The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.


MRTG::Config - Perl module for parsing MRTG configuration files


This module, while reliable right now, is still in ALPHA stages of development... The API/methods may change. Behaviors of methods will almost certainly change. The internal structure of data will change, as will many other things.

I will try to always release 'working' versions, but anyone who expects their code that uses this module to continue working shouldn't... until I remove this warning.


Ever have the need to parse an MRTG config file? I have. I needed to parse lots and lots of them. Using the functions built-in to MRTG_lib was too slow, too complex, and used too much RAM and CPU time for my poor web server to handle - and the data structures MRTG_lib built were way more complex than I needed.

MRTG::Config can load and parse MRTG and MRTG-style confiuguration files very quickly, and the parsed directives, targets and values can be located, extracted, and manipulated through an OO interface.

This module is intended to focus on correctly parsing the format of an MRTG configuration, regardless of whether or not the directives and values, etc. are valid for MRTG. I am using both the parsing behavior of MRTG_lib's readcfg() function and the description of the format on the MRTG website as my guidelines on how to correctly parse these configuration files. I am still a short way off that goal, but this module is currently being used in a production environment with great success!


I plan on adding to this documentation and making it better organized soon, but I'm willing to answer questions directly in the mean time. Also, this is my first module, written in a hurry to appease some disgruntled engineers. I do plan on continuing to improve it, so any input, positive or negative is certainly welcome!


  use MRTG::Config;

  my $cfgFile = 'mrtg.cfg';
  my $persist_file = 'mrtg.cfg.db'; 
  my $mrtgCfg = new MRTG::Config;
  # Want to store the parsed data for use later or by
  # another program?
  foreach my $tgtName (@{$mrtgCfg->targets()}) {
    my $tgtCfg = $mrtgCfg->target($tgtName);
    # Let's assume every target has a Title.
        print $tgtCfg->{title} . "\n"; 
  # globals() has some, um, interesting things you
  # should know. Please read about it below... 
  my $globalCfg = $mrtgCfg->globals();

  # Let's assume WorkDir is set.
  print $globalCfg->{workdir} . "\n"; 


I couldn't find any modules on CPAN that would parse MRTG config files, and Tobi's code in MRTG_lib is too slow and too complicated for my needs.

This module will load a given MRTG configuration file, following Include statements, and build a set of hashes representing the confiration keywords, targets, and values.

It's _much_ faster than Tobi's code, but it also does not build a data structure nearly as deep and complex.

It does, however, properly handle a number of facilities of the MRTG configuration file format that are specified in the MRTG documentation.

Multi-Line Values

The parsing code correctly handles directives where the value spans multiple lines (sucessive lines after the first begin with whitespace). Each line of the value is contatenated together, including newlines.

Include Directives

Include directives are also handled. When an Include is encountered, the value is used as the name of another MRTG configuration file to parse. Like in MRTG_lib, if the path is not absolute (beginning with / or C:\ or whatever your system uses) this file is looked for first in the same directory as the original configuration file, and then in the current working directory.

When an Included file is loaded, it's lines are inserted into the current position in the parsing buffer and then parsing continues, as if the contents of the included file were simply copied into that position in the original file.

While I have not yet tested it, I believe 'nested' includes are followed, and the same search and loading rules apply. The path of the first config file is always used when looking for included files.

WARNING: There is no loop-checking code. If File A includes File B and File B includes File A, the parser will run until your system goes p00f, eating up memory the whole way.

The [_] Target

This module understands directives for the [_] (default) target and will interpolate these directives into all the targets that follow the definition of a [_] directive and do not explicitly define the given directive.

From what I can tell, in Tobi's implementation, [_] directives are only applied to targets that follow the definition of that particular directive. This module does likewise. Also, if a [_] directive is redefined later in the configuration, it's new value is used for all future targets. Targets that have already had that directive interpolated are not updated.

For configs that use includes to span multiple files, definitions for the [_] target go 'out of scope' when parsing new files. This is buggy behavior that will soon be fixed, or become adjustable.

Duplicate Directives and Targets

WARNING: I don't remember if this is the behavior MRTG_lib applies. I need to revisit the code and docs.

If a particular target has a directive or directives defined more than once, the last definition in the file 'wins'. The same applies to the [_] target, and also to global directives. HOWEVER: globals and [_] definitions go 'out of scope' when another cfg file is included. Again, this is buggy behavior that will soon be fixed, or become adjustable.


This module is capable of some degree of persistience, by way of DBM::Deep. Using persistience will allow you to do all sorts of interesting things, which I will not get into right now, but if you're creative I'll bet you've already thought of some! Right now, only Global and target-specific directives are persisted.


Please note - I've found that performance with DBM::Deep varies WIDELY depending on what version of DBM::Deep you are using, and whether or not you allow cpan to upgrade it's dependencies - When I allowed cpan to update everything, performance dropped by AN ORDER OF MAGNITUDE.

For best performance, I suggest using DBM::Deep .94 and whatever versions of various core modules that come with Perl 5.8.8.

Testing (or lack thereof)

Most of my testing has been done on a stock Ubuntu 7.04. Some testing has been done on Windows XP SP2 with ActiveState Perl 5.8.8.

I'll try to write some more about my tests when I do some better testing.



Returns a list of the names of MRTG targets in the parsed config.

  my @targetNames = $mrtgCfg->targets();


Returns a reference to a hash of config directives for a specific Target, given it's name in lower-case.

  my $tgtCfg = $mrtgCfg->target($tgtName);
  print $tgtCfg->{maxbytes};


Returns the name of the loaded config file. In the case of Configs using Includes, returns the name of the first config file.

  my $cfgFile = $mrtgCfg->cfg_file();
  print $cfgFile;

If you are using Includes... you may want to know in which file a specific target was defined. You do that by passing in the name of the target, in lower-case.

  my $tgtCfgFile = $mrtgCfg->cfg_file($tgtName);
  print $tgtCfgFile;

NOTE: If directives for a target were specified in more than one file, the one that the target was specified in FIRST is the one returned.


Returns a reference to a hash of Global config directives

  my $cfgGlobals = $mrtgCfg->globals();
  print $cfgGlobals->{workdir};

If there were included files, the above code currently returns the globals found ONLY in the first, original file. If you want the globals found in an Included file, pass the name of that file (the value used in the Include directive that caused it to be loaded) as an argument:

  my $incFileGlobals = $mrtgCfg->globals($fileName);
  print $incFileGlobals->{workdir};

I'm fairly certain this behavior is NOT true to how MRTG operates, but it's what currently best serves my needs... I do plan on making alterations to support the correct behavior in a future update.


Please email me if you have any questions, complaints, comments, compliments, suggestions, requests, patches, or alcoholic beverages you'd like to share. The more feedback I can get, the better I can make this module!

Please make note, though - I can not be held responsible for any problems that bugs in this module or it's dependencies may cause. That being said, I'll do my best to prevent that possibility.


Nothing by default.


1. Fix bugs 2. Fix bugs. 3. Eat a sandwich. 4. Clean up code (stupid editor breaks my indentation) 5. Clean up code (stupid me writes some sloppy perl) 6. goto 1

Also - I need to start writing tests. This release (0.03) is currently un-tested.




Stephen R. Scaffidi (


Copyright (C) 2007 by Stephen R. Scaffidi

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself, either Perl version 5.8.8 or, at your option, any later version of Perl 5 you may have available.