of BackupPC's data (PC backup images, logs, configuration information) is stored
below this directory.

Below __TOPDIR__ are several directories:

__TOPDIR__/trash
Any directories and files below this directory are periodically deleted whenever
BackupPC_trashClean checks. When a backup is aborted or when an old backup
expires, BackupPC_dump simply moves the directory to __TOPDIR__/trash for later
removal by BackupPC_trashClean.

__TOPDIR__/pool
All uncompressed files from PC backups are stored below __TOPDIR__/pool. Each
file's name is based on the MD5 hex digest of the file contents. Specifically,
for files less than 256K, the file length and the entire file is used. For files
up to 1MB, the file length and the first and last 128K are used. Finally, for
files longer than 1MB, the file length, and the first and eighth 128K chunks for
the file are used.

Each file is stored in a subdirectory X/Y/Z, where X, Y, Z are the first 3 hex
digits of the MD5 digest.

For example, if a file has an MD5 digest of 123456789abcdef0, the file is stored
in __TOPDIR__/pool/1/2/3/123456789abcdef0.

The MD5 digest might not be unique (especially since not all the file's contents
are used for files bigger than 256K). Different files that have the same MD5
digest are stored with a trailing suffix "_n" where n is an incrementing number
starting at 0. So, for example, if two additional files were identical to the
first, except the last byte was different, and assuming the file was larger than
1MB (so the MD5 digests are the same but the files are actually different), the
three files would be stored as:

__TOPDIR__/pool/1/2/3/123456789abcdef0
__TOPDIR__/pool/1/2/3/123456789abcdef0_0
__TOPDIR__/pool/1/2/3/123456789abcdef0_1

Both BackupPC_dump (actually, BackupPC_tarExtract) and BackupPC_link are
responsible for checking newly backed up files against the pool. For each file,
the MD5 digest is used to generate a file name in the pool directory. If the file
exists in the pool, the contents are compared. If there is no match, additional
files ending in "_n" are checked. (Actually, BackupPC_tarExtract compares
multiple candidate files in parallel.) If the file contents exactly match, the
file is created by simply making a hard link to the pool file (this is done by
BackupPC_tarExtract as the backup proceeds). Otherwise, BackupPC_tarExtract writes
the new file to disk and a new hard link is made in the pool to the file (this is
done later by BackupPC_link).

Therefore, every file in the pool will have at least 2 hard links (one for the
pool file and one for the backup file below __TOPDIR__/pc). Identical files from
different backups or PCs will all be linked to the same file. When old backups
are deleted, some files in the pool might only have one link. BackupPC_nightly
checks the entire pool and removes all files that have only a single link, thereby
recovering the storage for that file.

One other issue: zero length files are not pooled, since there are a lot of these
files and on most file systems it doesn't save any disk space to turn these files
into hard links.

__TOPDIR__/cpool
All compressed files from PC backups are stored below __TOPDIR__/cpool. Its
layout is the same as __TOPDIR__/pool, and the hashing function is the same (and,
importantly, based on the uncompressed file, not the compressed file).

__TOPDIR__/pc/$host
For each PC $host, all the backups for that PC are stored below the directory
__TOPDIR__/pc/$host. This directory contains the following files:

LOG Current log file for this PC from BackupPC_dump.

LOG.DDMMYYYY or LOG.DDMMYYYY.z
Last month's log file. Log files are aged monthly and compressed (if
compression is enabled), and old LOG files are deleted. In earlier versions
of BackupPC these files used to have a suffix of 0, 1, ....

XferERR or XferERR.z
Output from the transport program (ie: smbclient, tar, rsync or ftp) for the
most recent failed backup.

new Subdirectory in which the current backup is stored. This directory is renamed
if the backup succeeds.

XferLOG or XferLOG.z
Output from the transport program (ie: smbclient, tar, rsync or ftp) for the
current backup.

nnn (an integer)
Successful backups are in directories numbered sequentially starting at 0.

XferLOG.nnn or XferLOG.nnn.z
Output from the transport program (ie: smbclient, tar, rsync or ftp)
corresponding to backup number nnn.

RestoreInfo.nnn
Information about restore request #nnn including who, what, when, and why.
This file is in Data::Dumper format. (Note that the restore numbers are not
related to the backup number.)

RestoreLOG.nnn.z
Output from smbclient, tar or rsync during restore #nnn. (Note that the
restore numbers are not related to the backup number.)

ArchiveInfo.nnn
Information about archive request #nnn including who, what, when, and why.
This file is in Data::Dumper format. (Note that the archive numbers are not
related to the restore or backup number.)

ArchiveLOG.nnn.z
Output from archive #nnn. (Note that the archive numbers are not related to
the backup or restore number.)

config.pl
Old location of optional configuration settings specific to this host.
Settings in this file override the main configuration file. In new versions
of BackupPC the per-host configuration files are stored in
__CONFDIR__/pc/HOST.pl.

backups
A tab-delimited ascii table listing information about each successful backup,
one per row. The columns are:

num The backup number, an integer that starts at 0 and increments for each
successive backup. The corresponding backup is stored in the directory
num (eg: if this field is 5, then the backup is stored in
__TOPDIR__/pc/$host/5).

type
Set to "full" or "incr" for full or incremental backup.

startTime
Start time of the backup in unix seconds.

endTime
Stop time of the backup in unix seconds.

nFiles
Number of files backed up (as reported by smbclient, tar, rsync or ftp).

size
Total file size backed up (as reported by smbclient, tar, rsync or ftp).

nFilesExist
Number of files that were already in the pool (as determined by
BackupPC_dump and BackupPC_link).

sizeExist
Total size of files that were already in the pool (as determined by
BackupPC_dump and BackupPC_link).

nFilesNew
Number of files that were not in the pool (as determined by
BackupPC_link).

sizeNew
Total size of files that were not in the pool (as determined by
BackupPC_link).

xferErrs
Number of errors or warnings from smbclient, tar, rsync or ftp.

xferBadFile
Number of errors from smbclient that were bad file errors (zero
otherwise).

xferBadShare
Number of errors from smbclient that were bad share errors (zero
otherwise).

tarErrs
Number of errors from BackupPC_tarExtract.

compress
The compression level used on this backup. Zero or empty means no
compression.

sizeExistComp
Total compressed size of files that were already in the pool (as
determined by BackupPC_dump and BackupPC_link).

sizeNewComp
Total compressed size of files that were not in the pool (as determined by
BackupPC_link).

noFill
Set if this backup has not been filled in with the most recent previous
filled or full backup. See $Conf{IncrFill}.

fillFromNum
If this backup was filled (ie: noFill is 0) then this is the number of the
backup that it was filled from

mangle
Set if this backup has mangled file names and attributes. Always true for
backups in v1.4.0 and above. False for all backups prior to v1.4.0.

xferMethod
Set to the value of $Conf{XferMethod} when this dump was done.

level
The level of this dump. A full dump is level 0. Currently incrementals
are 1. But when multi-level incrementals are supported this will reflect
each dump's incremental level.

restores
A tab-delimited ascii table listing information about each requested restore,
one per row. The columns are:

num Restore number (matches the suffix of the RestoreInfo.nnn and
RestoreLOG.nnn.z file), unrelated to the backup number.

startTime
Start time of the restore in unix seconds.

endTime
End time of the restore in unix seconds.

result
Result (ok or failed).

errorMsg
Error message if restore failed.

nFiles
Number of files restored.

size
Size in bytes of the restored files.

tarCreateErrs
Number of errors from BackupPC_tarCreate during restore.

xferErrs
Number of errors from smbclient, tar, rsync or ftp during restore.

archives
A tab-delimited ascii table listing information about each requested archive,
one per row. The columns are:

num Archive number (matches the suffix of the ArchiveInfo.nnn and
ArchiveLOG.nnn.z file), unrelated to the backup or restore number.

startTime
Start time of the restore in unix seconds.

endTime
End time of the restore in unix seconds.

result
Result (ok or failed).

errorMsg
Error message if archive failed.

Compressed file format
The compressed file format is as generated by Compress::Zlib::deflate with one minor, but
important, tweak. Since Compress::Zlib::inflate fully inflates its argument in memory, it
could take large amounts of memory if it was inflating a highly compressed file. For
example, a 200MB file of 0x0 bytes compresses to around 200K bytes. If
Compress::Zlib::inflate was called with this single 200K buffer, it would need to allocate
200MB of memory to return the result.

BackupPC watches how efficiently a file is compressing. If a big file has very high
compression (meaning it will use too much memory when it is inflated), BackupPC calls the
flush() method, which gracefully completes the current compression. BackupPC then starts
another deflate and simply appends the output file. So the BackupPC compressed file
format is one or more concatenated deflations/flushes. The specific ratios that BackupPC
uses is that if a 6MB chunk compresses to less than 64K then a flush will be done.

Back to the example of the 200MB file of 0x0 bytes. Adding flushes every 6MB adds only
200 or so bytes to the 200K output. So the storage cost of flushing is negligible.

To easily decompress a BackupPC compressed file, the script BackupPC_zcat can be found in
__INSTALLDIR__/bin. For each file name argument it inflates the file and writes it to
stdout.

Rsync checksum caching
An incremental backup with rsync compares attributes on the client with the last full
backup. Any files with identical attributes are skipped. A full backup with rsync sets
the --ignore-times option, which causes every file to be examined independent of
attributes.

Each file is examined by generating block checksums (default 2K blocks) on the receiving
side (that's the BackupPC side), sending those checksums to the client, where the remote
rsync matches those checksums with the corresponding file. The matching blocks and new
data is sent back, allowing the client file to be reassembled. A checksum for the entire
file is sent to as an extra check the the reconstructed file is correct.

This results in significant disk IO and computation for BackupPC: every file in a full
backup, or any file with non-matching attributes in an incremental backup, needs to be
uncompressed, block checksums computed and sent. Then the receiving side reassembles the
file and has to verify the whole-file checksum. Even if the file is identical, prior to
2.1.0, BackupPC had to read and uncompress the file twice, once to compute the block
checksums and later to verify the whole-file checksum.

Starting in 2.1.0, BackupPC supports optional checksum caching, which means the block and
file checksums only need to be computed once for each file. This results in a significant
performance improvement. This only works for compressed pool files. It is enabled by
adding

'--checksum-seed=32761',

to $Conf{RsyncArgs} and $Conf{RsyncRestoreArgs}.

Rsync versions prior to and including rsync-2.6.2 need a small patch to add support for
the --checksum-seed option. This patch is available in the cygwin-rsyncd package at
<http://backuppc.sourceforge.net>. This patch is already included in rsync CVS, so it
will be standard in future versions of rsync.

When this option is present, BackupPC will add block and file checksums to the compressed
pool file the next time a pool file is used and it doesn't already have cached checksums.
The first time a new file is written to the pool, the checksums are not appended. The
next time checksums are needed for a file, they are computed and added. So the full
performance benefit of checksum caching won't be noticed until the third time a pool file
is used (eg: the third full backup).

With checksum caching enabled, there is a risk that should a file's contents in the pool
be corrupted due to a disk problem, but the cached checksums are still correct, the
corruption will not be detected by a full backup, since the file contents are no longer
read and compared. To reduce the chance that this remains undetected, BackupPC can
recheck cached checksums for a fraction of the files. This fraction is set with the
$Conf{RsyncCsumCacheVerifyProb} setting. The default value of 0.01 means that 1% of the
time a file's checksums are read, the checksums are verified. This reduces performance
slightly, but, over time, ensures that files contents are in sync with the cached
checksums.

The format of the cached checksum data can be discovered by looking at the code.
Basically, the first byte of the compressed file is changed to denote that checksums are
appended. The block and file checksum data, plus some other information and magic word,
are appended to the compressed file. This allows the cache update to be done in-place.

File name mangling
Backup file names are stored in "mangled" form. Each node of a path is preceded by "f"
(mnemonic: file), and special characters (\n, \r, % and /) are URI-encoded as "%xx", where
xx is the ascii character's hex value. So c:/craig/example.txt is now stored as
fc/fcraig/fexample.txt.

This was done mainly so meta-data could be stored alongside the backup files without name
collisions. In particular, the attributes for the files in a directory are stored in a
file called "attrib", and mangling avoids file name collisions (I discarded the idea of
having a duplicate directory tree for every backup just to store the attributes). Other
meta-data (eg: rsync checksums) could be stored in file names preceded by, eg, "c". There
are two other benefits to mangling: the share name might contain "/" (eg: "/home/craig"
for tar transport), and I wanted that represented as a single level in the storage tree.
Secondly, as files are written to NewFileList for later processing by BackupPC_link,
embedded newlines in the file's path will cause problems which are avoided by mangling.

The CGI script undoes the mangling, so it is invisible to the user. Old (unmangled)
backups are still supported by the CGI interface.

Special files
Linux/unix file systems support several special file types: symbolic links, character and
block device files, fifos (pipes) and unix-domain sockets. All except unix-domain sockets
are supported by BackupPC (there's no point in backing up or restoring unix-domain sockets
since they only have meaning after a process creates them). Symbolic links are stored as a
plain file whose contents are the contents of the link (not the file it points to). This
file is compressed and pooled like any normal file. Character and block device files are
also stored as plain files, whose contents are two integers separated by a comma; the
numbers are the major and minor device number. These files are compressed and pooled like
any normal file. Fifo files are stored as empty plain files (which are not pooled since
they have zero size). In all cases, the original file type is stored in the attrib file so
it can be correctly restored.

Hardlinks are also supported. When GNU tar first encounters a file with more than one
link (ie: hardlinks) it dumps it as a regular file. When it sees the second and
subsequent hardlinks to the same file, it dumps just the hardlink information. BackupPC
correctly recognizes these hardlinks and stores them just like symlinks: a regular text
file whose contents is the path of the file linked to. The CGI script will download the
original file when you click on a hardlink.

Also, BackupPC_tarCreate has enough magic to re-create the hardlinks dynamically based on
whether or not the original file and hardlinks are both included in the tar file. For
example, imagine a/b/x is a hardlink to a/c/y. If you use BackupPC_tarCreate to restore
directory a, then the tar file will include a/b/x as the original file and a/c/y will be a
hardlink to a/b/x. If, instead you restore a/c, then the tar file will include a/c/y as
the original file, not a hardlink.

Attribute file format
The unix attributes for the contents of a directory (all the files and directories in that
directory) are stored in a file called attrib. There is a single attrib file for each
directory in a backup. For example, if c:/craig contains a single file
c:/craig/example.txt, that file would be stored as fc/fcraig/fexample.txt and there would
be an attribute file in fc/fcraig/attrib (and also fc/attrib and ./attrib). The file
fc/fcraig/attrib would contain a single entry containing the attributes for
fc/fcraig/fexample.txt.

The attrib file starts with a magic number, followed by the concatenation of the following
information for each file:

• File name length in perl's pack "w" format (variable length base 128).

• File name.

• The unix file type, mode, uid, gid and file size divided by 4GB and file size modulo
4GB (type mode uid gid sizeDiv4GB sizeMod4GB), in perl's pack "w" format (variable
length base 128).

• The unix mtime (unix seconds) in perl's pack "N" format (32 bit integer).

The attrib file is also compressed if compression is enabled. See the
lib/BackupPC/Attrib.pm module for full details.

Attribute files are pooled just like normal backup files. This saves space if all the
files in a directory have the same attributes across multiple backups, which is common.

Optimizations
BackupPC doesn't care about the access time of files in the pool since it saves attribute
meta-data separate from the files. Since BackupPC mostly does reads from disk,
maintaining the access time of files generates a lot of unnecessary disk writes. So,
provided BackupPC has a dedicated data disk, you should consider mounting BackupPC's data
directory with the noatime (or, with Linux kernels >=2.6.20, relatime) attribute (see
mount(1)).

Limitations
BackupPC isn't perfect (but it is getting better). Please see
<http://backuppc.sourceforge.net/faq/limitations.html> for a discussion of some of
BackupPC's limitations.

Security issues
Please see <http://backuppc.sourceforge.net/faq/security.html> for a discussion of some of
various security issues.

Configuration File

The BackupPC configuration file resides in __CONFDIR__/config.pl. Optional per-PC
configuration files reside in __CONFDIR__/pc/$host.pl (or __TOPDIR__/pc/$host/config.pl in
non-FHS versions of BackupPC). This file can be used to override settings just for a
particular PC.

Modifying the main configuration file
The configuration file is a perl script that is executed by BackupPC, so you should be
careful to preserve the file syntax (punctuation, quotes etc) when you edit it. It is
recommended that you use CVS, RCS or some other method of source control for changing
config.pl.

BackupPC reads or re-reads the main configuration file and the hosts file in three cases:

• Upon startup.

• When BackupPC is sent a HUP (-1) signal. Assuming you installed the init.d script,
you can also do this with "/etc/init.d/backuppc reload".

• When the modification time of config.pl file changes. BackupPC checks the
modification time once during each regular wakeup.

Whenever you change the configuration file you can either do a kill -HUP BackupPC_pid or
simply wait until the next regular wakeup period.

Each time the configuration file is re-read a message is reported in the LOG file, so you
can tail it (or view it via the CGI interface) to make sure your kill -HUP worked. Errors
in parsing the configuration file are also reported in the LOG file.

The optional per-PC configuration file (__CONFDIR__/pc/$host.pl or
__TOPDIR__/pc/$host/config.pl in non-FHS versions of BackupPC) is read whenever it is
needed by BackupPC_dump, BackupPC_link and others.

Configuration Parameters

       The configuration parameters are divided into five general groups.  The first group
       (general server configuration) provides general configuration for BackupPC.  The next two
       groups describe what to backup, when to do it, and how long to keep it.  The fourth group
       are settings for email reminders, and the final group contains settings for the CGI
       interface.

       All configuration settings in the second through fifth groups can be overridden by the
       per-PC config.pl file.

   General server configuration
       $Conf{ServerHost} = '';
           Host name on which the BackupPC server is running.

       $Conf{ServerPort} = -1;
           TCP port number on which the BackupPC server listens for and accepts connections.
           Normally this should be disabled (set to -1).  The TCP port is only needed if apache
           runs on a different machine from BackupPC.  In that case, set this to any spare port
           number over 1024 (eg: 2359).  If you enable the TCP port, make sure you set
           $Conf{ServerMesgSecret} too!

       $Conf{ServerMesgSecret} = '';
           Shared secret to make the TCP port secure.  Set this to a hard to guess string if you
           enable the TCP port (ie: $Conf{ServerPort} > 0).

           To avoid possible attacks via the TCP socket interface, every client message is
           protected by an MD5 digest. The MD5 digest includes four items:
             - a seed that is sent to the client when the connection opens
             - a sequence number that increments for each message
             - a shared secret that is stored in $Conf{ServerMesgSecret}
             - the message itself.

           The message is sent in plain text preceded by the MD5 digest.  A snooper can see the
           plain-text seed sent by BackupPC and plain-text message from the client, but cannot
           construct a valid MD5 digest since the secret $Conf{ServerMesgSecret} is unknown.  A
           replay attack is not possible since the seed changes on a per-connection and per-
           message basis.

       $Conf{MyPath} = '/bin';
           PATH setting for BackupPC.  An explicit value is necessary for taint mode.  Value
           shouldn't matter too much since all execs use explicit paths.  However, taint mode in
           perl will complain if this directory is world writable.

       $Conf{UmaskMode} = 027;
           Permission mask for directories and files created by BackupPC.  Default value prevents
           any access from group other, and prevents group write.

       $Conf{WakeupSchedule} = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18,
       19, 20, 21, 22, 23];
           Times at which we wake up, check all the PCs, and schedule necessary backups.  Times
           are measured in hours since midnight.  Can be fractional if necessary (eg: 4.25 means
           4:15am).

           If the hosts you are backing up are always connected to the network you might have
           only one or two wakeups each night.  This will keep the backup activity after hours.
           On the other hand, if you are backing up laptops that are only intermittently
           connected to the network you will want to have frequent wakeups (eg: hourly) to
           maximize the chance that each laptop is backed up.

           Examples:

               $Conf{WakeupSchedule} = [22.5];         # once per day at 10:30 pm.
               $Conf{WakeupSchedule} = [2,4,6,8,10,12,14,16,18,20,22];  # every 2 hours

           The default value is every hour except midnight.

           The first entry of $Conf{WakeupSchedule} is when BackupPC_nightly is run.  You might
           want to re-arrange the entries in $Conf{WakeupSchedule} (they don't have to be
           ascending) so that the first entry is when you want BackupPC_nightly to run (eg: when
           you don't expect a lot of regular backups to run).

       $Conf{MaxBackups} = 4;
           Maximum number of simultaneous backups to run.  If there are no user backup requests
           then this is the maximum number of simultaneous backups.

       $Conf{MaxUserBackups} = 4;
           Additional number of simultaneous backups that users can run.  As many as
           $Conf{MaxBackups} + $Conf{MaxUserBackups} requests can run at the same time.

       $Conf{MaxPendingCmds} = 15;
           Maximum number of pending link commands. New backups will only be started if there are
           no more than $Conf{MaxPendingCmds} plus $Conf{MaxBackups} number of pending link
           commands, plus running jobs.  This limit is to make sure BackupPC doesn't fall too far
           behind in running BackupPC_link commands.

       $Conf{CmdQueueNice} = 10;
           Nice level at which CmdQueue commands (eg: BackupPC_link and BackupPC_nightly) are run
           at.

       $Conf{MaxBackupPCNightlyJobs} = 2;
           How many BackupPC_nightly processes to run in parallel.

           Each night, at the first wakeup listed in $Conf{WakeupSchedule}, BackupPC_nightly is
           run.  Its job is to remove unneeded files in the pool, ie: files that only have one
           link.  To avoid race conditions, BackupPC_nightly and BackupPC_link cannot run at the
           same time.  Starting in v3.0.0, BackupPC_nightly can run concurrently with backups
           (BackupPC_dump).

           So to reduce the elapsed time, you might want to increase this setting to run several
           BackupPC_nightly processes in parallel (eg: 4, or even 8).

       $Conf{BackupPCNightlyPeriod} = 1;
           How many days (runs) it takes BackupPC_nightly to traverse the entire pool.  Normally
           this is 1, which means every night it runs, it does traverse the entire pool removing
           unused pool files.

           Other valid values are 2, 4, 8, 16.  This causes BackupPC_nightly to traverse 1/2,
           1/4, 1/8 or 1/16th of the pool each night, meaning it takes 2, 4, 8 or 16 days to
           completely traverse the pool.  The advantage is that each night the running time of
           BackupPC_nightly is reduced roughly in proportion, since the total job is split over
           multiple days.  The disadvantage is that unused pool files take longer to get deleted,
           which will slightly increase disk usage.

           Note that even when $Conf{BackupPCNightlyPeriod} > 1, BackupPC_nightly still runs
           every night.  It just does less work each time it runs.

           Examples:

              $Conf{BackupPCNightlyPeriod} = 1;   # entire pool is checked every night

              $Conf{BackupPCNightlyPeriod} = 2;   # two days to complete pool check
                                                  # (different half each night)

              $Conf{BackupPCNightlyPeriod} = 4;   # four days to complete pool check
                                                  # (different quarter each night)

       $Conf{MaxOldLogFiles} = 14;
           Maximum number of log files we keep around in log directory.  These files are aged
           nightly.  A setting of 14 means the log directory will contain about 2 weeks of old
           log files, in particular at most the files LOG, LOG.0, LOG.1, ... LOG.13 (except
           today's LOG, these files will have a .z extension if compression is on).

           If you decrease this number after BackupPC has been running for a while you will have
           to manually remove the older log files.

       $Conf{DfPath} = '';
           Full path to the df command.  Security caution: normal users should not allowed to
           write to this file or directory.

       $Conf{DfCmd} = '$dfPath $topDir';
           Command to run df.  The following variables are substituted at run-time:

             $dfPath      path to df ($Conf{DfPath})
             $topDir      top-level BackupPC data directory

           Note: all Cmds are executed directly without a shell, so the prog name needs to be a
           full path and you can't include shell syntax like redirection and pipes; put that in a
           script if you need it.

       $Conf{SplitPath} = '';
       $Conf{ParPath} = '';
       $Conf{CatPath} = '';
       $Conf{GzipPath} = '';
       $Conf{Bzip2Path} = '';
           Full path to various commands for archiving

       $Conf{DfMaxUsagePct} = 95;
           Maximum threshold for disk utilization on the __TOPDIR__ filesystem.  If the output
           from $Conf{DfPath} reports a percentage larger than this number then no new regularly
           scheduled backups will be run.  However, user requested backups (which are usually
           incremental and tend to be small) are still performed, independent of disk usage.
           Also, currently running backups will not be terminated when the disk usage exceeds
           this number.

       $Conf{TrashCleanSleepSec} = 300;
           How long BackupPC_trashClean sleeps in seconds between each check of the trash
           directory.  Once every 5 minutes should be reasonable.

       $Conf{DHCPAddressRanges} = [];
           List of DHCP address ranges we search looking for PCs to backup.  This is an array of
           hashes for each class C address range.  This is only needed if hosts in the conf/hosts
           file have the dhcp flag set.

           Examples:

              # to specify 192.10.10.20 to 192.10.10.250 as the DHCP address pool
              $Conf{DHCPAddressRanges} = [
                  {
                      ipAddrBase => '192.10.10',
                      first => 20,
                      last  => 250,
                  },
              ];
              # to specify two pools (192.10.10.20-250 and 192.10.11.10-50)
              $Conf{DHCPAddressRanges} = [
                  {
                      ipAddrBase => '192.10.10',
                      first => 20,
                      last  => 250,
                  },
                  {
                      ipAddrBase => '192.10.11',
                      first => 10,
                      last  => 50,
                  },
              ];

       $Conf{BackupPCUser} = '';
           The BackupPC user.

       $Conf{TopDir} = '';
       $Conf{ConfDir} = '';
       $Conf{LogDir} = '';
       $Conf{InstallDir} = '';
       $Conf{CgiDir} = '';
           Important installation directories:

             TopDir     - where all the backup data is stored
             ConfDir    - where the main config and hosts files resides
             LogDir     - where log files and other transient information
             InstallDir - where the bin, lib and doc installation dirs reside.
                          Note: you cannot change this value since all the
                          perl scripts include this path.  You must reinstall
                          with configure.pl to change InstallDir.
             CgiDir     - Apache CGI directory for BackupPC_Admin

           Note: it is STRONGLY recommended that you don't change the values here.  These are set
           at installation time and are here for reference and are used during upgrades.

           Instead of changing TopDir here it is recommended that you use a symbolic link to the
           new location, or mount the new BackupPC store at the existing $Conf{TopDir} setting.

       $Conf{BackupPCUserVerify} = 1;
           Whether BackupPC and the CGI script BackupPC_Admin verify that they are really running
           as user $Conf{BackupPCUser}.  If this flag is set and the effective user id (euid)
           differs from $Conf{BackupPCUser} then both scripts exit with an error.  This catches
           cases where BackupPC might be accidently started as root or the wrong user, or if the
           CGI script is not installed correctly.

       $Conf{HardLinkMax} = 31999;
           Maximum number of hardlinks supported by the $TopDir file system that BackupPC uses.
           Most linux or unix file systems should support at least 32000 hardlinks per file, or
           64000 in other cases.  If a pool file already has this number of hardlinks, a new pool
           file is created so that new hardlinks can be accommodated.  This limit will only be
           hit if an identical file appears at least this number of times across all the backups.

       $Conf{PerlModuleLoad} = undef;
           Advanced option for asking BackupPC to load additional perl modules.  Can be a list
           (array ref) of module names to load at startup.

       $Conf{ServerInitdPath} = '';
       $Conf{ServerInitdStartCmd} = '';
           Path to init.d script and command to use that script to start the server from the CGI
           interface.  The following variables are substituted at run-time:

             $sshPath           path to ssh ($Conf{SshPath})
             $serverHost        same as $Conf{ServerHost}
             $serverInitdPath   path to init.d script ($Conf{ServerInitdPath})

           Example:

           $Conf{ServerInitdPath}     = '/etc/init.d/backuppc'; $Conf{ServerInitdStartCmd} =
           '$sshPath -q -x -l root $serverHost'
                                      . ' $serverInitdPath start'
                                      . ' < /dev/null >& /dev/null';

           Note: all Cmds are executed directly without a shell, so the prog name needs to be a
           full path and you can't include shell syntax like redirection and pipes; put that in a
           script if you need it.

   What to backup and when to do it
       $Conf{FullPeriod} = 6.97;
           Minimum period in days between full backups. A full dump will only be done if at least
           this much time has elapsed since the last full dump, and at least $Conf{IncrPeriod}
           days has elapsed since the last successful dump.

           Typically this is set slightly less than an integer number of days. The time taken for
           the backup, plus the granularity of $Conf{WakeupSchedule} will make the actual backup
           interval a bit longer.

       $Conf{IncrPeriod} = 0.97;
           Minimum period in days between incremental backups (a user requested incremental
           backup will be done anytime on demand).

           Typically this is set slightly less than an integer number of days. The time taken for
           the backup, plus the granularity of $Conf{WakeupSchedule} will make the actual backup
           interval a bit longer.

       $Conf{FullKeepCnt} = 1;
           Number of full backups to keep.  Must be >= 1.

           In the steady state, each time a full backup completes successfully the oldest one is
           removed.  If this number is decreased, the extra old backups will be removed.

           If filling of incremental dumps is off the oldest backup always has to be a full (ie:
           filled) dump.  This might mean one or two extra full dumps are kept until the oldest
           incremental backups expire.

           Exponential backup expiry is also supported.  This allows you to specify:

             - num fulls to keep at intervals of 1 * $Conf{FullPeriod}, followed by
             - num fulls to keep at intervals of 2 * $Conf{FullPeriod},
             - num fulls to keep at intervals of 4 * $Conf{FullPeriod},
             - num fulls to keep at intervals of 8 * $Conf{FullPeriod},
             - num fulls to keep at intervals of 16 * $Conf{FullPeriod},

           and so on.  This works by deleting every other full as each expiry boundary is
           crossed.

           Exponential expiry is specified using an array for $Conf{FullKeepCnt}:

             $Conf{FullKeepCnt} = [4, 2, 3];

           Entry #n specifies how many fulls to keep at an interval of 2^n * $Conf{FullPeriod}
           (ie: 1, 2, 4, 8, 16, 32, ...).

           The example above specifies keeping 4 of the most recent full backups (1 week
           interval) two full backups at 2 week intervals, and 3 full backups at 4 week
           intervals, eg:

              full 0 19 weeks old   \
              full 1 15 weeks old    >---  3 backups at 4 * $Conf{FullPeriod}
              full 2 11 weeks old   /
              full 3  7 weeks old   \____  2 backups at 2 * $Conf{FullPeriod}
              full 4  5 weeks old   /
              full 5  3 weeks old   \
              full 6  2 weeks old    \___  4 backups at 1 * $Conf{FullPeriod}
              full 7  1 week old     /
              full 8  current       /

           On a given week the spacing might be less than shown as each backup ages through each
           expiry period.  For example, one week later, a new full is completed and the oldest is
           deleted, giving:

              full 0 16 weeks old   \
              full 1 12 weeks old    >---  3 backups at 4 * $Conf{FullPeriod}
              full 2  8 weeks old   /
              full 3  6 weeks old   \____  2 backups at 2 * $Conf{FullPeriod}
              full 4  4 weeks old   /
              full 5  3 weeks old   \
              full 6  2 weeks old    \___  4 backups at 1 * $Conf{FullPeriod}
              full 7  1 week old     /
              full 8  current       /

           You can specify 0 as a count (except in the first entry), and the array can be as long
           as you wish.  For example:

             $Conf{FullKeepCnt} = [4, 0, 4, 0, 0, 2];

           This will keep 10 full dumps, 4 most recent at 1 * $Conf{FullPeriod}, followed by 4 at
           an interval of 4 * $Conf{FullPeriod} (approx 1 month apart), and then 2 at an interval
           of 32 * $Conf{FullPeriod} (approx 7-8 months apart).

           Example: these two settings are equivalent and both keep just the four most recent
           full dumps:

              $Conf{FullKeepCnt} = 4;
              $Conf{FullKeepCnt} = [4];

       $Conf{FullKeepCntMin} = 1;
       $Conf{FullAgeMax} = 90;
           Very old full backups are removed after $Conf{FullAgeMax} days.  However, we keep at
           least $Conf{FullKeepCntMin} full backups no matter how old they are.

           Note that $Conf{FullAgeMax} will be increased to $Conf{FullKeepCnt} times
           $Conf{FullPeriod} if $Conf{FullKeepCnt} specifies enough full backups to exceed
           $Conf{FullAgeMax}.

       $Conf{IncrKeepCnt} = 6;
           Number of incremental backups to keep.  Must be >= 1.

           In the steady state, each time an incr backup completes successfully the oldest one is
           removed.  If this number is decreased, the extra old backups will be removed.

       $Conf{IncrKeepCntMin} = 1;
       $Conf{IncrAgeMax} = 30;
           Very old incremental backups are removed after $Conf{IncrAgeMax} days.  However, we
           keep at least $Conf{IncrKeepCntMin} incremental backups no matter how old they are.

       $Conf{IncrLevels} = [1];
           Level of each incremental.  "Level" follows the terminology of dump(1).  A full backup
           has level 0.  A new incremental of level N will backup all files that have changed
           since the most recent backup of a lower level.

           The entries of $Conf{IncrLevels} apply in order to each incremental after each full
           backup.  It wraps around until the next full backup.  For example, these two settings
           have the same effect:

                 $Conf{IncrLevels} = [1, 2, 3];
                 $Conf{IncrLevels} = [1, 2, 3, 1, 2, 3];

           This means the 1st and 4th incrementals (level 1) go all the way back to the full.
           The 2nd and 3rd (and 5th and 6th) backups just go back to the immediate preceeding
           incremental.

           Specifying a sequence of multi-level incrementals will usually mean more than
           $Conf{IncrKeepCnt} incrementals will need to be kept, since lower level incrementals
           are needed to merge a complete view of a backup.  For example, with

                 $Conf{FullPeriod}  = 7;
                 $Conf{IncrPeriod}  = 1;
                 $Conf{IncrKeepCnt} = 6;
                 $Conf{IncrLevels}  = [1, 2, 3, 4, 5, 6];

           there will be up to 11 incrementals in this case:

                 backup #0  (full, level 0, oldest)
                 backup #1  (incr, level 1)
                 backup #2  (incr, level 2)
                 backup #3  (incr, level 3)
                 backup #4  (incr, level 4)
                 backup #5  (incr, level 5)
                 backup #6  (incr, level 6)
                 backup #7  (full, level 0)
                 backup #8  (incr, level 1)
                 backup #9  (incr, level 2)
                 backup #10 (incr, level 3)
                 backup #11 (incr, level 4)
                 backup #12 (incr, level 5, newest)

           Backup #1 (the oldest level 1 incremental) can't be deleted since backups 2..6 depend
           on it.  Those 6 incrementals can't all be deleted since that would only leave 5
           (#8..12).  When the next incremental happens (level 6), the complete set of 6 older
           incrementals (#1..6) will be deleted, since that maintains the required number
           ($Conf{IncrKeepCnt}) of incrementals.  This situation is reduced if you set shorter
           chains of multi-level incrementals, eg:

                 $Conf{IncrLevels}  = [1, 2, 3];

           would only have up to 2 extra incremenals before all 3 are deleted.

           BackupPC as usual merges the full and the sequence of incrementals together so each
           incremental can be browsed and restored as though it is a complete backup.  If you
           specify a long chain of incrementals then more backups need to be merged when
           browsing, restoring, or getting the starting point for rsync backups.  In the example
           above (levels 1..6), browing backup #6 requires 7 different backups (#0..6) to be
           merged.

           Because of this merging and the additional incrementals that need to be kept, it is
           recommended that some level 1 incrementals be included in $Conf{IncrLevels}.

           Prior to version 3.0 incrementals were always level 1, meaning each incremental backed
           up all the files that changed since the last full.

       $Conf{BackupsDisable} = 0;
           Disable all full and incremental backups.  These settings are useful for a client that
           is no longer being backed up (eg: a retired machine), but you wish to keep the last
           backups available for browsing or restoring to other machines.

           There are three values for $Conf{BackupsDisable}:

             0    Backups are enabled.

             1    Don't do any regular backups on this client.  Manually
                  requested backups (via the CGI interface) will still occur.

             2    Don't do any backups on this client.  Manually requested
                  backups (via the CGI interface) will be ignored.

           In versions prior to 3.0 Backups were disabled by setting $Conf{FullPeriod} to -1 or
           -2.

       $Conf{PartialAgeMax} = 3;
           A failed full backup is saved as a partial backup.  The rsync XferMethod can take
           advantage of the partial full when the next backup is run. This parameter sets the age
           of the partial full in days: if the partial backup is older than this number of days,
           then rsync will ignore (not use) the partial full when the next backup is run.  If you
           set this to a negative value then no partials will be saved.  If you set this to 0,
           partials will be saved, but will not be used by the next backup.

           The default setting of 3 days means that a partial older than 3 days is ignored when
           the next full backup is done.

       $Conf{IncrFill} = 0;
           Whether incremental backups are filled.  "Filling" means that the most recent full (or
           filled) dump is merged into the new incremental dump using hardlinks.  This makes an
           incremental dump look like a full dump.  Prior to v1.03 all incremental backups were
           filled.  In v1.4.0 and later the default is off.

           BackupPC, and the cgi interface in particular, do the right thing on un-filled
           incremental backups.  It will correctly display the merged incremental backup with the
           most recent filled backup, giving the un-filled incremental backups a filled
           appearance.  That means it invisible to the user whether incremental dumps are filled
           or not.

           Filling backups takes a little extra disk space, and it does cost some extra disk
           activity for filling, and later removal.  Filling is no longer useful, since file
           mangling and compression doesn't make a filled backup very useful. It's likely the
           filling option will be removed from future versions: filling will be delegated to the
           display and extraction of backup data.

           If filling is off, BackupPC makes sure that the oldest backup is a full, otherwise the
           following incremental backups will be incomplete.  This might mean an extra full
           backup has to be kept until the following incremental backups expire.

           The default is off.  You can turn this on or off at any time without affecting
           existing backups.

       $Conf{RestoreInfoKeepCnt} = 10;
           Number of restore logs to keep.  BackupPC remembers information about each restore
           request.  This number per client will be kept around before the oldest ones are
           pruned.

           Note: files/dirs delivered via Zip or Tar downloads don't count as restores.  Only the
           first restore option (where the files and dirs are written to the host) count as
           restores that are logged.

       $Conf{ArchiveInfoKeepCnt} = 10;
           Number of archive logs to keep.  BackupPC remembers information about each archive
           request.  This number per archive client will be kept around before the oldest ones
           are pruned.

       $Conf{BackupFilesOnly} = undef;
           List of directories or files to backup.  If this is defined, only these directories or
           files will be backed up.

           For Smb, only one of $Conf{BackupFilesExclude} and $Conf{BackupFilesOnly} can be
           specified per share. If both are set for a particular share, then
           $Conf{BackupFilesOnly} takes precedence and $Conf{BackupFilesExclude} is ignored.

           This can be set to a string, an array of strings, or, in the case of multiple shares,
           a hash of strings or arrays.  A hash is used to give a list of directories or files to
           backup for each share (the share name is the key).  If this is set to just a string or
           array, and $Conf{SmbShareName} contains multiple share names, then the setting is
           assumed to apply all shares.

           If a hash is used, a special key "*" means it applies to all shares that don't have a
           specific entry.

           Examples:

              $Conf{BackupFilesOnly} = '/myFiles';
              $Conf{BackupFilesOnly} = ['/myFiles'];     # same as first example
              $Conf{BackupFilesOnly} = ['/myFiles', '/important'];
              $Conf{BackupFilesOnly} = {
                 'c' => ['/myFiles', '/important'],      # these are for 'c' share
                 'd' => ['/moreFiles', '/archive'],      # these are for 'd' share
              };
              $Conf{BackupFilesOnly} = {
                 'c' => ['/myFiles', '/important'],      # these are for 'c' share
                 '*' => ['/myFiles', '/important'],      # these are other shares
              };

       $Conf{BackupFilesExclude} = undef;
           List of directories or files to exclude from the backup.  For Smb, only one of
           $Conf{BackupFilesExclude} and $Conf{BackupFilesOnly} can be specified per share.  If
           both are set for a particular share, then $Conf{BackupFilesOnly} takes precedence and
           $Conf{BackupFilesExclude} is ignored.

           This can be set to a string, an array of strings, or, in the case of multiple shares,
           a hash of strings or arrays.  A hash is used to give a list of directories or files to
           exclude for each share (the share name is the key).  If this is set to just a string
           or array, and $Conf{SmbShareName} contains multiple share names, then the setting is
           assumed to apply to all shares.

           The exact behavior is determined by the underlying transport program, smbclient or
           tar.  For smbclient the exlclude file list is passed into the X option.  Simple shell
           wild-cards using "*" or "?" are allowed.

           For tar, if the exclude file contains a "/" it is assumed to be anchored at the start
           of the string.  Since all the tar paths start with "./", BackupPC prepends a "." if
           the exclude file starts with a "/".  Note that GNU tar version >= 1.13.7 is required
           for the exclude option to work correctly.  For linux or unix machines you should add
           "/proc" to $Conf{BackupFilesExclude} unless you have specified --one-file-system in
           $Conf{TarClientCmd} or --one-file-system in $Conf{RsyncArgs}.  Also, for tar, do not
           use a trailing "/" in the directory name: a trailing "/" causes the name to not match
           and the directory will not be excluded.

           Users report that for smbclient you should specify a directory followed by "/*", eg:
           "/proc/*", instead of just "/proc".

           FTP servers are traversed recursively so excluding directories will also exclude its
           contents.  You can use the wildcard characters "*" and "?" to define files for
           inclusion and exclusion.  Both attributes $Conf{BackupFilesOnly} and
           $Conf{BackupFilesExclude} can be defined for the same share.

           If a hash is used, a special key "*" means it applies to all shares that don't have a
           specific entry.

           Examples:

              $Conf{BackupFilesExclude} = '/temp';
              $Conf{BackupFilesExclude} = ['/temp'];     # same as first example
              $Conf{BackupFilesExclude} = ['/temp', '/winnt/tmp'];
              $Conf{BackupFilesExclude} = {
                 'c' => ['/temp', '/winnt/tmp'],         # these are for 'c' share
                 'd' => ['/junk', '/dont_back_this_up'], # these are for 'd' share
              };
              $Conf{BackupFilesExclude} = {
                 'c' => ['/temp', '/winnt/tmp'],         # these are for 'c' share
                 '*' => ['/junk', '/dont_back_this_up'], # these are for other shares
              };

       $Conf{BlackoutBadPingLimit} = 3;
       $Conf{BlackoutGoodCnt} = 7;
           PCs that are always or often on the network can be backed up after hours, to reduce
           PC, network and server load during working hours. For each PC a count of consecutive
           good pings is maintained. Once a PC has at least $Conf{BlackoutGoodCnt} consecutive
           good pings it is subject to "blackout" and not backed up during hours and days
           specified by $Conf{BlackoutPeriods}.

           To allow for periodic rebooting of a PC or other brief periods when a PC is not on the
           network, a number of consecutive bad pings is allowed before the good ping count is
           reset. This parameter is $Conf{BlackoutBadPingLimit}.

           Note that bad and good pings don't occur with the same interval. If a machine is
           always on the network, it will only be pinged roughly once every $Conf{IncrPeriod}
           (eg: once per day). So a setting for $Conf{BlackoutGoodCnt} of 7 means it will take
           around 7 days for a machine to be subject to blackout. On the other hand, if a ping is
           failed, it will be retried roughly every time BackupPC wakes up, eg, every one or two
           hours. So a setting for $Conf{BlackoutBadPingLimit} of 3 means that the PC will lose
           its blackout status after 3-6 hours of unavailability.

           To disable the blackout feature set $Conf{BlackoutGoodCnt} to a negative value.  A
           value of 0 will make all machines subject to blackout.  But if you don't want to do
           any backups during the day it would be easier to just set $Conf{WakeupSchedule} to a
           restricted schedule.

       $Conf{BlackoutPeriods} = [ ... ];
           One or more blackout periods can be specified.  If a client is subject to blackout
           then no regular (non-manual) backups will be started during any of these periods.
           hourBegin and hourEnd specify hours fro midnight and weekDays is a list of days of the
           week where 0 is Sunday, 1 is Monday etc.

           For example:

              $Conf{BlackoutPeriods} = [
                   {
                       hourBegin =>  7.0,
                       hourEnd   => 19.5,
                       weekDays  => [1, 2, 3, 4, 5],
                   },
              ];

           specifies one blackout period from 7:00am to 7:30pm local time on Mon-Fri.

           The blackout period can also span midnight by setting hourBegin > hourEnd, eg:

              $Conf{BlackoutPeriods} = [
                   {
                       hourBegin =>  7.0,
                       hourEnd   => 19.5,
                       weekDays  => [1, 2, 3, 4, 5],
                   },
                   {
                       hourBegin => 23,
                       hourEnd   =>  5,
                       weekDays  => [5, 6],
                   },
              ];

           This specifies one blackout period from 7:00am to 7:30pm local time on Mon-Fri, and a
           second period from 11pm to 5am on Friday and Saturday night.

       $Conf{BackupZeroFilesIsFatal} = 1;
           A backup of a share that has zero files is considered fatal. This is used to catch
           miscellaneous Xfer errors that result in no files being backed up.  If you have shares
           that might be empty (and therefore an empty backup is valid) you should set this flag
           to 0.

   How to backup a client
       $Conf{XferMethod} = 'smb';
           What transport method to use to backup each host.  If you have a mixed set of WinXX
           and linux/unix hosts you will need to override this in the per-PC config.pl.

           The valid values are:

             - 'smb':     backup and restore via smbclient and the SMB protocol.
                          Easiest choice for WinXX.

             - 'rsync':   backup and restore via rsync (via rsh or ssh).
                          Best choice for linux/unix.  Good choice also for WinXX.

             - 'rsyncd':  backup and restore via rsync daemon on the client.
                          Best choice for linux/unix if you have rsyncd running on
                          the client.  Good choice also for WinXX.

             - 'tar':    backup and restore via tar, tar over ssh, rsh or nfs.
                         Good choice for linux/unix.

             - 'archive': host is a special archive host.  Backups are not done.
                          An archive host is used to archive other host's backups
                          to permanent media, such as tape, CDR or DVD.

       $Conf{XferLogLevel} = 1;
           Level of verbosity in Xfer log files.  0 means be quiet, 1 will give will give one
           line per file, 2 will also show skipped files on incrementals, higher values give more
           output.

       $Conf{ClientCharset} = '';
           Filename charset encoding on the client.  BackupPC uses utf8 on the server for
           filename encoding.  If this is empty, then utf8 is assumed and client filenames will
           not be modified.  If set to a different encoding then filenames will converted to/from
           utf8 automatically during backup and restore.

           If the file names displayed in the browser (eg: accents or special characters) don't
           look right then it is likely you haven't set $Conf{ClientCharset} correctly.

           If you are using smbclient on a WinXX machine, smbclient will convert to the "unix
           charset" setting in smb.conf.  The default is utf8, in which case leave
           $Conf{ClientCharset} empty since smbclient does the right conversion.

           If you are using rsync on a WinXX machine then it does no conversion.  A typical WinXX
           encoding for latin1/western europe is 'cp1252', so in this case set
           $Conf{ClientCharset} to 'cp1252'.

           On a linux or unix client, run "locale charmap" to see the client's charset.  Set
           $Conf{ClientCharset} to this value.  A typical value for english/US is 'ISO-8859-1'.

           Do "perldoc Encode::Supported" to see the list of possible charset values.  The FAQ at
           http://www.cl.cam.ac.uk/~mgk25/unicode.html is excellent, and
           http://czyborra.com/charsets/iso8859.html provides more information on the iso-8859
           charsets.

       $Conf{ClientCharsetLegacy} = 'iso-8859-1';
           Prior to 3.x no charset conversion was done by BackupPC.  Backups were stored in what
           ever charset the XferMethod provided - typically utf8 for smbclient and the client's
           locale settings for rsync and tar (eg: cp1252 for rsync on WinXX and perhaps
           iso-8859-1 with rsync on linux).  This setting tells BackupPC the charset that was
           used to store file names in old backups taken with BackupPC 2.x, so that non-ascii
           file names in old backups can be viewed and restored.

   Samba Configuration
       $Conf{SmbShareName} = 'C$';
           Name of the host share that is backed up when using SMB.  This can be a string or an
           array of strings if there are multiple shares per host.  Examples:

             $Conf{SmbShareName} = 'c';          # backup 'c' share
             $Conf{SmbShareName} = ['c', 'd'];   # backup 'c' and 'd' shares

           This setting only matters if $Conf{XferMethod} = 'smb'.

       $Conf{SmbShareUserName} = '';
           Smbclient share user name.  This is passed to smbclient's -U argument.

           This setting only matters if $Conf{XferMethod} = 'smb'.

       $Conf{SmbSharePasswd} = '';
           Smbclient share password.  This is passed to smbclient via its PASSWD environment
           variable.  There are several ways you can tell BackupPC the smb share password.  In
           each case you should be very careful about security.  If you put the password here,
           make sure that this file is not readable by regular users!  See the "Setting up
           config.pl" section in the documentation for more information.

           This setting only matters if $Conf{XferMethod} = 'smb'.

       $Conf{SmbClientPath} = '';
           Full path for smbclient. Security caution: normal users should not allowed to write to
           this file or directory.

           smbclient is from the Samba distribution. smbclient is used to actually extract the
           incremental or full dump of the share filesystem from the PC.

           This setting only matters if $Conf{XferMethod} = 'smb'.

       $Conf{SmbClientFullCmd} = '$smbClientPath \\\\$host\\$shareName' ...
           Command to run smbclient for a full dump.  This setting only matters if
           $Conf{XferMethod} = 'smb'.

           The following variables are substituted at run-time:

              $smbClientPath   same as $Conf{SmbClientPath}
              $host            host to backup/restore
              $hostIP          host IP address
              $shareName       share name
              $userName        user name
              $fileList        list of files to backup (based on exclude/include)
              $I_option        optional -I option to smbclient
              $X_option        exclude option (if $fileList is an exclude list)
              $timeStampFile   start time for incremental dump

           Note: all Cmds are executed directly without a shell, so the prog name needs to be a
           full path and you can't include shell syntax like redirection and pipes; put that in a
           script if you need it.

       $Conf{SmbClientIncrCmd} = '$smbClientPath \\\\$host\\$shareName' ...
           Command to run smbclient for an incremental dump.  This setting only matters if
           $Conf{XferMethod} = 'smb'.

           Same variable substitutions are applied as $Conf{SmbClientFullCmd}.

           Note: all Cmds are executed directly without a shell, so the prog name needs to be a
           full path and you can't include shell syntax like redirection and pipes; put that in a
           script if you need it.

       $Conf{SmbClientRestoreCmd} = '$smbClientPath \\\\$host\\$shareName' ...
           Command to run smbclient for a restore.  This setting only matters if
           $Conf{XferMethod} = 'smb'.

           Same variable substitutions are applied as $Conf{SmbClientFullCmd}.

           If your smb share is read-only then direct restores will fail.  You should set
           $Conf{SmbClientRestoreCmd} to undef and the corresponding CGI restore option will be
           removed.

           Note: all Cmds are executed directly without a shell, so the prog name needs to be a
           full path and you can't include shell syntax like redirection and pipes; put that in a
           script if you need it.

   Tar Configuration
       $Conf{TarShareName} = '/';
           Which host directories to backup when using tar transport.  This can be a string or an
           array of strings if there are multiple directories to backup per host.  Examples:

             $Conf{TarShareName} = '/';                    # backup everything
             $Conf{TarShareName} = '/home';                # only backup /home
             $Conf{TarShareName} = ['/home', '/src'];      # backup /home and /src

           The fact this parameter is called 'TarShareName' is for historical consistency with
           the Smb transport options.  You can use any valid directory on the client: there is no
           need for it to correspond to any Smb share or device mount point.

           Note also that you can also use $Conf{BackupFilesOnly} to specify a specific list of
           directories to backup.  It's more efficient to use this option instead of
           $Conf{TarShareName} since a new tar is run for each entry in $Conf{TarShareName}.

           On the other hand, if you add --one-file-system to $Conf{TarClientCmd} you can backup
           each file system separately, which makes restoring one bad file system easier.  In
           this case you would list all of the mount points here, since you can't get the same
           result with $Conf{BackupFilesOnly}:

               $Conf{TarShareName} = ['/', '/var', '/data', '/boot'];

           This setting only matters if $Conf{XferMethod} = 'tar'.

       $Conf{TarClientCmd} = '$sshPath -q -x -n -l root $host' ...
           Command to run tar on the client.  GNU tar is required.  You will need to fill in the
           correct paths for ssh2 on the local host (server) and GNU tar on the client.  Security
           caution: normal users should not allowed to write to these executable files or
           directories.

           $Conf{TarClientCmd} is appended with with either $Conf{TarFullArgs} or
           $Conf{TarIncrArgs} to create the final command that is run.

           See the documentation for more information about setting up ssh2 keys.

           If you plan to use NFS then tar just runs locally and ssh2 is not needed.  For
           example, assuming the client filesystem is mounted below /mnt/hostName, you could use
           something like:

              $Conf{TarClientCmd} = '$tarPath -c -v -f - -C /mnt/$host/$shareName'
                                  . ' --totals';

           In the case of NFS or rsh you need to make sure BackupPC's privileges are sufficient
           to read all the files you want to backup.  Also, you will probably want to add "/proc"
           to $Conf{BackupFilesExclude}.

           The following variables are substituted at run-time:

             $host        host name
             $hostIP      host's IP address
             $incrDate    newer-than date for incremental backups
             $shareName   share name to backup (ie: top-level directory path)
             $fileList    specific files to backup or exclude
             $tarPath     same as $Conf{TarClientPath}
             $sshPath     same as $Conf{SshPath}

           If a variable is followed by a "+" it is shell escaped.  This is necessary for the
           command part of ssh or rsh, since it ends up getting passed through the shell.

           This setting only matters if $Conf{XferMethod} = 'tar'.

           Note: all Cmds are executed directly without a shell, so the prog name needs to be a
           full path and you can't include shell syntax like redirection and pipes; put that in a
           script if you need it.

       $Conf{TarFullArgs} = '$fileList+';
           Extra tar arguments for full backups.  Several variables are substituted at run-time.
           See $Conf{TarClientCmd} for the list of variable substitutions.

           If you are running tar locally (ie: without rsh or ssh) then remove the "+" so that
           the argument is no longer shell escaped.

           This setting only matters if $Conf{XferMethod} = 'tar'.

       $Conf{TarIncrArgs} = '--newer=$incrDate+ $fileList+';
           Extra tar arguments for incr backups.  Several variables are substituted at run-time.
           See $Conf{TarClientCmd} for the list of variable substitutions.

           Note that GNU tar has several methods for specifying incremental backups, including:

             --newer-mtime $incrDate+
                    This causes a file to be included if the modification time is
                    later than $incrDate (meaning its contents might have changed).
                    But changes in the ownership or modes will not qualify the
                    file to be included in an incremental.

             --newer=$incrDate+
                    This causes the file to be included if any attribute of the
                    file is later than $incrDate, meaning either attributes or
                    the modification time.  This is the default method.  Do
                    not use --atime-preserve in $Conf{TarClientCmd} above,
                    otherwise resetting the atime (access time) counts as an
                    attribute change, meaning the file will always be included
                    in each new incremental dump.

           If you are running tar locally (ie: without rsh or ssh) then remove the "+" so that
           the argument is no longer shell escaped.

           This setting only matters if $Conf{XferMethod} = 'tar'.

       $Conf{TarClientRestoreCmd} = '$sshPath -q -x -l root $host' ...
           Full command to run tar for restore on the client.  GNU tar is required.  This can be
           the same as $Conf{TarClientCmd}, with tar's -c replaced by -x and ssh's -n removed.

           See $Conf{TarClientCmd} for full details.

           This setting only matters if $Conf{XferMethod} = "tar".

           If you want to disable direct restores using tar, you should set
           $Conf{TarClientRestoreCmd} to undef and the corresponding CGI restore option will be
           removed.

           Note: all Cmds are executed directly without a shell, so the prog name needs to be a
           full path and you can't include shell syntax like redirection and pipes; put that in a
           script if you need it.

       $Conf{TarClientPath} = '';
           Full path for tar on the client. Security caution: normal users should not allowed to
           write to this file or directory.

           This setting only matters if $Conf{XferMethod} = 'tar'.

   Rsync/Rsyncd Configuration
       $Conf{RsyncClientPath} = '';
           Path to rsync executable on the client

       $Conf{RsyncClientCmd} = '$sshPath -q -x -l root $host $rsyncPath $argList+';
           Full command to run rsync on the client machine.  The following variables are
           substituted at run-time:

                  $host           host name being backed up
                  $hostIP         host's IP address
                  $shareName      share name to backup (ie: top-level directory path)
                  $rsyncPath      same as $Conf{RsyncClientPath}
                  $sshPath        same as $Conf{SshPath}
                  $argList        argument list, built from $Conf{RsyncArgs},
                                  $shareName, $Conf{BackupFilesExclude} and
                                  $Conf{BackupFilesOnly}

           This setting only matters if $Conf{XferMethod} = 'rsync'.

       $Conf{RsyncClientRestoreCmd} = '$sshPath -q -x -l root $host $rsyncPath $argList+';
           Full command to run rsync for restore on the client.  The following variables are
           substituted at run-time:

                  $host           host name being backed up
                  $hostIP         host's IP address
                  $shareName      share name to backup (ie: top-level directory path)
                  $rsyncPath      same as $Conf{RsyncClientPath}
                  $sshPath        same as $Conf{SshPath}
                  $argList        argument list, built from $Conf{RsyncArgs},
                                  $shareName, $Conf{BackupFilesExclude} and
                                  $Conf{BackupFilesOnly}

           This setting only matters if $Conf{XferMethod} = 'rsync'.

           Note: all Cmds are executed directly without a shell, so the prog name needs to be a
           full path and you can't include shell syntax like redirection and pipes; put that in a
           script if you need it.

       $Conf{RsyncShareName} = '/';
           Share name to backup.  For $Conf{XferMethod} = "rsync" this should be a file system
           path, eg '/' or '/home'.

           For $Conf{XferMethod} = "rsyncd" this should be the name of the module to backup (ie:
           the name from /etc/rsynd.conf).

           This can also be a list of multiple file system paths or modules.  For example, by
           adding --one-file-system to $Conf{RsyncArgs} you can backup each file system
           separately, which makes restoring one bad file system easier.  In this case you would
           list all of the mount points:

               $Conf{RsyncShareName} = ['/', '/var', '/data', '/boot'];

       $Conf{RsyncdClientPort} = 873;
           Rsync daemon port on the client, for $Conf{XferMethod} = "rsyncd".

       $Conf{RsyncdUserName} = '';
           Rsync daemon user name on client, for $Conf{XferMethod} = "rsyncd".  The user name and
           password are stored on the client in whatever file the "secrets file" parameter in
           rsyncd.conf points to (eg: /etc/rsyncd.secrets).

       $Conf{RsyncdPasswd} = '';
           Rsync daemon user name on client, for $Conf{XferMethod} = "rsyncd".  The user name and
           password are stored on the client in whatever file the "secrets file" parameter in
           rsyncd.conf points to (eg: /etc/rsyncd.secrets).

       $Conf{RsyncdAuthRequired} = 1;
           Whether authentication is mandatory when connecting to the client's rsyncd.  By
           default this is on, ensuring that BackupPC will refuse to connect to an rsyncd on the
           client that is not password protected.  Turn off at your own risk.

       $Conf{RsyncCsumCacheVerifyProb} = 0.01;
           When rsync checksum caching is enabled (by adding the --checksum-seed=32761 option to
           $Conf{RsyncArgs}), the cached checksums can be occasionally verified to make sure the
           file contents matches the cached checksums.  This is to avoid the risk that disk
           problems might cause the pool file contents to get corrupted, but the cached checksums
           would make BackupPC think that the file still matches the client.

           This setting is the probability (0 means never and 1 means always) that a file will be
           rechecked.  Setting it to 0 means the checksums will not be rechecked (unless there is
           a phase 0 failure).  Setting it to 1 (ie: 100%) means all files will be checked, but
           that is not a desirable setting since you are better off simply turning caching off
           (ie: remove the --checksum-seed option).

           The default of 0.01 means 1% (on average) of the files during a full backup will have
           their cached checksum re-checked.

           This setting has no effect unless checksum caching is turned on.

       $Conf{RsyncArgs} = [ ... ];
           Arguments to rsync for backup.  Do not edit the first set unless you have a thorough
           understanding of how File::RsyncP works.

       $Conf{RsyncArgsExtra} = [];
           Additional arguments added to RsyncArgs.  This can be used in conbination with
           $Conf{RsyncArgs} to allow customization of the rsync arguments on a part-client basis.
           The standard arguments go in $Conf{RsyncArgs} and $Conf{RsyncArgsExtra} can be set on
           a per-client basis.

           Examples of additional arguments that should work are --exclude/--include, eg:

               $Conf{RsyncArgsExtra} = [
                     '--exclude', '/proc',
                     '--exclude', '*.tmp',
               ];

           Both $Conf{RsyncArgs} and $Conf{RsyncArgsExtra} are subject to the following variable
           substitutions:

                  $client       client name being backed up
                  $host         host name (could be different from client name if
                                           $Conf{ClientNameAlias} is set)
                  $hostIP       IP address of host
                  $confDir      configuration directory path

           This allows settings of the form:

               $Conf{RsyncArgsExtra} = [
                       '--exclude-from=$confDir/pc/$host.exclude',
               ];

       $Conf{RsyncRestoreArgs} = [ ... ];
           Arguments to rsync for restore.  Do not edit the first set unless you have a thorough
           understanding of how File::RsyncP works.

           If you want to disable direct restores using rsync (eg: is the module is read-only),
           you should set $Conf{RsyncRestoreArgs} to undef and the corresponding CGI restore
           option will be removed.

           $Conf{RsyncRestoreArgs} is subject to the following variable substitutions:

                  $client       client name being backed up
                  $host         host name (could be different from client name if
                                           $Conf{ClientNameAlias} is set)
                  $hostIP       IP address of host
                  $confDir      configuration directory path

           Note: $Conf{RsyncArgsExtra} doesn't apply to $Conf{RsyncRestoreArgs}.

   FTP Configuration
       $Conf{FtpShareName} = '';
           Which host directories to backup when using FTP.  This can be a string or an array of
           strings if there are multiple shares per host.

           This value must be specified in one of two ways: either as a subdirectory of the
           'share root' on the server, or as the absolute path of the directory.

           In the following example, if the directory /home/username is the root share of the ftp
           server with the given username, the following two values will back up the same
           directory:

              $Conf{FtpShareName} = 'www';                # www directory
              $Conf{FtpShareName} = '/home/username/www'; # same directory

           Path resolution is not supported; i.e.; you may not have an ftp share path defined as
           '../otheruser' or '~/games'.

            Multiple shares may also be specified, as with other protocols:

              $Conf{FtpShareName} = [ 'www',
                                      'bin',
                                      'config' ];

           Note also that you can also use $Conf{BackupFilesOnly} to specify a specific list of
           directories to backup.  It's more efficient to use this option instead of
           $Conf{FtpShareName} since a new tar is run for each entry in $Conf{FtpShareName}.

           This setting only matters if $Conf{XferMethod} = 'ftp'.

       $Conf{FtpUserName} = '';
           FTP user name.  This is used to log into the server.

           This setting is used only if $Conf{XferMethod} = 'ftp'.

       $Conf{FtpPasswd} = '';
           FTP user password.  This is used to log into the server.

           This setting is used only if $Conf{XferMethod} = 'ftp'.

       $Conf{FtpPassive} = 1;
           Whether passive mode is used.  The correct setting depends upon whether local or
           remote ports are accessible from the other machine, which is affected by any firewall
           or routers between the FTP server on the client and the BackupPC server.

           This setting is used only if $Conf{XferMethod} = 'ftp'.

       $Conf{FtpBlockSize} = 10240;
           Transfer block size. This sets the size of the amounts of data in each frame. While
           undefined, this value takes the default value.

           This setting is used only if $Conf{XferMethod} = 'ftp'.

       $Conf{FtpPort} = 21;
           The port of the ftp server.  If undefined, 21 is used.

           This setting is used only if $Conf{XferMethod} = 'ftp'.

       $Conf{FtpTimeout} = 120;
           Connection timeout for FTP.  When undefined, the default is 120 seconds.

           This setting is used only if $Conf{XferMethod} = 'ftp'.

       $Conf{FtpFollowSymlinks} = 0;
           Behaviour when BackupPC encounters symlinks on the FTP share.

           Symlinks cannot be restored via FTP, so the desired behaviour will be different
           depending on the setup of the share. The default for this behavor is 1.  Directory
           shares with more complicated directory structures should consider other protocols.

   Archive Configuration
       $Conf{ArchiveDest} = '/tmp';
           Archive Destination

           The Destination of the archive e.g. /tmp for file archive or /dev/nst0 for device
           archive

       $Conf{ArchiveComp} = 'gzip';
           Archive Compression type

           The valid values are:

             - 'none':  No Compression

             - 'gzip':  Medium Compression. Recommended.

             - 'bzip2': High Compression but takes longer.

       $Conf{ArchivePar} = 0;
           Archive Parity Files

           The amount of Parity data to generate, as a percentage of the archive size.  Uses the
           commandline par2 (par2cmdline) available from http://parchive.sourceforge.net

           Only useful for file dumps.

           Set to 0 to disable this feature.

       $Conf{ArchiveSplit} = 0;
           Archive Size Split

           Only for file archives. Splits the output into the specified size * 1,000,000.  e.g.
           to split into 650,000,000 bytes, specify 650 below.

           If the value is 0, or if $Conf{ArchiveDest} is an existing file or device (e.g. a
           streaming tape drive), this feature is disabled.

       $Conf{ArchiveClientCmd} = '$Installdir/bin/BackupPC_archiveHost' ...
           Archive Command

           This is the command that is called to actually run the archive process for each host.
           The following variables are substituted at run-time:

             $Installdir    The installation directory of BackupPC
             $tarCreatePath The path to BackupPC_tarCreate
             $splitpath     The path to the split program
             $parpath       The path to the par2 program
             $host          The host to archive
             $backupnumber  The backup number of the host to archive
             $compression   The path to the compression program
             $compext       The extension assigned to the compression type
             $splitsize     The number of bytes to split archives into
             $archiveloc    The location to put the archive
             $parfile       The amount of parity data to create (percentage)

           Note: all Cmds are executed directly without a shell, so the prog name needs to be a
           full path and you can't include shell syntax like redirection and pipes; put that in a
           script if you need it.

       $Conf{SshPath} = '';
           Full path for ssh. Security caution: normal users should not allowed to write to this
           file or directory.

       $Conf{NmbLookupPath} = '';
           Full path for nmblookup. Security caution: normal users should not allowed to write to
           this file or directory.

           nmblookup is from the Samba distribution. nmblookup is used to get the netbios name,
           necessary for DHCP hosts.

       $Conf{NmbLookupCmd} = '$nmbLookupPath -A $host';
           NmbLookup command.  Given an IP address, does an nmblookup on that IP address.  The
           following variables are substituted at run-time:

             $nmbLookupPath      path to nmblookup ($Conf{NmbLookupPath})
             $host               IP address

           This command is only used for DHCP hosts: given an IP address, this command should try
           to find its NetBios name.

           Note: all Cmds are executed directly without a shell, so the prog name needs to be a
           full path and you can't include shell syntax like redirection and pipes; put that in a
           script if you need it.

       $Conf{NmbLookupFindHostCmd} = '$nmbLookupPath $host';
           NmbLookup command.  Given a netbios name, finds that host by doing a NetBios lookup.
           Several variables are substituted at run-time:

             $nmbLookupPath      path to nmblookup ($Conf{NmbLookupPath})
             $host               NetBios name

           In some cases you might need to change the broadcast address, for example if nmblookup
           uses 192.168.255.255 by default and you find that doesn't work, try 192.168.1.255 (or
           your equivalent class C address) using the -B option:

              $Conf{NmbLookupFindHostCmd} = '$nmbLookupPath -B 192.168.1.255 $host';

           If you use a WINS server and your machines don't respond to multicast NetBios requests
           you can use this (replace 1.2.3.4 with the IP address of your WINS server):

              $Conf{NmbLookupFindHostCmd} = '$nmbLookupPath -R -U 1.2.3.4 $host';

           This is preferred over multicast since it minimizes network traffic.

           Experiment manually for your site to see what form of nmblookup command works.

           Note: all Cmds are executed directly without a shell, so the prog name needs to be a
           full path and you can't include shell syntax like redirection and pipes; put that in a
           script if you need it.

       $Conf{FixedIPNetBiosNameCheck} = 0;
           For fixed IP address hosts, BackupPC_dump can also verify the netbios name to ensure
           it matches the host name.  An error is generated if they do not match.  Typically this
           flag is off.  But if you are going to transition a bunch of machines from fixed host
           addresses to DHCP, setting this flag is a great way to verify that the machines have
           their netbios name set correctly before turning on DCHP.

       $Conf{PingPath} = '';
           Full path to the ping command.  Security caution: normal users should not be allowed
           to write to this file or directory.

           If you want to disable ping checking, set this to some program that exits with 0
           status, eg:

               $Conf{PingPath} = '/bin/echo';

       $Conf{PingCmd} = '$pingPath -c 1 $host';
           Ping command.  The following variables are substituted at run-time:

             $pingPath      path to ping ($Conf{PingPath})
             $host          host name

           Wade Brown reports that on solaris 2.6 and 2.7 ping -s returns the wrong exit status
           (0 even on failure).  Replace with "ping $host 1", which gets the correct exit status
           but we don't get the round-trip time.

           Note: all Cmds are executed directly without a shell, so the prog name needs to be a
           full path and you can't include shell syntax like redirection and pipes; put that in a
           script if you need it.

       $Conf{PingMaxMsec} = 20;
           Maximum round-trip ping time in milliseconds.  This threshold is set to avoid backing
           up PCs that are remotely connected through WAN or dialup connections.  The output from
           ping -s (assuming it is supported on your system) is used to check the round-trip
           packet time.  On your local LAN round-trip times should be much less than 20msec.  On
           most WAN or dialup connections the round-trip time will be typically more than 20msec.
           Tune if necessary.

       $Conf{CompressLevel} = 0;
           Compression level to use on files.  0 means no compression.  Compression levels can be
           from 1 (least cpu time, slightly worse compression) to 9 (most cpu time, slightly
           better compression).  The recommended value is 3.  Changing to 5, for example, will
           take maybe 20% more cpu time and will get another 2-3% additional compression. See the
           zlib documentation for more information about compression levels.

           Changing compression on or off after backups have already been done will require both
           compressed and uncompressed pool files to be stored.  This will increase the pool
           storage requirements, at least until all the old backups expire and are deleted.

           It is ok to change the compression value (from one non-zero value to another non-zero
           value) after dumps are already done.  Since BackupPC matches pool files by comparing
           the uncompressed versions, it will still correctly match new incoming files against
           existing pool files.  The new compression level will take effect only for new files
           that are newly compressed and added to the pool.

           If compression was off and you are enabling compression for the first time you can use
           the BackupPC_compressPool utility to compress the pool.  This avoids having the pool
           grow to accommodate both compressed and uncompressed backups.  See the documentation
           for more information.

           Note: compression needs the Compress::Zlib perl library.  If the Compress::Zlib
           library can't be found then $Conf{CompressLevel} is forced to 0 (compression off).

       $Conf{ClientTimeout} = 72000;
           Timeout in seconds when listening for the transport program's (smbclient, tar etc)
           stdout. If no output is received during this time, then it is assumed that something
           has wedged during a backup, and the backup is terminated.

           Note that stdout buffering combined with huge files being backed up could cause
           longish delays in the output from smbclient that BackupPC_dump sees, so in rare cases
           you might want to increase this value.

           Despite the name, this parameter sets the timeout for all transport methods (tar, smb
           etc).

       $Conf{MaxOldPerPCLogFiles} = 12;
           Maximum number of log files we keep around in each PC's directory (ie: pc/$host).
           These files are aged monthly.  A setting of 12 means there will be at most the files
           LOG, LOG.0, LOG.1, ... LOG.11 in the pc/$host directory (ie: about a years worth).
           (Except this month's LOG, these files will have a .z extension if compression is on).

           If you decrease this number after BackupPC has been running for a while you will have
           to manually remove the older log files.

       $Conf{DumpPreUserCmd} = undef;
       $Conf{DumpPostUserCmd} = undef;
       $Conf{DumpPreShareCmd} = undef;
       $Conf{DumpPostShareCmd} = undef;
       $Conf{RestorePreUserCmd} = undef;
       $Conf{RestorePostUserCmd} = undef;
       $Conf{ArchivePreUserCmd} = undef;
       $Conf{ArchivePostUserCmd} = undef;
           Optional commands to run before and after dumps and restores, and also before and
           after each share of a dump.

           Stdout from these commands will be written to the Xfer (or Restore) log file.  One
           example of using these commands would be to shut down and restart a database server,
           dump a database to files for backup, or doing a snapshot of a share prior to a backup.
           Example:

              $Conf{DumpPreUserCmd} = '$sshPath -q -x -l root $host /usr/bin/dumpMysql';

           The following variable substitutions are made at run time for $Conf{DumpPreUserCmd},
           $Conf{DumpPostUserCmd}, $Conf{DumpPreShareCmd} and $Conf{DumpPostShareCmd}:

                  $type         type of dump (incr or full)
                  $xferOK       1 if the dump succeeded, 0 if it didn't
                  $client       client name being backed up
                  $host         host name (could be different from client name if
                                           $Conf{ClientNameAlias} is set)
                  $hostIP       IP address of host
                  $user         user name from the hosts file
                  $moreUsers    list of additional users from the hosts file
                  $share        the first share name (or current share for
                                  $Conf{DumpPreShareCmd} and $Conf{DumpPostShareCmd})
                  $shares       list of all the share names
                  $XferMethod   value of $Conf{XferMethod} (eg: tar, rsync, smb)
                  $sshPath      value of $Conf{SshPath},
                  $cmdType      set to DumpPreUserCmd or DumpPostUserCmd

           The following variable substitutions are made at run time for $Conf{RestorePreUserCmd}
           and $Conf{RestorePostUserCmd}:

                  $client       client name being backed up
                  $xferOK       1 if the restore succeeded, 0 if it didn't
                  $host         host name (could be different from client name if
                                           $Conf{ClientNameAlias} is set)
                  $hostIP       IP address of host
                  $user         user name from the hosts file
                  $moreUsers    list of additional users from the hosts file
                  $share        the first share name
                  $XferMethod   value of $Conf{XferMethod} (eg: tar, rsync, smb)
                  $sshPath      value of $Conf{SshPath},
                  $type         set to "restore"
                  $bkupSrcHost  host name of the restore source
                  $bkupSrcShare share name of the restore source
                  $bkupSrcNum   backup number of the restore source
                  $pathHdrSrc   common starting path of restore source
                  $pathHdrDest  common starting path of destination
                  $fileList     list of files being restored
                  $cmdType      set to RestorePreUserCmd or RestorePostUserCmd

           The following variable substitutions are made at run time for $Conf{ArchivePreUserCmd}
           and $Conf{ArchivePostUserCmd}:

                  $client       client name being backed up
                  $xferOK       1 if the archive succeeded, 0 if it didn't
                  $host         Name of the archive host
                  $user         user name from the hosts file
                  $share        the first share name
                  $XferMethod   value of $Conf{XferMethod} (eg: tar, rsync, smb)
                  $HostList     list of hosts being archived
                  $BackupList   list of backup numbers for the hosts being archived
                  $archiveloc   location where the archive is sent to
                  $parfile      amount of parity data being generated (percentage)
                  $compression  compression program being used (eg: cat, gzip, bzip2)
                  $compext      extension used for compression type (eg: raw, gz, bz2)
                  $splitsize    size of the files that the archive creates
                  $sshPath      value of $Conf{SshPath},
                  $type         set to "archive"
                  $cmdType      set to ArchivePreUserCmd or ArchivePostUserCmd

           Note: all Cmds are executed directly without a shell, so the prog name needs to be a
           full path and you can't include shell syntax like redirection and pipes; put that in a
           script if you need it.

       $Conf{UserCmdCheckStatus} = 0;
           Whether the exit status of each PreUserCmd and PostUserCmd is checked.

           If set and the Dump/Restore/Archive Pre/Post UserCmd returns a non-zero exit status
           then the dump/restore/archive is aborted.  To maintain backward compatibility (where
           the exit status in early versions was always ignored), this flag defaults to 0.

           If this flag is set and the Dump/Restore/Archive PreUserCmd fails then the matching
           Dump/Restore/Archive PostUserCmd is not executed.  If DumpPreShareCmd returns a non-
           exit status, then DumpPostShareCmd is not executed, but the DumpPostUserCmd is still
           run (since DumpPreUserCmd must have previously succeeded).

           An example of a DumpPreUserCmd that might fail is a script that snapshots or dumps a
           database which fails because of some database error.

       $Conf{ClientNameAlias} = undef;
           Override the client's host name.  This allows multiple clients to all refer to the
           same physical host.  This should only be set in the per-PC config file and is only
           used by BackupPC at the last moment prior to generating the command used to backup
           that machine (ie: the value of $Conf{ClientNameAlias} is invisible everywhere else in
           BackupPC).  The setting can be a host name or IP address, eg:

                   $Conf{ClientNameAlias} = 'realHostName';
                   $Conf{ClientNameAlias} = '192.1.1.15';

           will cause the relevant smb/tar/rsync backup/restore commands to be directed to
           realHostName, not the client name.

           Note: this setting doesn't work for hosts with DHCP set to 1.

   Email reminders, status and messages
       $Conf{SendmailPath} = '';
           Full path to the sendmail command.  Security caution: normal users should not allowed
           to write to this file or directory.

       $Conf{EMailNotifyMinDays} = 2.5;
           Minimum period between consecutive emails to a single user.  This tries to keep
           annoying email to users to a reasonable level.  Email checks are done nightly, so this
           number is effectively rounded up (ie: 2.5 means a user will never receive email more
           than once every 3 days).

       $Conf{EMailFromUserName} = '';
           Name to use as the "from" name for email.  Depending upon your mail handler this is
           either a plain name (eg: "admin") or a fully-qualified name (eg:
           "admin@mydomain.com").

       $Conf{EMailAdminUserName} = '';
           Destination address to an administrative user who will receive a nightly email with
           warnings and errors.  If there are no warnings or errors then no email will be sent.
           Depending upon your mail handler this is either a plain name (eg: "admin") or a fully-
           qualified name (eg: "admin@mydomain.com").

       $Conf{EMailUserDestDomain} = '';
           Destination domain name for email sent to users.  By default this is empty, meaning
           email is sent to plain, unqualified addresses.  Otherwise, set it to the destintation
           domain, eg:

              $Cong{EMailUserDestDomain} = '@mydomain.com';

           With this setting user email will be set to 'user@mydomain.com'.

       $Conf{EMailNoBackupEverSubj} = undef;
       $Conf{EMailNoBackupEverMesg} = undef;
           This subject and message is sent to a user if their PC has never been backed up.

           These values are language-dependent.  The default versions can be found in the
           language file (eg: lib/BackupPC/Lang/en.pm).  If you need to change the message, copy
           it here and edit it, eg:

             $Conf{EMailNoBackupEverMesg} = <<'EOF';
             To: $user$domain
             cc:
             Subject: $subj

             Dear $userName,

             This is a site-specific email message.
             EOF

       $Conf{EMailNotifyOldBackupDays} = 7.0;
           How old the most recent backup has to be before notifying user.  When there have been
           no backups in this number of days the user is sent an email.

       $Conf{EMailNoBackupRecentSubj} = undef;
       $Conf{EMailNoBackupRecentMesg} = undef;
           This subject and message is sent to a user if their PC has not recently been backed up
           (ie: more than $Conf{EMailNotifyOldBackupDays} days ago).

           These values are language-dependent.  The default versions can be found in the
           language file (eg: lib/BackupPC/Lang/en.pm).  If you need to change the message, copy
           it here and edit it, eg:

             $Conf{EMailNoBackupRecentMesg} = <<'EOF';
             To: $user$domain
             cc:
             Subject: $subj

             Dear $userName,

             This is a site-specific email message.
             EOF

       $Conf{EMailNotifyOldOutlookDays} = 5.0;
           How old the most recent backup of Outlook files has to be before notifying user.

       $Conf{EMailOutlookBackupSubj} = undef;
       $Conf{EMailOutlookBackupMesg} = undef;
           This subject and message is sent to a user if their Outlook files have not recently
           been backed up (ie: more than $Conf{EMailNotifyOldOutlookDays} days ago).

           These values are language-dependent.  The default versions can be found in the
           language file (eg: lib/BackupPC/Lang/en.pm).  If you need to change the message, copy
           it here and edit it, eg:

             $Conf{EMailOutlookBackupMesg} = <<'EOF';
             To: $user$domain
             cc:
             Subject: $subj

             Dear $userName,

             This is a site-specific email message.
             EOF

       $Conf{EMailHeaders} = <<EOF;
           Additional email headers.  This sets to charset to utf8.

   CGI user interface configuration settings
       $Conf{CgiAdminUserGroup} = '';
       $Conf{CgiAdminUsers} = '';
           Normal users can only access information specific to their host.  They can
           start/stop/browse/restore backups.

           Administrative users have full access to all hosts, plus overall status and log
           information.

           The administrative users are the union of the unix/linux group
           $Conf{CgiAdminUserGroup} and the manual list of users, separated by spaces, in
           $Conf{CgiAdminUsers}. If you don't want a group or manual list of users set the
           corresponding configuration setting to undef or an empty string.

           If you want every user to have admin privileges (careful!), set $Conf{CgiAdminUsers} =
           '*'.

           Examples:

              $Conf{CgiAdminUserGroup} = 'admin';
              $Conf{CgiAdminUsers}     = 'craig celia';
              --> administrative users are the union of group admin, plus
                craig and celia.

              $Conf{CgiAdminUserGroup} = '';
              $Conf{CgiAdminUsers}     = 'craig celia';
              --> administrative users are only craig and celia'.

       $Conf{CgiURL} = undef;
           URL of the BackupPC_Admin CGI script.  Used for email messages.

       $Conf{Language} = 'en';
           Language to use.  See lib/BackupPC/Lang for the list of supported languages, which
           include English (en), French (fr), Spanish (es), German (de), Italian (it), Dutch
           (nl), Polish (pl), Portuguese Brazillian (pt_br) and Chinese (zh_CH).

           Currently the Language setting applies to the CGI interface and email messages sent to
           users.  Log files and other text are still in English.

       $Conf{CgiUserHomePageCheck} = '';
       $Conf{CgiUserUrlCreate} = 'mailto:%s';
           User names that are rendered by the CGI interface can be turned into links into their
           home page or other information about the user.  To set this up you need to create two
           sprintf() strings, that each contain a single '%s' that will be replaced by the user
           name.  The default is a mailto: link.

           $Conf{CgiUserHomePageCheck} should be an absolute file path that is used to check (via
           "-f") that the user has a valid home page.  Set this to undef or an empty string to
           turn off this check.

           $Conf{CgiUserUrlCreate} should be a full URL that points to the user's home page.  Set
           this to undef or an empty string to turn off generation of URLs for user names.

           Example:

              $Conf{CgiUserHomePageCheck} = '/var/www/html/users/%s.html';
              $Conf{CgiUserUrlCreate}     = 'http://myhost/users/%s.html';
              --> if /var/www/html/users/craig.html exists, then 'craig' will
                be rendered as a link to http://myhost/users/craig.html.

       $Conf{CgiDateFormatMMDD} = 1;
           Date display format for CGI interface.  A value of 1 uses US-style dates (MM/DD), a
           value of 2 uses full YYYY-MM-DD format, and zero for international dates (DD/MM).

       $Conf{CgiNavBarAdminAllHosts} = 1;
           If set, the complete list of hosts appears in the left navigation bar pull-down for
           administrators.  Otherwise, just the hosts for which the user is listed in the host
           file (as either the user or in moreUsers) are displayed.

       $Conf{CgiSearchBoxEnable} = 1;
           Enable/disable the search box in the navigation bar.

       $Conf{CgiNavBarLinks} = [ ... ];
           Additional navigation bar links.  These appear for both regular users and
           administrators.  This is a list of hashes giving the link (URL) and the text (name)
           for the link.  Specifying lname instead of name uses the language specific string (ie:
           $Lang->{lname}) instead of just literally displaying name.

       $Conf{CgiStatusHilightColor} = { ...
           Hilight colors based on status that are used in the PC summary page.

       $Conf{CgiHeaders} = '<meta http-equiv="pragma" content="no-cache">';
           Additional CGI header text.

       $Conf{CgiImageDir} = '';
           Directory where images are stored.  This directory should be below Apache's
           DocumentRoot.  This value isn't used by BackupPC but is used by configure.pl when you
           upgrade BackupPC.

           Example:

               $Conf{CgiImageDir} = '/var/www/htdocs/BackupPC';

       $Conf{CgiExt2ContentType} = { };
           Additional mappings of file name extenions to Content-Type for individual file
           restore.  See $Ext2ContentType in BackupPC_Admin for the default setting.  You can add
           additional settings here, or override any default settings.  Example:

               $Conf{CgiExt2ContentType} = {
                           'pl'  => 'text/plain',
                    };

       $Conf{CgiImageDirURL} = '';
           URL (without the leading http://host) for BackupPC's image directory.  The CGI script
           uses this value to serve up image files.

           Example:

               $Conf{CgiImageDirURL} = '/BackupPC';

       $Conf{CgiCSSFile} = 'BackupPC_stnd.css';
           CSS stylesheet "skin" for the CGI interface.  It is stored in the $Conf{CgiImageDir}
           directory and accessed via the $Conf{CgiImageDirURL} URL.

           For BackupPC v3.x several color, layout and font changes were made.  The previous v2.x
           version is available as BackupPC_stnd_orig.css, so if you prefer the old skin, change
           this to BackupPC_stnd_orig.css.

       $Conf{CgiUserConfigEditEnable} = 1;
           Whether the user is allowed to edit their per-PC config.

       $Conf{CgiUserConfigEdit} = { ...
           Which per-host config variables a non-admin user is allowed to edit.  Admin users can
           edit all per-host config variables, even if disabled in this list.

           SECURITY WARNING: Do not let users edit any of the Cmd config variables!  That's
           because a user could set a Cmd to a shell script of their choice and it will be run as
           the BackupPC user.  That script could do all sorts of bad things.

Version Numbers

       Starting with v1.4.0 BackupPC uses a X.Y.Z version numbering system, instead of X.0Y. The
       first digit is for major new releases, the middle digit is for significant feature
       releases and improvements (most of the releases have been in this category), and the last
       digit is for bug fixes. You should think of the old 1.00, 1.01, 1.02 and 1.03 as 1.0.0,
       1.1.0, 1.2.0 and 1.3.0.

       Additionally, patches might be made available.  A patched version number is of the form
       X.Y.ZplN (eg: 2.1.0pl2), where N is the patch level.

Author

       Craig Barratt  <cbarratt@users.sourceforge.net>

       See <http://backuppc.sourceforge.net>.

Copyright

       Copyright (C) 2001-2013 Craig Barratt

Credits

Ryan Kucera contributed the directory navigation code and images for v1.5.0. He
contributed the first skeleton of BackupPC_restore. He also added a significant revision
to the CGI interface, including CSS tags, in v2.1.0, and designed the BackupPC logo.

Xavier Nicollet, with additions from Guillaume Filion, added the internationalization
(i18n) support to the CGI interface for v2.0.0. Xavier provided the French translation
fr.pm, with additions from Guillaume.

Guillaume Filion wrote BackupPC_zipCreate and added the CGI support for zip download, in
addition to some CGI cleanup, for v1.5.0. Guillaume continues to support fr.pm updates
for each new version.

Josh Marshall implemented the Archive feature in v2.1.0.

Ludovic Drolez supports the BackupPC Debian package.

Javier Gonzalez provided the Spanish translation, es.pm for v2.0.0.

Manfred Herrmann provided the German translation, de.pm for v2.0.0. Manfred continues to
support de.pm updates for each new version, together with some help from Ralph Passgang.

Lorenzo Cappelletti provided the Italian translation, it.pm for v2.1.0. Giuseppe Iuculano
and Vittorio Macchi updated it for 3.0.0.

Lieven Bridts provided the Dutch translation, nl.pm, for v2.1.0, with some tweaks from
Guus Houtzager, and updates for 3.0.0.

Reginaldo Ferreira provided the Portuguese Brazillian translation pt_br.pm for v2.2.0.

Rich Duzenbury provided the RSS feed option to the CGI interface.

Jono Woodhouse from CapeSoft Software (www.capesoft.com) provided a new CSS skin for 3.0.0
with several layout improvements. Sean Cameron (also from CapeSoft) designed new and more
compact file icons for 3.0.0.

Youlin Feng provided the Chinese translation for 3.1.0.

Karol 'Semper' Stelmaczonek provided the Polish translation for 3.1.0.

Jeremy Tietsort provided the host summary table sorting feature for 3.1.0.

Paul Mantz contributed the ftp Xfer method for 3.2.0.

Petr Pokorny provided the Czech translation for 3.2.1.

Rikiya Yamamoto provided the Japanese translation for 3.3.0.

Yakim provided the Ukrainian translation for 3.3.0.

Sergei Butakov provided the Russian translation for 3.3.0.

Many people have reported bugs, made useful suggestions and helped with testing; see the
ChangeLog and the mailing lists.

Your name could appear here in the next version!

License

       This program is free software; you can redistribute it and/or modify it under the terms of
       the GNU General Public License as published by the Free Software Foundation; either
       version 2 of the License, or (at your option) any later version.

       This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY;
       without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
       See the GNU General Public License for more details.

       You should have received a copy of the GNU General Public License in the LICENSE file
       along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple
       Place, Suite 330, Boston, MA 02111-1307 USA.