<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
"http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd" [
<!ENTITY mdash "—" >
]>
<chapter id="hooks-libraries">
<title>Hooks Libraries</title>
<section id="hooks-libraries-introduction">
<title>Introduction</title>
<para>
Although Kea offers a lot of flexibility, there may be cases where
its behavior needs customisation. To accommodate this possibility,
Kea includes the idea of "Hooks". This feature lets Kea load one
or more dynamically-linked libraries (known as "hooks libraries")
and, at various points in its processing ("hook points"), call
functions in them. Those functions perform whatever custom
processing is required.
</para>
<para>
Hooks libraries are attached to individual Kea processes, not to
Kea as a whole. This means (for example) that it is possible
to associate one set of libraries with the DHCP4 server and a
different set to the DHCP6 server.
</para>
<para>
Another point to note is that it is possible for a process to
load multiple libraries. When processing reaches a hook point,
Kea calls the hooks library functions attached to it. If multiple
libraries have attached a function to a given hook point, Kea calls
all of them, in the order in which the libraries are specified in
the configuration file. The order may be important: consult the
documentation of the libraries to see if this is the case.
</para>
<para>
The next section describes how to configure hooks libraries. If you
are interested in writing your own hooks library, information can be
found in the <ulink url="https://jenkins.isc.org/job/Fedora20_32_doxygen_doc/doxygen/">Kea
Developer's Guide</ulink>.
</para>
</section> <!-- end Introduction -->
<section>
<title>Configuring Hooks Libraries</title>
<para>
The hooks libraries for a given process are configured using the
<command>hooks-libraries</command> keyword in the
configuration for that process. (Note that
the word "hooks" is plural). The value of the keyword
is an array of map structures, each structure corresponding to a hooks
library. For example, to set up two hooks libraries for the DHCPv4
server, the configuration would be:
<screen>
<userinput>"Dhcp4": {
:
"hooks-libraries": [
{
"library": "/opt/charging.so"
},
{
"library": "/opt/local/notification.so",
"parameters": {
"mail": "spam@example.com",
"floor": 13,
"debug": false,
"users": [ "alice", "bob", "charlie" ],
"languages": {
"french": "bonjour",
"klingon": "yl'el"
}
}
}
]
:
}</userinput>
</screen>
</para>
<note><para>
This is a change to the syntax used in Kea 0.9.2 and earlier, where
hooks-libraries was a list of strings, each string being the name of
a library. The change was made in Kea 1.0 to facilitate the
specification of library-specific parameters, a capability
available in Kea 1.1.0 onwards.
</para></note>
<note>
<para>
The library reloading behavior has changed in Kea 1.1. Libraries are
reloaded, even if their list hasn't changed. Kea does that, because
the parameters specified for the library (or the files those
parameters point to) may have changed.
</para>
</note>
<para>
Libraries may have additional parameters. Those are not mandatory in the
sense that there may be libraries that don't require them. However, for
specific library there is often specific requirement for specify certain
set of parameters. Please consult the documentation for your library
for details. In the example above, the first library has no parameters.
The second library has five parameters, specifying mail (string
parameter), floor (integer parameter), debug (boolean parameter) and
even lists (list of strings) and maps (containing strings). Nested
parameters could be used if the library supports it. This topic is
explained in detail in the Hooks Developer's Guide in the "Configuring
Hooks Libraries" section.
</para>
<para>
Notes:
<itemizedlist mark='bullet'>
<listitem><para>
The full path to each library should be given.
</para></listitem>
<listitem><para>
As noted above, order may be important - consult the documentation for
each library.
</para></listitem>
<listitem><para>
An empty list has the same effect as omitting the
<command>hooks-libraries</command> configuration element all together.
</para>
<note><para>
There is one case where this is not true: if Kea
is running with a configuration that contains a
<command>hooks-libraries</command> item, and that item is
removed and the configuration reloaded, the removal will be
ignored and the libraries remain loaded. As a workaround,
instead of removing the <command>hooks-libraries</command>
item, change it to an empty list. This will be fixed in a
future version of Kea.
</para></note>
</listitem>
</itemizedlist>
</para>
<para>
At the present time, only the kea-dhcp4 and kea-dhcp6 processes support
hooks libraries.
</para>
</section>
<section>
<title>Available Hooks Libraries</title>
<para>
As described above, the hooks functionality provides a way to customize
a Kea server without modifying the core code. ISC has chosen to take
advantage of this feature to provide functions that may only be useful
to a subset of Kea users. To this end ISC has created some hooks
libraries; these discussed in the following sections.
</para>
<note><para>
Some of these libraries will be available with the base code while others
will be shared with organizations supporting development of Kea
, possibly as a 'benefit' or 'thank you' for helping to sustain
the larger Kea project. If you would like to get access to those
libraries, please consider taking out a support contract: this includes
professional support, advance security notifications, input into our
roadmap planning, and many other benefits, while helping
making Kea sustainable in the long term.
</para></note>
<para>Currently the following libraries are available or planned from ISC:
<table frame="all" id="hook-libs">
<title>List of available hooks libraries</title>
<tgroup cols='3'>
<colspec colname='name' />
<colspec colname='avail' />
<colspec colname='description' />
<thead>
<row>
<entry>Name</entry>
<entry>Availability</entry>
<entry>Since</entry>
<entry>Description</entry>
</row>
</thead>
<tbody>
<row>
<entry>user_chk</entry>
<entry>Kea sources</entry>
<entry>Kea 0.8</entry>
<entry>Reads known users list from a file. Unknown users
will be assigned a
lease from the last subnet defined in the configuration file,
e.g. to redirect them a captive portal. This demonstrates how an
external source of information can be used to influence the Kea
allocation engine. This hook is part of the Kea source code and is
available in the src/hooks/dhcp/user_chk directory.</entry>
</row>
<row>
<entry>Forensic Logging</entry>
<entry>Support customers</entry>
<entry>Kea 1.1.0</entry>
<entry>This library provides hooks that record a detailed log of
lease assignments and renewals into a set of log files. In many
legal jurisdictions companies, especially ISPs, must record
information about the addresses they have leased to DHCP
clients. This library is designed to help with that
requirement. If the information that it records is sufficient it
may be used directly. If your jurisdiction requires that you save
a different set of information, you may use it as a template or
example and create your own custom logging hooks.</entry>
</row>
<row>
<entry>Flexible Identifier</entry>
<entry>Support customers</entry>
<entry>Kea 1.2.0 beta</entry>
<entry>Kea software provides a way to handle host reservations
that include addresses, prefixes, options, client classes and
other features. The reservation can be based on hardware address,
DUID, circuit-id or client-id in DHCPv4 and using hardware address
or DUID in DHCPv6. However, there are sometimes scenarios where the
reservation is more complex, e.g. uses other options that
mentioned above, uses part of specific options or perhaps even a
combination of several options and fields to uniquely identify a
client. Those scenarios are addressed by the Flexible Identifiers
hook application. It allows defining an expression, similar to
the one used in client classification,
e.g. substring(relay6[0].option[37],0,6). Each incoming packet is
evaluated against that expression and its value is then searched
in the reservations database.
</entry>
</row>
</tbody>
</tgroup>
</table>
</para>
<para>
ISC hopes to see more hooks libraries become available as time
progresses, both developed internally and externally. Since
this list may evolve dynamically, we decided to keep it on a
wiki page, available at this link: <ulink
url="http://kea.isc.org/wiki/Hooks">http://kea.isc.org/wiki/Hooks</ulink>.
If you are a developer or are aware of any hooks libraries not
listed there, please send a note to the kea-users or kea-dev
mailing lists and someone will update it.
</para>
<section>
<title>user_chk: Checking User Access</title>
<para>
The user_chk library is the first hooks library published by ISC. It
attempts to serve several purposes:
<itemizedlist>
<listitem>
<para>To assign "new" or "unregistered" users to a
restricted subnet, while "known" or "registered" users are assigned
to unrestricted subnets.</para>
</listitem>
<listitem>
<para>To allow DHCP response options or vendor option
values to be customized based upon user identity. </para>
</listitem>
<listitem>
<para>To provide a real time record of the user registration
activity which can be sampled by an external consumer.</para>
</listitem>
<listitem>
<para> To serve as a demonstration of various capabilities
possible using the hooks interface.</para>
</listitem>
</itemizedlist>
</para>
<para>
Once loaded, the library allows segregating incoming requests into
known and unknown clients. For known clients, the packets are
processed mostly as usual, except it is possible to override certain
options being sent. That can be done on a per host basis. Clients
that are not on the known hosts list will be treated as unknown and
will be assigned to the last subnet defined in the configuration file.
</para>
<para>
As an example of use, this behavior may be used to put unknown users into a
separate subnet that leads to a walled garden, where they can only
access a registration portal. Once they fill in necessary data, their
details are added to the known clients file and they get a proper
address after their device is restarted.
</para>
<note><para>This library was developed several years before the host
reservation mechanism has become available. Currently host reservation is
much more
powerful and flexible, but nevertheless the user_chk capability to consult
and external source of information about clients and alter Kea's
behavior is useful and remains of educational value.
</para></note>
<para>
The library reads the /tmp/user_chk_registry.txt file while being
loaded and each time an incoming packet is processed. The file is expected
to have each line contain a self-contained JSON snippet which must
have the following two entries:
<itemizedlist>
<listitem><para><command>type</command>, whose value
is "HW_ADDR" for IPv4 users or "DUID" for IPv6
users</para></listitem>
<listitem><para><command>id</command>, whose value is
either the hardware address or the DUID from the request
formatted as a string of hex digits, with or without
":" delimiters.</para></listitem>
</itemizedlist>
and may have the zero or more of the following entries:
<itemizedlist>
<listitem><para><command>bootfile</command> whose value
is the pathname of the desired file</para></listitem>
<listitem><para><command>tftp_server</command> whose
value is the hostname or IP address of the desired
server</para></listitem>
</itemizedlist>
A sample user registry file is shown below:
<screen>{ "type" : "HW_ADDR", "id" : "0c:0e:0a:01:ff:04", "bootfile" : "/tmp/v4bootfile" }
{ "type" : "HW_ADDR", "id" : "0c:0e:0a:01:ff:06", "tftp_server" : "tftp.v4.example.com" }
{ "type" : "DUID", "id" : "00:01:00:01:19:ef:e6:3b:00:0c:01:02:03:04", "bootfile" : "/tmp/v6bootfile" }
{ "type" : "DUID", "id" : "00:01:00:01:19:ef:e6:3b:00:0c:01:02:03:06", "tftp_server" : "tftp.v6.example.com" }</screen>
</para>
<para>As with any other hooks libraries provided by ISC, internals of the
user_chk code are well documented. You can take a look at the <ulink
url="https://jenkins.isc.org/job/Fedora20_32_doxygen_doc/doxygen/d8/db2/libdhcp_user_chk.html">Kea Developer's Guide section dedicated to the user_chk library</ulink>
that discusses how the code works internally. That, together with
our general entries in <ulink
url="https://jenkins.isc.org/job/Fedora20_32_doxygen_doc/doxygen/">Hooks
Framework section</ulink> should give you some pointers how to extend
this library and perhaps even write your own from scratch.</para>
</section>
<section>
<title>legal_log: Forensic Logging Hooks</title>
<para>
This section describes the forensic log hooks library. This library
provides hooks that record a detailed log of lease assignments
and renewals into a set of log files. Currently this library
is only available to ISC customers with a support contract.
</para>
<para>
In many legal jurisdictions companies, especially ISPs, must record
information about the addresses they have leased to DHCP clients.
This library is designed to help with that requirement. If the
information that it records is sufficient it may be used directly.
If your jurisdiction requires that you save a different set of
information you may use it as a template or example and create your
own custom logging hooks.
</para>
<para>
This logging is done as a set of hooks to allow it to be customized
to any particular need. Modifying a hooks library is easier and
safer than updating the core code. In addition by using the hooks
features those users who don't need to log this information can
leave it out and avoid any performance penalties.
</para>
<section>
<title>Log File Naming</title>
<para>
The names for the log files have the following form:
</para>
<screen>
path/base-name.CCYYMMDD.txt
</screen>
<para>
The "path" and "base-name" are supplied in the
configuration as described below see
<xref linkend="forensic-log-configuration"/>. The next part of the name is
the date the log file was started, with four digits for year, two digits
for month and two digits for day. The file is rotated on a daily basis.
</para>
<note><para>
When running Kea servers for both DHCPv4 and DHCPv6 the log names must
be distinct. See the examples in <xref linkend="forensic-log-configuration"/>.
</para></note>
</section>
<section>
<title>DHCPv4 Log Entries</title>
<para>
For DHCPv4 the library creates entries based on DHCPREQUEST messages
and corresponding DHCPv4 leases intercepted by lease4_select
(for new leases) and lease4_renew (for renewed leases) hooks.
</para>
<para>
An entry is a single string with no embedded end-of-line markers
and has the following sections:
<screen>
address duration device-id {client-info} {relay-info}
</screen>
</para>
<para>
Where:
<itemizedlist>
<listitem><para>
address - the leased IPv4 address given out and whether it was
assigned or renewed.
</para></listitem>
<listitem><para>
duration - the lease lifetime expressed in days (if present),
hours, minutes and seconds. A lease lifetime of 0xFFFFFFFF will be
denoted with the text "infinite duration".
</para></listitem>
<listitem><para>
device-id - the client's hardware address shown as numerical type
and hex digit string.
</para></listitem>
<listitem><para>
client-info - the DHCP client id option (61) if present, shown as
a hex string.
</para></listitem>
<listitem><para>
relay-info - for relayed packets the giaddr and the RAI circuit-id,
remote-id and subscriber-id options (option 82 sub
options: 1, 2 and 6) if present.
The circuit id and remote id are presented as hex strings
</para></listitem>
</itemizedlist>
</para>
<para>
For instance (line breaks added for readability, they would not
be present in the log file).
<screen>
Address: 192.2.1.100 has been renewed for 1 hrs 52 min 15 secs to a device with
hardware address: hwtype=1 08:00:2b:02:3f:4e, client-id: 17:34:e2:ff:09:92:54
connected via relay at address: 192.2.16.33, identified by circuit-id:
68:6f:77:64:79 and remote-id: 87:f6:79:77:ef
</screen>
</para>
</section>
<section>
<title>DHCPv6 Log Entries</title>
<para>
For DHCPv6 the library creates entries based on lease management
actions intercepted by the lease6_select (for new leases), lease6_renew
(for renewed leases) and lease6_rebind (for rebound leases).
</para>
<para>
An entry is a single string with no embedded end-of-line markers
and has the following sections:
<screen>
address duration device-id {relay-info}*
</screen>
</para>
<para>
Where:
<itemizedlist>
<listitem><para>
address - the leased IPv6 address or prefix given out and whether
it was assigned or renewed.
</para></listitem>
<listitem><para>
duration - the lease lifetime expressed in days (if present),
hours, minutes and seconds. A lease lifetime of 0xFFFFFFFF will be
denoted with the text "infinite duration".
</para></listitem>
<listitem><para>
device-id - the client's DUID and hardware address (if present).
</para></listitem>
<listitem><para>
relay-info - for relayed packets the content of relay agent
messages, remote-id (code 37), subscriber-id (code 38) and
interface-id (code 18) options if present. Note that
interface-id option, if present, identifies the whole interface the
relay agent received the message on. This typically translates to a
single link in your network, but it depends on your specific network
topology. Nevertheless, this is useful information to better scope
down the location of the device, so it is being recorded, if present.
</para></listitem>
</itemizedlist>
</para>
<para>
For instance (line breaks added for readability, they would not
be present in the log file).
<screen>
Address:2001:db8:1:: has been assigned for 0 hrs 11 mins 53 secs to a device with
DUID: 17:34:e2:ff:09:92:54 and hardware address: hwtype=1 08:00:2b:02:3f:4e
(from Raw Socket) connected via relay at address: fe80::abcd for client on
link address: 3001::1, hop count: 1, identified by remote-id:
01:02:03:04:0a:0b:0c:0d:0e:0f and subscriber-id: 1a:2b:3c:4d:5e:6f
</screen>
</para>
</section>
<section id="forensic-log-configuration">
<title>Configuring the Forensic Log Hooks</title>
<para>
To use this functionality the hook library must be included in the
configuration of the desired DHCP server modules. The legal_log
library is installed alongside the Kea libraries in
<filename>[kea-install-dir]/lib</filename> where
<filename>kea-install-dir</filename> is determined by the
"--prefix" option of the configure script. It defaults to
<filename>/usr/local</filename>. Assuming the
default value then, configuring kea-dhcp4 to load the legal_log
library could be done with the following Kea4 configuration:
<screen>
"Dhcp4": { <userinput>
"hooks-libraries": [
{
"library": "/usr/local/lib/libdhcp_legal_log.so",
"parameters": {
"path": "/var/kea/var",
"base-name": "kea-forensic4"
}
},
...
] </userinput>
}
</screen>
</para>
<para>
To configure it for kea-dhcp6, the commands are simply as shown below:
<screen>
"Dhcp6": { <userinput>
"hooks-libraries": [
{
"library": "/usr/local/lib/libdhcp_legal_log.so",
"parameters": {
"path": "/var/kea/var",
"base-name": "kea-forensic6"
}
},
...
] </userinput>
}
</screen>
</para>
<para>
Two Hook Library parameters are supported:
<itemizedlist>
<listitem><para>
path - the directory in which the forensic file(s) will be written. The
default value is
<filename>[prefix]/kea/var</filename>. The directory must exist.
</para></listitem>
<listitem><para>
base-name - an arbitrary value which is used in conjunction with
the current system date to form the current forensic file name. It defaults
to <filename>kea-legal</filename>.
</para></listitem>
</itemizedlist>
</para>
</section>
</section>
<section>
<title>flex_id: Flexible Identifiers for Host Reservations</title>
<para>
This section describes a hook application dedicated to generate
flexible identifiers for host reservation. Kea software provides a way
to handle host reservations that include addresses, prefixes, options,
client classes and other features. The reservation can be based on
hardware address, DUID, circuit-id or client-id in DHCPv4 and using
hardware address or DUID in DHCPv6. However, there are sometimes
scenario where the reservation is more complex, e.g. uses other
options that mentioned above, uses part of specific options or perhaps
even a combination of several options and fields to uniquely identify
a client. Those scenarios are addressed by the Flexible Identifiers
hook application.</para>
<para>The library allows defining an expression, using notation
initially used for client classification only. See <xref
linkend="classification-using-expressions" /> for detailed description
of the syntax available. One notable difference is that for client
classification the expression currently has to evaluate to either true
or false, while the flexible identifier expression is expected to
evaluate to a string that will be used as identifier. It is a valid case
for the expression to evaluate to empty string (e.g. in cases where a
client does not sent specific options). This expression is then
evaluated for each incoming packet. This evaluation generates an
identifier that is used to identify the client. In particular, there may
be host reservations that are tied to specific values of the flexible
identifier.</para>
<para>
The library can be loaded in similar way as other hook libraries. It
takes one mandatory parameter identifier-expression:
<screen>
"Dhcp6": { <userinput>
"hooks-libraries": [
{
"library": "/path/libdhcp_flex_id.so",
"parameters": {
"identifier-expression": "<userinput>expression</userinput>"
}
},
...
] </userinput>
}
</screen>
</para>
<para>
The flexible identifier library supports both DHCPv4 and DHCPv6.
</para>
<para>
EXAMPLE: Let's consider a case of an IPv6 network that has an
independent interface for each of the connected customers. Customers
are able to plug in whatever device they want, so any type of
identifier (e.g. a client-id) is unreliable. Therefore the operator
may decide to use an option inserted by a relay agent to differentiate
between clients. In this particular deployment, the operator verified
that the interface-id is unique for each customer facing
interface. Therefore it is suitable for usage as reservation. However,
only the first 6 bytes of the interface-id are interesting, because
remaining bytes are either randomly changed or not unique between
devices. Therefore the customer decided to use first 6 bytes of the
interface-id option inserted by the relay agent. This could be
achieved by using the following configuration:
<screen>
"Dhcp6": {
"subnet6": [{ ..., // subnet definition starts here
"reservations": [
<userinput>"flex-id": "'port1234'"</userinput>, // value of the first 8 bytes of the interface-id
"ip-addresses": [ "2001:db8::1" ]
],
}], // end of subnet definitions
"hooks-libraries": [
{
"library": "/path/libdhcp_flex_id.so",
"parameters": {
"identifier-expression": "<userinput>substring(relay6[0].option[18],0,8)</userinput>"
}
},
...
]
}
</screen>
</para>
<para>
NOTE: Care should be taken when adjusting the expression. If the
expression changes, then all the flex-id values may change, possibly
rendering all reservations based on flex-id unusable until they're
manually updated. Therefore it is strongly recommended to start with
the expression and a handful reservations, adjust the expression as
needed and only after it was confirmed the expression does exactly
what is expected out of it go forward with host reservations on any
broader scale.
</para>
<para>
flex-id values in host reservations can be specified in two
ways. First, they can be expressed as hex string, e.g. bar string
can be represented as 626174. Alternatively, it can be expressed
as quoted value (using double and single quotes), e.g. "'bar'".
The former is more convenient for printable characters, while hex
string values are more convenient for non-printable characters.
</para>
</section>
</section>
<section id="user-context">
<title>User contexts</title>
<para>Hook libraries can have their own configuration parameters. That is
convenient if the parameter applies to the whole library. However,
sometimes it is very useful if certain configuration entities are extended
with additional configuration data. This is where the concept of user
contexts comes in. A sysadmin can define an arbitrary set of data and
attach it to Kea structures, as long as the data is specified as JSON map.
In particular, it is possible to define fields that are integers, strings,
boolean, lists and maps. It is possible to define nested structures of
arbitrary complexity. Kea does not use that data on its own, simply stores
and makes it available for the hook libraries.
</para>
<para>
As of Kea 1.2, the only structures that allow user contexts are address
and prefix pools, but it is expected to extend other structures with the
user context capability.
</para>
</section>
</chapter> <!-- hooks-libraries -->