doc/arm/libdns.xml

<?xml version="1.0" encoding="utf-8"?>
<!--
 - Copyright (C) 2010, 2014  Internet Systems Consortium, Inc. ("ISC")
 -
 - Permission to use, copy, modify, and/or distribute this software for any
 - purpose with or without fee is hereby granted, provided that the above
 - copyright notice and this permission notice appear in all copies.
 -
 - THE SOFTWARE IS PROVIDED "AS IS" AND ISC DISCLAIMS ALL WARRANTIES WITH
 - REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY
 - AND FITNESS.  IN NO EVENT SHALL ISC BE LIABLE FOR ANY SPECIAL, DIRECT,
 - INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM
 - LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE
 - OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
 - PERFORMANCE OF THIS SOFTWARE.
-->

<sect1 id="bind9.library">
  <title>BIND 9 DNS Library Support</title>
  <para>This version of BIND 9 "exports" its internal libraries so
  that they can be used by third-party applications more easily (we
  call them "export" libraries in this document). In addition to
  all major DNS-related APIs BIND 9 is currently using, the export
  libraries provide the following features:</para>
  <itemizedlist>
    <listitem>
      <para>The newly created "DNS client" module. This is a higher
      level API that provides an interface to name resolution,
      single DNS transaction with a particular server, and dynamic
      update. Regarding name resolution, it supports advanced
      features such as DNSSEC validation and caching. This module
      supports both synchronous and asynchronous mode.</para>
    </listitem>
    <listitem>
      <para>The new "IRS" (Information Retrieval System) library.
      It provides an interface to parse the traditional resolv.conf
      file and more advanced, DNS-specific configuration file for
      the rest of this package (see the description for the
      dns.conf file below).</para>
    </listitem>
    <listitem>
      <para>As part of the IRS library, newly implemented standard
      address-name mapping functions, getaddrinfo() and
      getnameinfo(), are provided. They use the DNSSEC-aware
      validating resolver backend, and could use other advanced
      features of the BIND 9 libraries such as caching. The
      getaddrinfo() function resolves both A and AAAA RRs
      concurrently (when the address family is unspecified).</para>
    </listitem>
    <listitem>
      <para>An experimental framework to support other event
      libraries than BIND 9's internal event task system.</para>
    </listitem>
  </itemizedlist>
  <sect2>
    <title>Prerequisite</title>
  <para>GNU make is required to build the export libraries (other
  part of BIND 9 can still be built with other types of make). In
  the reminder of this document, "make" means GNU make. Note that
  in some platforms you may need to invoke a different command name
  than "make" (e.g. "gmake") to indicate it's GNU make.</para>
  </sect2>
  <sect2>
    <title>Compilation</title>
  <screen>
$ <userinput>./configure --enable-exportlib <replaceable>[other flags]</replaceable></userinput>
$ <userinput>make</userinput>
</screen>
  <para>
  This will create (in addition to usual BIND 9 programs) and a
  separate set of libraries under the lib/export directory. For
  example, <filename>lib/export/dns/libdns.a</filename> is the archive file of the
  export version of the BIND 9 DNS library. Sample application
  programs using the libraries will also be built under the
  lib/export/samples directory (see below).</para>
  </sect2>
  <sect2>
    <title>Installation</title>
  <screen>
$ <userinput>cd lib/export</userinput>
$ <userinput>make install</userinput>
</screen>
  <para>
  This will install library object files under the directory
  specified by the --with-export-libdir configure option (default:
  EPREFIX/lib/bind9), and header files under the directory
  specified by the --with-export-includedir configure option
  (default: PREFIX/include/bind9).
  Root privilege is normally required.
  "<command>make install</command>" at the top directory will do the
  same.
  </para>
  <para>
  To see how to build your own
  application after the installation, see
  <filename>lib/export/samples/Makefile-postinstall.in</filename>.</para>
  </sect2>
  <sect2>
    <title>Known Defects/Restrictions</title>
  <itemizedlist>
    <listitem>
<!-- TODO: what about AIX? -->
      <para>Currently, win32 is not supported for the export
      library. (Normal BIND 9 application can be built as
      before).</para>
    </listitem>
    <listitem>
      <para>The "fixed" RRset order is not (currently) supported in
      the export library. If you want to use "fixed" RRset order
      for, e.g. <command>named</command> while still building the
      export library even without the fixed order support, build
      them separately:
      <screen>
$ <userinput>./configure --enable-fixed-rrset <replaceable>[other flags, but not --enable-exportlib]</replaceable></userinput>
$ <userinput>make</userinput>
$ <userinput>./configure --enable-exportlib <replaceable>[other flags, but not --enable-fixed-rrset]</replaceable></userinput>
$ <userinput>cd lib/export</userinput>
$ <userinput>make</userinput>
</screen>
    </para>
    </listitem>
    <listitem>
      <para>The client module and the IRS library currently do not
      support DNSSEC validation using DLV (the underlying modules
      can handle it, but there is no tunable interface to enable
      the feature).</para>
    </listitem>
    <listitem>
      <para>RFC 5011 is not supported in the validating stub
      resolver of the export library. In fact, it is not clear
      whether it should: trust anchors would be a system-wide
      configuration which would be managed by an administrator,
      while the stub resolver will be used by ordinary applications
      run by a normal user.</para>
    </listitem>
    <listitem>
      <para>Not all common <filename>/etc/resolv.conf</filename>
      options are supported
      in the IRS library. The only available options in this
      version are "debug" and "ndots".</para>
    </listitem>
  </itemizedlist>
  </sect2>
  <sect2>
    <title>The dns.conf File</title>
  <para>The IRS library supports an "advanced" configuration file
  related to the DNS library for configuration parameters that
  would be beyond the capability of the
  <filename>resolv.conf</filename> file.
  Specifically, it is intended to provide DNSSEC related
  configuration parameters. By default the path to this
  configuration file is <filename>/etc/dns.conf</filename>.
  This module is very
  experimental and the configuration syntax or library interfaces
  may change in future versions. Currently, only the
  <command>trusted-keys</command>
  statement is supported, whose syntax is the same as the same name
  of statement for <filename>named.conf</filename>. (See
  <xref linkend="trusted-keys" /> for details.)</para>
  </sect2>
  <sect2>
    <title>Sample Applications</title>
  <para>Some sample application programs using this API are
  provided for reference. The following is a brief description of
  these applications.
  </para>
  <sect3>
    <title>sample: a simple stub resolver utility</title>
  <para>
  It sends a query of a given name (of a given optional RR type) to a
  specified recursive server, and prints the result as a list of
  RRs. It can also act as a validating stub resolver if a trust
  anchor is given via a set of command line options.</para>
  <para>
  Usage: sample [options] server_address hostname
  </para>
  <para>
  Options and Arguments:
  </para>
  <variablelist>
  <varlistentry>
  <term>
  -t RRtype
  </term>
  <listitem><para>
        specify the RR type of the query.  The default is the A RR.
  </para></listitem>
  </varlistentry>
  <varlistentry>
  <term>
  [-a algorithm] [-e] -k keyname -K keystring
  </term>
  <listitem><para>
        specify a command-line DNS key to validate the answer.  For
        example, to specify the following DNSKEY of example.com:
<literallayout>
                example.com. 3600 IN DNSKEY 257 3 5 xxx
</literallayout>
        specify the options as follows:
<screen>
<userinput>
          -e -k example.com -K "xxx"
</userinput>
</screen>
        -e means that this key is a zone's "key signing key" (as known
        as "secure Entry point").
        When -a is omitted rsasha1 will be used by default.
  </para></listitem>
  </varlistentry>
  <varlistentry>
  <term>
  -s domain:alt_server_address
  </term>
  <listitem><para>
         specify a separate recursive server address for the specific
        "domain".  Example: -s example.com:2001:db8::1234
  </para></listitem>
  </varlistentry>
  <varlistentry>
  <term>server_address</term>
  <listitem><para>
        an IP(v4/v6) address of the recursive server to which queries
        are sent.
  </para></listitem>
  </varlistentry>
  <varlistentry>
  <term>hostname</term>
  <listitem><para>
        the domain name for the query
  </para></listitem>
  </varlistentry>
  </variablelist>
  </sect3>
  <sect3>
    <title>sample-async: a simple stub resolver, working asynchronously</title>
  <para>
  Similar to "sample", but accepts a list
  of (query) domain names as a separate file and resolves the names
  asynchronously.</para>
  <para>
    Usage: sample-async [-s server_address] [-t RR_type] input_file</para>
  <para>
 Options and Arguments:
  </para>
  <variablelist>
  <varlistentry>
   <term>
   -s server_address
   </term>
  <listitem>
   an IPv4 address of the recursive server to which queries are sent.
  (IPv6 addresses are not supported in this implementation)
  </listitem>
  </varlistentry>
  <varlistentry>
  <term>
   -t RR_type
  </term>
  <listitem>
  specify the RR type of the queries. The default is the A
  RR.
  </listitem>
  </varlistentry>
  <varlistentry>
  <term>
   input_file
  </term>
  <listitem>
   a list of domain names to be resolved. each line
  consists of a single domain name. Example:
  <literallayout>
  www.example.com
  mx.example.net
  ns.xxx.example
</literallayout>
  </listitem>
    </varlistentry>
    </variablelist>
  </sect3>
  <sect3>
    <title>sample-request: a simple DNS transaction client</title>
  <para>
  It sends a query to a specified server, and
  prints the response with minimal processing. It doesn't act as a
  "stub resolver": it stops the processing once it gets any
  response from the server, whether it's a referral or an alias
  (CNAME or DNAME) that would require further queries to get the
  ultimate answer. In other words, this utility acts as a very
  simplified <command>dig</command>.
  </para>
  <para>
  Usage: sample-request [-t RRtype] server_address hostname
  </para>
  <para>
    Options and Arguments:
  </para>
  <variablelist>
  <varlistentry>
   <term>
   -t RRtype
  </term>
  <listitem>
  <para>
  specify the RR type of
  the queries. The default is the A RR.
  </para>
  </listitem>
  </varlistentry>
  <varlistentry>
  <term>
  server_address
  </term>
  <listitem>
  <para>
   an IP(v4/v6)
  address of the recursive server to which the query is sent.
  </para>
  </listitem>
  </varlistentry>
  <varlistentry>
  <term>
  hostname
  </term>
  <listitem>
  <para>
  the domain name for the query
  </para>
  </listitem>
  </varlistentry>
  </variablelist>
  </sect3>
  <sect3>
    <title>sample-gai: getaddrinfo() and getnameinfo() test code</title>
  <para>
  This is a test program
  to check getaddrinfo() and getnameinfo() behavior. It takes a
  host name as an argument, calls getaddrinfo() with the given host
  name, and calls getnameinfo() with the resulting IP addresses
  returned by getaddrinfo(). If the dns.conf file exists and
  defines a trust anchor, the underlying resolver will act as a
  validating resolver, and getaddrinfo()/getnameinfo() will fail
  with an EAI_INSECUREDATA error when DNSSEC validation fails.
  </para>
  <para>
  Usage: sample-gai hostname
  </para>
  </sect3>
  <sect3>
    <title>sample-update: a simple dynamic update client program</title>
  <para>
  It accepts a single update command as a
  command-line argument, sends an update request message to the
  authoritative server, and shows the response from the server. In
  other words, this is a simplified <command>nsupdate</command>.
  </para>
  <para>
   Usage: sample-update [options] (add|delete) "update data"
  </para>
  <para>
  Options and Arguments:
  </para>
  <variablelist>
  <varlistentry>
   <term>
  -a auth_server
   </term>
   <listitem><para>
        An IP address of the authoritative server that has authority
        for the zone containing the update name.  This should normally
        be the primary authoritative server that accepts dynamic
        updates.  It can also be a secondary server that is configured
        to forward update requests to the primary server.
   </para></listitem>
   </varlistentry>
   <varlistentry>
   <term>
  -k keyfile
   </term>
   <listitem><para>
        A TSIG key file to secure the update transaction.  The keyfile
        format is the same as that for the nsupdate utility.
   </para></listitem>
   </varlistentry>
   <varlistentry>
   <term>
  -p prerequisite
   </term>
   <listitem><para>
        A prerequisite for the update (only one prerequisite can be
        specified).  The prerequisite format is the same as that is
        accepted by the nsupdate utility.
   </para></listitem>
   </varlistentry>
   <varlistentry>
   <term>
  -r recursive_server
   </term>
   <listitem><para>
        An IP address of a recursive server that this utility will
        use.  A recursive server may be necessary to identify the
        authoritative server address to which the update request is
        sent.
   </para></listitem>
   </varlistentry>
   <varlistentry>
   <term>
  -z zonename
   </term>
   <listitem><para>
        The domain name of the zone that contains
   </para></listitem>
   </varlistentry>
   <varlistentry>
   <term>
  (add|delete)
   </term>
   <listitem><para>
        Specify the type of update operation.  Either "add" or "delete"
        must be specified.
   </para></listitem>
   </varlistentry>
   <varlistentry>
   <term>
  "update data"
   </term>
   <listitem><para>
        Specify the data to be updated.  A typical example of the data
        would look like "name TTL RRtype RDATA".
  </para></listitem>
  </varlistentry>
  </variablelist>

   <note>In practice, either -a or -r must be specified.  Others can
   be optional; the underlying library routine tries to identify the
   appropriate server and the zone name for the update.</note>

   <para>
   Examples: assuming the primary authoritative server of the
   dynamic.example.com zone has an IPv6 address 2001:db8::1234,
   </para>
   <screen>
$ <userinput>sample-update -a sample-update -k Kxxx.+nnn+mmmm.key add "foo.dynamic.example.com 30 IN A 192.168.2.1"</userinput></screen>
   <para>
     adds an A RR for foo.dynamic.example.com using the given key.
   </para>
   <screen>
$ <userinput>sample-update -a sample-update -k Kxxx.+nnn+mmmm.key delete "foo.dynamic.example.com 30 IN A"</userinput></screen>
   <para>
     removes all A RRs for foo.dynamic.example.com using the given key.
   </para>
   <screen>
$ <userinput>sample-update -a sample-update -k Kxxx.+nnn+mmmm.key delete "foo.dynamic.example.com"</userinput></screen>
   <para>
     removes all RRs for foo.dynamic.example.com using the given key.
   </para>
  </sect3>
  <sect3>
    <title>nsprobe: domain/name server checker in terms of RFC 4074</title>
  <para>
  It checks a set
  of domains to see the name servers of the domains behave
  correctly in terms of RFC 4074. This is included in the set of
  sample programs to show how the export library can be used in a
  DNS-related application.
  </para>
  <para>
 Usage: nsprobe [-d] [-v [-v...]] [-c cache_address] [input_file]
  </para>
  <para>
   Options
  </para>

  <variablelist>
  <varlistentry>
  <term>
  -d
  </term>
  <listitem><para>
        run in the "debug" mode.  with this option nsprobe will dump
        every RRs it receives.
  </para></listitem>
  </varlistentry>
  <varlistentry>
  <term>
  -v
  </term>
  <listitem><para>
        increase verbosity of other normal log messages.  This can be
        specified multiple times
  </para></listitem>
  </varlistentry>
  <varlistentry>
  <term>
  -c cache_address
  </term>
  <listitem><para>
        specify an IP address of a recursive (caching) name server.
        nsprobe uses this server to get the NS RRset of each domain and
        the A and/or AAAA RRsets for the name servers.  The default
        value is 127.0.0.1.
  </para></listitem>
  </varlistentry>
  <varlistentry>
  <term>
  input_file
  </term>
  <listitem><para>
        a file name containing a list of domain (zone) names to be
        probed.  when omitted the standard input will be used.  Each
        line of the input file specifies a single domain name such as
        "example.com".  In general this domain name must be the apex
        name of some DNS zone (unlike normal "host names" such as
        "www.example.com").  nsprobe first identifies the NS RRsets for
        the given domain name, and sends A and AAAA queries to these
        servers for some "widely used" names under the zone;
        specifically, adding "www" and "ftp" to the zone name.
  </para></listitem>
  </varlistentry>
  </variablelist>
  </sect3>
  </sect2>
  <sect2>
    <title>Library References</title>
  <para>As of this writing, there is no formal "manual" of the
  libraries, except this document, header files (some of them
  provide pretty detailed explanations), and sample application
  programs.</para>
  </sect2>
</sect1>
<!-- Id: libdns.xml,v 1.3 2010/02/03 23:49:07 tbox Exp  -->