1. About this Guide

This guide will help you modifying PacketFence to your particular needs. It also contains information on how to add support for new switches.

The latest version of this guide is available at https://packetfence.org/documentation/.

1.1. Other sources of information

Network Devices Configuration Guide

Covers switch, controllers and access points configuration.

Administration Guide

Covers PacketFence installation, configuration and administration.

NEWS

Covers noteworthy features, improvements and bug fixes by release.

UPGRADE

Covers compatibility related changes, manual instructions and general notes about upgrading.

ChangeLog

Covers all changes to the source code.

These files are included in the package and release tarballs.

2. Documentation

The in-depth or more technical documentation is always as close to the code as possible. Always look at the POD doc [1]. To do so, the preferred way is using the perldoc command as follows:

perldoc lib/pfconfig/cached.pm

3. Code conventions

3.1. Code style

Caution
Work in progress.

We are slowly migrating away from an automated perltidy code style. The reason we are not doing another pass of tidy is that it messes up code history and makes maintainer’s job more complicated than it should be. Every new change uses the new guidelines so over time the old code style will slowly disappear.

  • Lines of 120 character width maximum

  • No tab characters

  • Stay consistent with surrounding white spaces

  • Document each subroutine in POD format (perldoc perlpod)

  • Use constants instead of hard coded strings or numbers (use constant or Readonly modules)

  • in object-oriented modules we use CamelCase [2] notation (ex: $radiusRequest->getVoIpAttributes();)

  • in procedural modules we use Perl’s usual notation (ex: $node_info{'pid'} = $current_request{'pid'};)

  • regular expressions should be documented (with the /x modifier)

    if ($phone_number =~ /
        ^\(?([2-9]\d{2})\)?  # captures first 3 digits allows parens
        (?:-|.|\s)?          # separator -, ., space or nothing
        (\d{3})              # captures 3 digits
        (?:-|.|\s)?          # separator -, ., space or nothing
        (\d{4})$             # captures last 4 digits
        /x) {
            return "$1$2$3";
    }
  • SQL should be capitalized, properly indented and always use named fields (no *)

    $node_statements->{'node_add_sql'} = get_db_handle()->prepare(<<'SQL');
        INSERT INTO node (
            mac, pid, category_id, status, voip, bypass_vlan,
            detect_date, regdate, unregdate, lastskip,
            user_agent, computername, dhcp_fingerprint,
            last_arp, last_dhcp,
            notes,
        ) VALUES (
            ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?
        )
    SQL

4. HTTP JSON API

PacketFence provides an HTTP JSON API which exposes most of its functionality.

The full API specification for:

The API is exposed on 127.0.0.1 (localhost) as well as the management interface of the server for remote calls. Users access the API on TCP port 9999 over a secure connection (HTTPS).

Authentication can be done using either the webservices credentials or any credentials that are valid on the web admin interface.

4.1. How to use the API

Note
The API specification mentionned above provides ready-to-use curl commands through the "Try it out" button.

First, get an authentication token with the webservices credentials or an admin account:

curl -X POST "https://PF_MANAGEMENT_IP:9999/api/v1/login" \
-H  "accept: application/json" \
-H  "Content-Type: application/json" \
-d "{\"password\":\"admin\",\"username\":\"admin\"}" \
--insecure |
python -m json.tool

You will get following response :

{
   "token": "MY_TOKEN"
}

Then use this token to query the API. The following example fetches the domains configuration:

curl -X GET "https://PF_MANAGEMENT_IP:9999/api/v1/config/domains" \
-H  "accept: application/json" \
-H  "Authorization: MY_TOKEN" \
--insecure |
python -m json.tool

Although the API should mostly stay the same, backward compatibility is not 100% guaranteed for the moment until the v1 API reaches full maturity.

5. Customizing PacketFence

5.1. Captive Portal

5.1.1. Presentation

XHTML Templates

Captive portal content use Template Toolkit templates. The default template files are located in /usr/local/pf/html/captive-portal/templates. You can freely edit the HTML code in these files.

Each template relies on layout.html for the common top and bottom portions of the page.

Internationalized AUP

In the event that you need an Acceptable Usage Policy that should be translated in different languages in the captive portal, you can create specially named templates that will be used for different languages.

For example, if the browser locale is es_ES, creating a template named aup_text.es.html will display this AUP when it detects this locale. Same goes for a browser with locale en_US, where creating a template named aup_text.en.html will be used for English browsers.

The template aup_text.html is used as the final fallback for all locales.

Note that you cannot use the full locale in the template name (i.e. aup_text.en_US.html) as only the two letter prefix should be used.

CSS

The default stylesheet is the result of processing some Sass files structured around the solid inuitcss architectural foundation.

  • /usr/local/pf/html/common/Gruntfile.js — the task runner configuration file

  • /usr/local/pf/html/common/package.js — the npm dependencies required to process the Sass files

  • /usr/local/pf/html/common/scss/ — the directory containing the Sass source files

  • /usr/local/pf/html/common/styles.css — the generated stylesheet

In order to prepare your environment to process the Sass files, you need to install grunt (the task runner) and additional nodejs modules, including the various inuitcss components:

sudo npm install -g grunt-cli
cd /usr/local/pf/html/common
npm install

Once all modules are downloaded, you can customize the Sass files and generate a new stylesheet:

grunt dist

To change the color palette of the captive portal, modify the file scss/_settings.colors.scss.

Workflow

When a HTTP request is received by the Apache web server, the following workflow happens:

  1. URL is compared against the redirection instructions in /usr/local/pf/conf/httpd.conf.d/captive-portal-cleanurls.conf

  2. Requested CGI script in /usr/local/pf/html/captive-portal/ is executed

  3. CGI script calls a generate_<type> which is defined in /usr/local/pf/lib/pf/web.pm

  4. The generate_<type> function populate the proper template in /usr/local/pf/html/captive-portal/templates in order to render the page

Remediation Pages

The remediation page shown to the user during isolation are specified through the URL parameter of the given violation in /usr/local/pf/conf/violations.conf. In its default configuration, PacketFence uses Template Toolkit to render text provided in the directory /usr/local/pf/html/captive-portal/templates/violations and obeys to everything mentioned in the Presentation section.

Translations

The language of the user registration pages is selected through the general.locale configuration parameter. Translatable strings are handled differently for the Remediation pages and the rest of the captive portal:

  • Remediation pages

    Strings defined in the violation pages (in /usr/local/pf/html/captive-portal/templates/violations) will be looked up in the translation files in /usr/local/pf/conf/locale/.. and if a translation is available the translated string will be the one visible on the captive portal.

    Also, if you create a violation template with the name of your locale in /usr/local/pf/html/captive-portal/templates/violations in the format: <template_name>.<locale_name>.html. It will be loaded instead of the default <template_name>.html and so you can put strings and HTML directly in your target language without the hassle of escaping everything properly as you would need to do with gettext.

    For example, if malware.es_ES.html exists and you are using the es_ES (Spanish) locale then it will be loaded instead of malware.html on a violation set to load the malware template.

  • Rest of the captive portal

    In the templates, if a string is in a i18n() call it will be translated. Also pf::web takes care of performing some of the other translations.

5.2. Adding custom fields to the database

You can, if needed, add additional fields to the PacketFence database. Keep in mind though that this might lead to more work when you upgrade to the next PacketFence version. Depending on the degree of integration of these fields with PacketFence, you’ll have to execute one or more of the following steps

5.2.1. Adding a field to the database only

In this case, the field is part of one of the main PacketFence tables, but PacketFence is unaware of it. PacketFence won’t consult the field and won’t be able to modify it. A possible usage scenario would be a 3rd party application which maintains this field.

Since PacketFence doesn’t have to know about the field, all you have to do is execute your SQL ALTER TABLE query and you are done.

5.2.2. Adding a field and giving PacketFence read-only access

In this case, PacketFence can show the contents of the table using both pfcmd but won’t be able to modify the contents of the field.

Start by modifying the database table using an SQL ALTER TABLE query.

Then, modify the Perl module having the same name as the table you have added the field to, i.e. If you added the field to the node table, then edit /usr/local/pf/lib/pf/node.pm. You’ll have to modify the SQL SELECT queries at the beginning of the file to include your new field and, possibly the functions using these queries. If your new field should be used in reports, the dashboard or graphs, you’ll also have to modify the queries in /usr/local/pf/lib/pf/pfcmd/graph.pm, /usr/local/pf/lib/pf/pfcmd/report.pm and /usr/local/pf/lib/pf/pfcmd/dashboard.pm.

5.2.3. Adding a field and giving PacketFence read-write access

Start by creating the read-only field as described above.

Then, modify the SQL UPDATE and INSERT queries in the database tables' Perl module, as well as the associated functions.

5.3. VLAN assignment

PacketFence uses the getRegisteredRole function defined in pf::role::custom to determine a node’s VLAN. Here’s the default function:

sub getRegisteredRole {
    #$switch is the switch object (pf::Switch)
    #$ifIndex is the ifIndex of the computer connected to
    #$mac is the mac connected
    #$node_info is the node info hashref (result of pf::node's node_view on $mac)
    #$conn_type is set to the connection type expressed as the constant in pf::config
    #$user_name is set to the RADIUS User-Name attribute (802.1X Username or MAC address under MAC Authentication)
    #$ssid is the name of the SSID (Be careful: will be empty string if radius non-wireless and undef if not radius)
    my ($self, $switch, $ifIndex, $mac, $node_info, $connection_type, $user_name, $ssid) = @_;

    my $logger = Log::Log4perl->get_logger();

    return $switch->getVlanByName('normalVlan');
}

As you can see, the function receives several parameters (such as the switch and full node details) which allow you to return the VLAN in a way that matches exactly your needs!

6. SNMP

6.1. Introduction

Good places to start reading about SNMP are http://en.wikipedia.org/wiki/SNMP and http://www.net-snmp.org/.

When working with SNMP, you’ll sooner or later (in fact more sooner than later) be confronted with having to translate between OIDs and variable names. When the OIDs are part of the Cisco MIBs, you can use the following tool to do the translation: http://tools.cisco.com/Support/SNMP/public.jsp. Otherwise, you’ll have to use snmptranslate for example and setup your own collection of MIBs, provided (hopefully) by the manufacturer of your network equipment.

6.2. Obtaining switch and port information

Below are some example of how to obtain simple switch and port information using SNMP. We’ll assume that your switch understands SNMP v2, has the read community public defined and is reachable at 192.168.1.10.

6.2.1. Switch Type

snmpwalk -v 2c -c public 192.168.1.10 sysDescr

6.2.2. Switchport indexes and descriptions

snmpwalk -v 2c -c public 192.168.1.10 ifDescr

6.2.3. Switchport types

snmpwalk -v 2c -c public 192.168.1.10 ifType

6.2.4. Switchport status

snmpwalk -v 2c -c public 192.168.1.10 ifAdminStatus
snmpwalk -v 2c -c public 192.168.1.10 ifOperStatus

7. Supporting new network hardware

PacketFence is designed to ease the addition of support for new network hardware referred to as Network Devices. All supported network devices are represented through Perl objects with an extensive use of inheritance. Adding support for a new product comes down to extending the pf::Switch class (in /usr/local/pf/lib/pf).

The starting point to adding support for a new network device should be the vendor’s documentation! First of all, you’ll have to figure out the exact capabilities of the switch and how these capabilities will fit into PacketFence. Is it a Switch, an Access-Point or a Wireless Controller?

7.1. Switch

Will you be able to use only link change traps? Does your switch allow you to use MAC notification traps? Port Security? MAC Authentication? 802.1X?

You need to define a new class which inherits from pf::Switch and defines at least the following functions:

  • getMacAddrVlan

  • getVersion

  • getVlan

  • getVlans

  • isDefinedVlan

  • parseTrap

  • _getMacAtIfIndex

  • _setVlan

The parseTrap function will need to return an hash with keys trapType and trapIfIndex. The associated values must be up or down for trapType and the traps' ifIndex for trapIfIndex. See a similar switch’s implementation for inspiration. Usually recent modules are better coded than older ones.

7.1.2. MAC notification capabilities

In addition to the functions mentioned for link change, you need to define the following function:

  • isLearntTrapsEnabled

Also, your parseTrap function will need to return trapOperation, trapVlan and trapMac keys in addition to trapType equals mac. See a similar switch’s implementation for inspiration. Usually recent modules are better coded than older ones.

7.1.3. Port security capabilities

In addition to the functions mentioned for link change, you need to define the following functions:

  • isPortSecurityEnabled

  • authorizeMAC

In this case, the parseTrap function needs to return secureMacAddrViolation for the trapType key. See a similar switch’s implementation for inspiration. Usually recent modules are better coded than older ones.

7.1.4. MAC Authentication

Note
Work in progress
NAS-Port translation

Often the ifIndex provided by the switch in a RADIUS Access-Request is not the same as it’s real world physical equivalent. For example in Cisco requests are in the 50xxx while physical ifIndex are 10xxx. In order for PacketFence to properly shut the port or request re-authentication a translation between the two is required. To do so provide an implementation of the following interface:

  • NasPortToIfIndex

MAC Authentication re-evaluation

MAC Authentication re-evaluation is necessary in order to provoke a VLAN change in the PacketFence system. This happens for instance when a node is isolated based on an IDS event or when the user successfully authenticates through the captive portal. The default implementation in pf::Switch will bounce the port if there is no Voice over IP (VoIP) devices connected to the port. Otherwise it will do nothing and send an email. If your device has specific needs (for example it doesn’t support RADIUS Dynamic VLAN Assignments) override:

  • handleReAssignVlanTrapForWiredMacAuth

Please note that the default implementation works 99% of the time. If you are unsure whether to override, it means you don’t need to override.

Once the MAC Authentication works, add the Wired MAC Auth capability to the switch’s code with:

sub supportsWiredMacAuth { return $TRUE; }

7.1.5. 802.1X

Note
Work in progress
NAS-Port translation

Often the ifIndex provided by the switch in a RADIUS Access-Request is not the same as it’s real world physical equivalent. For example in Cisco requests are in the 50xxx while physical ifIndex are 10xxx. In order for PacketFence to properly shut the port or request re-authentication a translation between the two is required. To do so provide an implementation of the following interface:

  • NasPortToIfIndex

So far the implementation has been the same for MAC Authentication and 802.1X.

Force 802.1X re-authentication

802.1X re-authentication is necessary in order to provoke a VLAN change in the PacketFence system. This happens for instance when a node is isolated based on an IDS event or when the user successfully authenticates through the captive portal. The default implementation in pf::Switch uses SNMP and the standard IEEE8021-PAE-MIB and is generally well supported. If the default implementation to force 802.1X re-authentication doesn’t work override:

  • dot1xPortReauthenticate

Proper 802.1X implementations will perform re-authentication while still allowing traffic to go through for supplicants under re-evaluation.

Once the 802.1X works, add the Wired Dot1X capability to the switch’s code with:

sub supportsWiredDot1x { return $TRUE; }

7.1.6. RADIUS Dynamic Authorization (RFC3576)

Note
RADIUS Dynamic Authorization implementation is not recommended on the wired side at this point.

RADIUS Dynamic Authorization also known as RADIUS Change of Authorization (CoA) or RADIUS Disconnect Messages is supported by PacketFence starting with version 3.1.

On wired network devices CoA can be used to change the security posture of a MAC and perform other functions like bounce a port. So far we only encountered support for CoA on the wired side on the Cisco hardware. For an implementation example check _radiusBounceMac in pf::Switch::Cisco.

7.1.7. Floating Network Devices Support

Floating Network Devices are described in the Administration Guide under "Floating Network Devices" in the "Optional Components" section. Refer to this documentation if you don’t know what Floating Network Devices are.

In order to support Floating Network Devices on a switch, you need to implement the following methods:

  • setPortSecurityEnableByIfIndex($ifIndex, $enable)

  • isTrunkPort($ifIndex)

  • setModeTrunk($ifIndex, $enable)

  • setTaggedVlans($ifIndex, $switch_locker_ref, @vlans)

  • removeAllTaggedVlans($ifIndex, $switch_locker_ref)

You might need to implement the following:

  • enablePortConfigAsTrunk($mac, $switch_port, $switch_locker, $taggedVlans)

    Provided by pf::Switch core as the glue between setModeTrunk(), setTaggedVlans() and removeAllTaggedVlans(). Override if necessary.

  • disablePortConfigAsTrunk($switch_port)

    Provided by pf::Switch core as the glue between setModeTrunk(), setTaggedVlans() and removeAllTaggedVlans(). Override if necessary.

  • enablePortSecurityByIfIndex($ifIndex)

    Provided by pf::Switch core as a slim accessor to setPortSecurityEnableByIfIndex(). Override if necessary.

  • disablePortSecurityByIfIndex($ifIndex)

    Provided by pf::Switch core as a slim accessor to setPortSecurityEnableByIfIndex(). Override if necessary.

  • enableIfLinkUpDownTraps($ifIndex)

    Provided by pf::Switch core as a slim accessor to setIfLinkUpDownTrapEnable. Override if necessary.

  • disableIfLinkUpDownTraps($ifIndex)

    Provided by pf::Switch core as a slim accessor to setIfLinkUpDownTrapEnable. Override if necessary.

Once all the required methods are implemented, enable the capability in the switch’s code with:

sub supportsFloatingDevice { return $TRUE; }

7.2. Wireless Access-Points or Controllers

7.2.1. Minimum hardware requirements

PacketFence’s minimum requirements regarding Wireless hardware is:

  • definition of several SSID with several VLANs inside every SSID (minimum of 2 VLANs per SSID)

  • RADIUS authentication (MAC Authentication / 802.1X)

  • Dynamic VLAN assignment through RADIUS attributes

  • a means to de-associate or de-authenticate a client through CLI (Telnet or SSH), SNMP, RADIUS Dyn-Auth [3] or WebServices

Most of these features are available on enterprise grade Access Points (AP) or Controllers. Where the situation starts to vary wildly is for deauthentication support.

7.2.2. De-authentication techniques

CLI (SSH or Telnet)

An error prone interface and requires preparation for the SSH access or is insecure for Telnet. Not recommended if you can avoid it.

SNMP

SNMP de-authentication works well when available. However Vendor support is not consistent and the OID to use are not standard.

RADIUS Dynamic Authorization (RFC3576)

RADIUS Dynamic Authorization also known as RADIUS Change of Authorization (CoA) or RADIUS Disconnect Messages is supported by PacketFence starting with version 3.1. When supported it is the preferred technique to perform de-authentication. It is standard and requires less configuration from the user.

An actual implementation can be found in pf::Switch::Aruba.

7.2.3. Template module

Start with a copy of the template module pf/lib/pf/Switch/WirelessModuleTemplate.pm and fill in appropriate documentation and code.

7.2.4. Required methods

You need to implement at least:

getVersion()

Fetches firmware version

parseTrap()

Parses the SNMP Traps sent by the hardware. For wireless hardware an empty method like the one in pf::Switch::WirelessModuleTemplate is ok.

deauthenticateMac()

Performs deauthentication

supportsWirelessMacAuth()

Return $TRUE if MAC-Authentication is supported

supportsWirelessDot1x()

Return $TRUE if 802.1X (aka WPA-Enterprise) is supported

7.2.5. Override methods

If default implementation of the following methods doesn’t work you will need to override them:

extractSsid()

Extract SSID from RADIUS Request

7.2.6. Special case: bridged versus tunneled modes and deauthentication

It is important to validate the Access-Point (AP) to Controller relationship when operating in bridged mode versus when operating in tunneled mode. For example, some hardware will send the RADIUS Access-Request from the AP when in bridged mode even though it is controlled by a controller. This behavior impacts deauthentication because it still needs to be performed on the controller. To support this behavior a switches.conf parameter was introduced: controller_ip.

When adding a new Wireless module try to validate the bridged versus tunneled behavior and modify deauthenticateMac() to honor controller_ip if required.

7.3. The "adding a new network device module in PacketFence" checklist

Here’s a quick rundown of the several files you need to edit in order to add a new switch into PacketFence. There’s a plan to reduce this amount of work in progress see issue #1085.

  • Tested model and firmware version should be documented in module’s POD

  • Any bugs and limitations should be documented in module’s POD

  • Add it to pf/html/admin/configuration/switches_add.php and switches_edit.php

  • Make sure that all tests pass

  • Add configuration documentation to the Network Devices Guide

  • Add switch to the Network Devices Guide’s switch chart

  • Add switch to the chart in README.network-devices

8. Developer recipes

8.1. Running development version

8.1.1. Bleeding edge

For day to day development one can run a checkout of the current development branch in /usr/local/pf/ and develop there within a working setup.

Care should be taken not to commit local configuration files changes and files not in the repository.

8.1.2. Not so bleeding edge

Using the development yum repository and upgrade PacketFence often is a good way to proceed. Check our snapshots download page for instructions.

Make sure you read the UPGRADE document after every upgrades to avoid any surprises.

8.2. New Exception handling techniques under testing

Little attention was given to error-handling in PacketFence’s early design. This is understandable as it wasn’t probably the most bang-for-the-buck thing to do. However we must now live with a large code base that explodes at runtime or that doesn’t differentiate an erroneous condition from an undefined or 0 value. Refactoring to improve error-handling will be gradual but new code should follow these tips:

  1. use Try::Tiny

  2. wrap stuff in try {...} catch {...}; (and optionally a finally {...};)

  3. in the code use die(...); to throw an exception and make the error message meaningful

  4. in the catch block, use $logger->logcarp("explanation: $_") if I want output to the CLI, otherwise, choose wisely

This catches a lot of errors (including runtime crashers) and allows us to recover from these conditions.

So far, it is mandatory to wrap the Web Services enabled network devices modules' code since SOAP::Lite will die on you if host is unreachable for example (actually it’s LWP::UserAgent who will).

9. Contributing

Here are some golden rules of contributing to PacketFence:

The place to be if you want to contribute to the PacketFence project is our developers mailing list: https://lists.sourceforge.net/lists/listinfo/packetfence-devel. Let us know your issues, what you are working on and how you want to solve your problems. The more you collaborate the greater the chances that your work will be incorporated in a timely fashion.

Good chances that the bug you want to fix or the feature you want to implement is already filed and that information in the ticket will help you.

  • Please provide small, focused and manageable patches or pull-requests

If you plan on doing a lot of code, use git and track our current stable branch called stable. Develop the feature in small chunks and stay in touch with us. This way it’ll be merged quickly in our code base. Ideally there would be no big code dumps after finishing a feature.

9.1. Creating patches

Note
Since we migrated to git / github, using these tools is recommended over sending patches by hand.

Patches should be sent in unified diff format. This can be obtained from the diff or git tools.

diff -u oldfile newfile

or from a checkout of the PacketFence source code from git:

git diff

9.2. Translations

The internationalization process uses gettext. If you are new to gettext, please consult http://www.gnu.org/software/gettext/manual/gettext.html#Overview for a quick introduction.

The PO files are stored in /usr/local/pf/conf/locale. List that directory to see the languages we currently have translations for.

9.2.1. Online using Transifex

We use the hosted service Transifex to translate PacketFence’s PO files. It offers the possibility to translate all the strings online as well as providing a command-line tool to push your changes. It’s very convenient.

To use Transifex, you must first sign up for a free account here: https://www.transifex.net/plans/signup/free/

If you need further help about using Transifex, you might want to have a look here.

9.2.2. Using traditional method

If you want to add support for a new language, please follow these steps:

  1. create a new language subdirectory in /usr/local/pf/conf/locale

  2. change into your newly created directory

  3. create a new subdirectory LC_MESSAGES

  4. change into your newly created directory

  5. copy the file /usr/local/pf/conf/locale/en/LC_MESSAGES/packetfence.po into your directory

  6. translate the message strings in packetfence.po

  7. create the MO file by executing:

    /usr/bin/msgfmt packetfence.po

Submit your new translation to the PacketFence project by contacting us at packetfence-devel@lists.sourceforge.net.

9.3. Admin GUI

The Admin GUI can be customized via Controllers and templates.

9.3.1. Templates

You can add customize templates to the /usr/local/pf/html/pfappserver/root-custom/ folder to override current templates.

However when doing this you must keep track of any changes to those templates across versions.

9.3.2. Stylesheets and JavaScript

The main stylesheet is the result of processing some Sass files structured around the excellent Bootstrap framework.

  • /usr/local/pf/html/pfappserver/root/static/Gruntfile.js — the task runner configuration file

  • /usr/local/pf/html/pfappserver/root/static/package.js — the npm dependencies required to process the Sass files and to package the JavaScript code

  • /usr/local/pf/html/pfappserver/root/static/bower.json — the list of required components

  • /usr/local/pf/html/pfappserver/root/static/scss/ — the directory containing the Sass source files

  • /usr/local/pf/html/pfappserver/root/static/css/styles.css — the generated stylesheet

  • /usr/local/pf/html/pfappserver/root/static/js/vendor/ — the directory containing the external JavaScript libraries

In order to prepare your environment, you need to install grunt (the task runner) and bower (the package manager):

sudo npm install -g grunt-cli
sudo npm install -g bower
cd /usr/local/pf/html/pfappserver/root/static/
npm install
bower install

Once all modules and components are downloaded, you can customize the Sass files or JavaScript files and generate a new version of the static files:

grunt dist

9.3.3. Controllers

You change the behavior of the controllers by modifying the controllers in the path.

/usr/local/pf/html/pfappserver/lib/pfappserver/Controller/

However when doing this you must keep track of any changes to those Controllers across versions.

10. Additional Information

For more information, please consult the mailing archives or post your questions to it. For details, see:

11. Commercial Support and Contact Information

For any questions or comments, do not hesitate to contact us by writing an email to: support@inverse.ca.

Inverse (http://inverse.ca) offers professional services around PacketFence to help organizations deploy the solution, customize, migrate versions or from another system, performance tuning or aligning with best practices.

Hourly rates or support packages are offered to best suit your needs.

Please visit http://inverse.ca/ for details.

12. GNU Free Documentation License

Please refer to http://www.gnu.org/licenses/fdl-1.2.txt for the full license.


1. Perl’s Plain Old Documentation: http://perldoc.perl.org/perlpod.html
2. http://en.wikipedia.org/wiki/CamelCase
3. RADIUS Dynamic Authorization (RFC 3576) aka Change of Authorization (CoA) or Disconnect-Messages (DM aka PoD)