xref: /netbsd-src/external/mpl/bind/dist/doc/arm/reference.rst (revision 7d62b00eb9ad855ffcd7da46b41e23feb5476fac)
1.. Copyright (C) Internet Systems Consortium, Inc. ("ISC")
2..
3.. SPDX-License-Identifier: MPL-2.0
4..
5.. This Source Code Form is subject to the terms of the Mozilla Public
6.. License, v. 2.0.  If a copy of the MPL was not distributed with this
7.. file, you can obtain one at https://mozilla.org/MPL/2.0/.
8..
9.. See the COPYRIGHT file distributed with this work for additional
10.. information regarding copyright ownership.
11
12.. Reference:
13
14BIND 9 Configuration Reference
15==============================
16
17.. _configuration_file_elements:
18
19Configuration File Elements
20---------------------------
21
22Following is a list of elements used throughout the BIND configuration
23file documentation:
24
25.. glossary::
26
27    ``acl_name``
28        The name of an ``address_match_list`` as defined by the ``acl`` statement.
29
30    ``address_match_list``
31        A list of one or more ``ip_addr``, ``ip_prefix``, ``key_id``, or ``acl_name`` elements; see :ref:`address_match_lists`.
32
33    ``remoteserver_list``
34        A named list of one or more ``ip_addr`` with optional ``key_id`` and/or ``ip_port``. A ``remoteserver_list`` may include other ``remoteserver_list``.
35
36    ``domain_name``
37        A quoted string which is used as a DNS name; for example. ``my.test.domain``.
38
39    ``namelist``
40        A list of one or more ``domain_name`` elements.
41
42    ``dotted_decimal``
43        One to four integers valued 0 through 255 separated by dots (``.``), such as ``123.45.67`` or ``89.123.45.67``.
44
45    ``ip4_addr``
46        An IPv4 address with exactly four elements in ``dotted_decimal`` notation.
47
48    ``ip6_addr``
49        An IPv6 address, such as ``2001:db8::1234``. IPv6-scoped addresses that have ambiguity on their scope zones must be disambiguated by an appropriate zone ID with the percent character (``%``) as a delimiter. It is strongly recommended to use string zone names rather than numeric identifiers, to be robust against system configuration changes. However, since there is no standard mapping for such names and identifier values, only interface names as link identifiers are supported, assuming one-to-one mapping between interfaces and links. For example, a link-local address ``fe80::1`` on the link attached to the interface ``ne0`` can be specified as ``fe80::1%ne0``. Note that on most systems link-local addresses always have ambiguity and need to be disambiguated.
50
51    ``ip_addr``
52        An ``ip4_addr`` or ``ip6_addr``.
53
54    ``ip_dscp``
55        A ``number`` between 0 and 63, used to select a differentiated services code point (DSCP) value for use with outgoing traffic on operating systems that support DSCP.
56
57    ``ip_port``
58        An IP port ``number``. The ``number`` is limited to 0 through 65535, with values below 1024 typically restricted to use by processes running as root. In some cases, an asterisk (``*``) character can be used as a placeholder to select a random high-numbered port.
59
60    ``ip_prefix``
61        An IP network specified as an ``ip_addr``, followed by a slash (``/``) and then the number of bits in the netmask. Trailing zeros in an``ip_addr`` may be omitted. For example, ``127/8`` is the network ``127.0.0.0``with netmask ``255.0.0.0`` and ``1.2.3.0/28`` is network ``1.2.3.0`` with netmask ``255.255.255.240``.
62        When specifying a prefix involving a IPv6-scoped address, the scope may be omitted. In that case, the prefix matches packets from any scope.
63
64    ``key_id``
65        A ``domain_name`` representing the name of a shared key, to be used for transaction security.
66
67    ``key_list``
68        A list of one or more ``key_id``, separated by semicolons and ending with a semicolon.
69
70    ``number``
71        A non-negative 32-bit integer (i.e., a number between 0 and 4294967295, inclusive). Its acceptable value might be further limited by the context in which it is used.
72
73    ``fixedpoint``
74        A non-negative real number that can be specified to the nearest one-hundredth. Up to five digits can be specified before a decimal point, and up to two digits after, so the maximum value is 99999.99. Acceptable values might be further limited by the contexts in which they are used.
75
76    ``path_name``
77        A quoted string which is used as a pathname, such as ``zones/master/my.test.domain``.
78
79    ``port_list``
80        A list of an ``ip_port`` or a port range. A port range is specified in the form of ``range`` followed by two ``ip_port``s, ``port_low`` and ``port_high``, which represents port numbers from ``port_low`` through ``port_high``, inclusive. ``port_low`` must not be larger than ``port_high``. For example, ``range 1024 65535`` represents ports from 1024 through 65535. In either case an asterisk (``*``) character is not allowed as a valid ``ip_port``.
81
82    ``size_spec``
83        A 64-bit unsigned integer, or the keywords ``unlimited`` or ``default``. Integers may take values 0 <= value <= 18446744073709551615, though certain parameters (such as ``max-journal-size``) may use a more limited range within these extremes. In most cases, setting a value to 0 does not literally mean zero; it means "undefined" or "as big as possible," depending on the context. See the explanations of particular parameters that use ``size_spec`` for details on how they interpret its use. Numeric values can optionally be followed by a scaling factor: ``K`` or ``k`` for kilobytes, ``M`` or ``m`` for megabytes, and ``G`` or ``g`` for gigabytes, which scale by 1024, 1024*1024, and 1024*1024*1024 respectively.
84        ``unlimited`` generally means "as big as possible," and is usually the best way to safely set a very large number.
85        ``default`` uses the limit that was in force when the server was started.
86
87    ``size_or_percent``
88         A ``size_spec`` or integer value followed by ``%`` to represent percent. The behavior is exactly the same as ``size_spec``, but ``size_or_percent`` also allows specifying a positive integer value followed by the ``%`` sign to represent percent.
89
90    ``yes_or_no``
91        Either ``yes`` or ``no``. The words ``true`` and ``false`` are also accepted, as are the numbers ``1`` and ``0``.
92
93    ``dialup_option``
94        One of ``yes``, ``no``, ``notify``, ``notify-passive``, ``refresh``, or  ``passive``. When used in a zone, ``notify-passive``, ``refresh``, and ``passive`` are restricted to secondary and stub zones.
95
96.. _address_match_lists:
97
98Address Match Lists
99~~~~~~~~~~~~~~~~~~~
100
101Syntax
102^^^^^^
103
104::
105
106   address_match_list = address_match_list_element ; ...
107
108   address_match_list_element = [ ! ] ( ip_address | ip_prefix |
109        key key_id | acl_name | { address_match_list } )
110
111Definition and Usage
112^^^^^^^^^^^^^^^^^^^^
113
114Address match lists are primarily used to determine access control for
115various server operations. They are also used in the ``listen-on`` and
116``sortlist`` statements. The elements which constitute an address match
117list can be any of the following:
118
119-  an IP address (IPv4 or IPv6)
120
121-  an IP prefix (in ``/`` notation)
122
123-  a key ID, as defined by the ``key`` statement
124
125-  the name of an address match list defined with the ``acl`` statement
126
127-  a nested address match list enclosed in braces
128
129Elements can be negated with a leading exclamation mark (``!``), and the
130match list names "any", "none", "localhost", and "localnets" are
131predefined. More information on those names can be found in the
132description of the ``acl`` statement.
133
134The addition of the key clause made the name of this syntactic element
135something of a misnomer, since security keys can be used to validate
136access without regard to a host or network address. Nonetheless, the
137term "address match list" is still used throughout the documentation.
138
139When a given IP address or prefix is compared to an address match list,
140the comparison takes place in approximately O(1) time. However, key
141comparisons require that the list of keys be traversed until a matching
142key is found, and therefore may be somewhat slower.
143
144The interpretation of a match depends on whether the list is being used
145for access control, defining ``listen-on`` ports, or in a ``sortlist``,
146and whether the element was negated.
147
148When used as an access control list, a non-negated match allows access
149and a negated match denies access. If there is no match, access is
150denied. The clauses ``allow-notify``, ``allow-recursion``,
151``allow-recursion-on``, ``allow-query``, ``allow-query-on``,
152``allow-query-cache``, ``allow-query-cache-on``, ``allow-transfer``,
153``allow-update``, ``allow-update-forwarding``, ``blackhole``, and
154``keep-response-order`` all use address match lists. Similarly, the
155``listen-on`` option causes the server to refuse queries on any of
156the machine's addresses which do not match the list.
157
158Order of insertion is significant. If more than one element in an ACL is
159found to match a given IP address or prefix, preference is given to
160the one that came *first* in the ACL definition. Because of this
161first-match behavior, an element that defines a subset of another
162element in the list should come before the broader element, regardless
163of whether either is negated. For example, in ``1.2.3/24; ! 1.2.3.13;``
164the 1.2.3.13 element is completely useless because the algorithm
165matches any lookup for 1.2.3.13 to the 1.2.3/24 element. Using
166``! 1.2.3.13; 1.2.3/24`` fixes that problem by blocking 1.2.3.13
167via the negation, but all other 1.2.3.\* hosts pass through.
168
169.. _comment_syntax:
170
171Comment Syntax
172~~~~~~~~~~~~~~
173
174The BIND 9 comment syntax allows comments to appear anywhere that
175whitespace may appear in a BIND configuration file. To appeal to
176programmers of all kinds, they can be written in the C, C++, or
177shell/perl style.
178
179Syntax
180^^^^^^
181
182::
183
184   /* This is a BIND comment as in C */
185
186::
187
188   // This is a BIND comment as in C++
189
190::
191
192   # This is a BIND comment as in common Unix shells
193   # and perl
194
195Definition and Usage
196^^^^^^^^^^^^^^^^^^^^
197
198Comments may appear anywhere that whitespace may appear in a BIND
199configuration file.
200
201C-style comments start with the two characters /\* (slash, star) and end
202with \*/ (star, slash). Because they are completely delimited with these
203characters, they can be used to comment only a portion of a line or to
204span multiple lines.
205
206C-style comments cannot be nested. For example, the following is not
207valid because the entire comment ends with the first \*/:
208
209::
210
211   /* This is the start of a comment.
212      This is still part of the comment.
213   /* This is an incorrect attempt at nesting a comment. */
214      This is no longer in any comment. */
215
216C++-style comments start with the two characters // (slash, slash) and
217continue to the end of the physical line. They cannot be continued
218across multiple physical lines; to have one logical comment span
219multiple lines, each line must use the // pair. For example:
220
221::
222
223   // This is the start of a comment.  The next line
224   // is a new comment, even though it is logically
225   // part of the previous comment.
226
227Shell-style (or perl-style) comments start with the
228character ``#`` (number sign) and continue to the end of the physical
229line, as in C++ comments. For example:
230
231::
232
233   # This is the start of a comment.  The next line
234   # is a new comment, even though it is logically
235   # part of the previous comment.
236
237..
238
239.. warning::
240
241   The semicolon (``;``) character cannot start a comment, unlike
242   in a zone file. The semicolon indicates the end of a
243   configuration statement.
244
245.. _Configuration_File_Grammar:
246
247Configuration File Grammar
248--------------------------
249
250A BIND 9 configuration consists of statements and comments. Statements
251end with a semicolon; statements and comments are the only elements that
252can appear without enclosing braces. Many statements contain a block of
253sub-statements, which are also terminated with a semicolon.
254
255The following statements are supported:
256
257    ``acl``
258        Defines a named IP address matching list, for access control and other uses.
259
260    ``controls``
261        Declares control channels to be used by the ``rndc`` utility.
262
263    ``dnssec-policy``
264        Describes a DNSSEC key and signing policy for zones. See :ref:`dnssec-policy Grammar <dnssec_policy_grammar>` for details.
265
266    ``include``
267        Includes a file.
268
269    ``key``
270        Specifies key information for use in authentication and authorization using TSIG.
271
272    ``logging``
273        Specifies what information the server logs and where the log messages are sent.
274
275    ``masters``
276        Synonym for ``primaries``.
277
278    ``options``
279        Controls global server configuration options and sets defaults for other statements.
280
281    ``parental-agents``
282        Defines a named list of servers for inclusion in primary and secondary zones' ``parental-agents`` lists.
283
284    ``primaries``
285        Defines a named list of servers for inclusion in stub and secondary zones' ``primaries`` or ``also-notify`` lists. (Note: this is a synonym for the original keyword ``masters``, which can still be used, but is no longer the preferred terminology.)
286
287    ``server``
288        Sets certain configuration options on a per-server basis.
289
290    ``statistics-channels``
291        Declares communication channels to get access to ``named`` statistics.
292
293    ``trust-anchors``
294        Defines DNSSEC trust anchors: if used with the ``initial-key`` or ``initial-ds`` keyword, trust anchors are kept up-to-date using :rfc:`5011` trust anchor maintenance; if used with ``static-key`` or ``static-ds``, keys are permanent.
295
296    ``managed-keys``
297        Is identical to ``trust-anchors``; this option is deprecated in favor of ``trust-anchors`` with the ``initial-key`` keyword, and may be removed in a future release.
298
299    ``trusted-keys``
300        Defines permanent trusted DNSSEC keys; this option is deprecated in favor of ``trust-anchors`` with the ``static-key`` keyword, and may be removed in a future release.
301
302    ``view``
303        Defines a view.
304
305    ``zone``
306        Defines a zone.
307
308The ``logging`` and ``options`` statements may only occur once per
309configuration.
310
311.. _acl_grammar:
312
313``acl`` Statement Grammar
314~~~~~~~~~~~~~~~~~~~~~~~~~
315
316.. include:: ../misc/acl.grammar.rst
317
318.. _acl:
319
320``acl`` Statement Definition and Usage
321~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
322
323The ``acl`` statement assigns a symbolic name to an address match list.
324It gets its name from one of the primary uses of address match lists: Access
325Control Lists (ACLs).
326
327The following ACLs are built-in:
328
329    ``any``
330        Matches all hosts.
331
332    ``none``
333        Matches no hosts.
334
335    ``localhost``
336        Matches the IPv4 and IPv6 addresses of all network interfaces on the system. When addresses are added or removed, the ``localhost`` ACL element is updated to reflect the changes.
337
338    ``localnets``
339        Matches any host on an IPv4 or IPv6 network for which the system has an interface. When addresses are added or removed, the ``localnets`` ACL element is updated to reflect the changes. Some systems do not provide a way to determine the prefix lengths of local IPv6 addresses; in such cases, ``localnets`` only matches the local IPv6 addresses, just like ``localhost``.
340
341.. _controls_grammar:
342
343``controls`` Statement Grammar
344~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
345
346.. include:: ../misc/controls.grammar.rst
347
348.. _controls_statement_definition_and_usage:
349
350``controls`` Statement Definition and Usage
351~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
352
353The ``controls`` statement declares control channels to be used by
354system administrators to manage the operation of the name server. These
355control channels are used by the ``rndc`` utility to send commands to
356and retrieve non-DNS results from a name server.
357
358An ``inet`` control channel is a TCP socket listening at the specified
359``ip_port`` on the specified ``ip_addr``, which can be an IPv4 or IPv6
360address. An ``ip_addr`` of ``*`` (asterisk) is interpreted as the IPv4
361wildcard address; connections are accepted on any of the system's
362IPv4 addresses. To listen on the IPv6 wildcard address, use an
363``ip_addr`` of ``::``. If ``rndc`` is only used on the local host,
364using the loopback address (``127.0.0.1`` or ``::1``) is recommended for
365maximum security.
366
367If no port is specified, port 953 is used. The asterisk ``*`` cannot
368be used for ``ip_port``.
369
370The ability to issue commands over the control channel is restricted by
371the ``allow`` and ``keys`` clauses. Connections to the control channel
372are permitted based on the ``address_match_list``. This is for simple IP
373address-based filtering only; any ``key_id`` elements of the
374``address_match_list`` are ignored.
375
376A ``unix`` control channel is a Unix domain socket listening at the
377specified path in the file system. Access to the socket is specified by
378the ``perm``, ``owner``, and ``group`` clauses. Note that on some platforms
379(SunOS and Solaris), the permissions (``perm``) are applied to the parent
380directory as the permissions on the socket itself are ignored.
381
382The primary authorization mechanism of the command channel is the
383``key_list``, which contains a list of ``key_id``s. Each ``key_id`` in
384the ``key_list`` is authorized to execute commands over the control
385channel. See :ref:`admin_tools` for information about
386configuring keys in ``rndc``.
387
388If the ``read-only`` clause is enabled, the control channel is limited
389to the following set of read-only commands: ``nta -dump``, ``null``,
390``status``, ``showzone``, ``testgen``, and ``zonestatus``. By default,
391``read-only`` is not enabled and the control channel allows read-write
392access.
393
394If no ``controls`` statement is present, ``named`` sets up a default
395control channel listening on the loopback address 127.0.0.1 and its IPv6
396counterpart, ::1. In this case, and also when the ``controls`` statement
397is present but does not have a ``keys`` clause, ``named`` attempts
398to load the command channel key from the file ``rndc.key`` in ``/etc``
399(or whatever ``sysconfdir`` was specified when BIND was built). To
400create an ``rndc.key`` file, run ``rndc-confgen -a``.
401
402To disable the command channel, use an empty ``controls`` statement:
403``controls { };``.
404
405.. _include_grammar:
406
407``include`` Statement Grammar
408~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
409
410::
411
412   include filename;
413
414.. _include_statement:
415
416``include`` Statement Definition and Usage
417~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
418
419The ``include`` statement inserts the specified file (or files if a valid glob
420expression is detected) at the point where the ``include`` statement is
421encountered. The ``include`` statement facilitates the administration of
422configuration files by permitting the reading or writing of some things but not
423others. For example, the statement could include private keys that are readable
424only by the name server.
425
426.. _key_grammar:
427
428``key`` Statement Grammar
429~~~~~~~~~~~~~~~~~~~~~~~~~
430
431.. include:: ../misc/key.grammar.rst
432
433.. _key_statement:
434
435``key`` Statement Definition and Usage
436~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
437
438The ``key`` statement defines a shared secret key for use with TSIG (see
439:ref:`tsig`) or the command channel (see :ref:`controls_statement_definition_and_usage`).
440
441The ``key`` statement can occur at the top level of the configuration
442file or inside a ``view`` statement. Keys defined in top-level ``key``
443statements can be used in all views. Keys intended for use in a
444``controls`` statement (see :ref:`controls_statement_definition_and_usage`)
445must be defined at the top level.
446
447The ``key_id``, also known as the key name, is a domain name that uniquely
448identifies the key. It can be used in a ``server`` statement to cause
449requests sent to that server to be signed with this key, or in address
450match lists to verify that incoming requests have been signed with a key
451matching this name, algorithm, and secret.
452
453The ``algorithm_id`` is a string that specifies a security/authentication
454algorithm. The ``named`` server supports ``hmac-md5``, ``hmac-sha1``,
455``hmac-sha224``, ``hmac-sha256``, ``hmac-sha384``, and ``hmac-sha512``
456TSIG authentication. Truncated hashes are supported by appending the
457minimum number of required bits preceded by a dash, e.g.,
458``hmac-sha1-80``. The ``secret_string`` is the secret to be used by the
459algorithm, and is treated as a Base64-encoded string.
460
461.. _logging_grammar:
462
463``logging`` Statement Grammar
464~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
465
466.. include:: ../misc/logging.grammar.rst
467
468.. _logging_statement:
469
470``logging`` Statement Definition and Usage
471~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
472
473The ``logging`` statement configures a wide variety of logging options
474for the name server. Its ``channel`` phrase associates output methods,
475format options, and severity levels with a name that can then be used
476with the ``category`` phrase to select how various classes of messages
477are logged.
478
479Only one ``logging`` statement is used to define as many channels and
480categories as desired. If there is no ``logging`` statement, the
481logging configuration is:
482
483::
484
485   logging {
486        category default { default_syslog; default_debug; };
487        category unmatched { null; };
488   };
489
490If ``named`` is started with the ``-L`` option, it logs to the specified
491file at startup, instead of using syslog. In this case the logging
492configuration is:
493
494::
495
496   logging {
497        category default { default_logfile; default_debug; };
498        category unmatched { null; };
499   };
500
501The logging configuration is only established when the entire
502configuration file has been parsed. When the server starts up, all
503logging messages regarding syntax errors in the configuration file go to
504the default channels, or to standard error if the ``-g`` option was
505specified.
506
507.. _channel:
508
509The ``channel`` Phrase
510^^^^^^^^^^^^^^^^^^^^^^
511
512All log output goes to one or more ``channels``; there is no limit to
513the number of channels that can be created.
514
515Every channel definition must include a destination clause that says
516whether messages selected for the channel go to a file, go to a particular
517syslog facility, go to the standard error stream, or are discarded. The definition can
518optionally also limit the message severity level that is accepted
519by the channel (the default is ``info``), and whether to include a
520``named``-generated time stamp, the category name, and/or the severity level
521(the default is not to include any).
522
523The ``null`` destination clause causes all messages sent to the channel
524to be discarded; in that case, other options for the channel are
525meaningless.
526
527The ``file`` destination clause directs the channel to a disk file. It
528can include additional arguments to specify how large the file is
529allowed to become before it is rolled to a backup file (``size``), how
530many backup versions of the file are saved each time this happens
531(``versions``), and the format to use for naming backup versions
532(``suffix``).
533
534The ``size`` option is used to limit log file growth. If the file ever
535exceeds the specified size, then ``named`` stops writing to the file
536unless it has a ``versions`` option associated with it. If backup
537versions are kept, the files are rolled as described below. If there is
538no ``versions`` option, no more data is written to the log until
539some out-of-band mechanism removes or truncates the log to less than the
540maximum size. The default behavior is not to limit the size of the file.
541
542File rolling only occurs when the file exceeds the size specified with
543the ``size`` option. No backup versions are kept by default; any
544existing log file is simply appended. The ``versions`` option specifies
545how many backup versions of the file should be kept. If set to
546``unlimited``, there is no limit.
547
548The ``suffix`` option can be set to either ``increment`` or
549``timestamp``. If set to ``timestamp``, then when a log file is rolled,
550it is saved with the current timestamp as a file suffix. If set to
551``increment``, then backup files are saved with incrementing numbers as
552suffixes; older files are renamed when rolling. For example, if
553``versions`` is set to 3 and ``suffix`` to ``increment``, then when
554``filename.log`` reaches the size specified by ``size``,
555``filename.log.1`` is renamed to ``filename.log.2``, ``filename.log.0``
556is renamed to ``filename.log.1``, and ``filename.log`` is renamed to
557``filename.log.0``, whereupon a new ``filename.log`` is opened.
558
559Here is an example using the ``size``, ``versions``, and ``suffix`` options:
560
561::
562
563   channel an_example_channel {
564       file "example.log" versions 3 size 20m suffix increment;
565       print-time yes;
566       print-category yes;
567   };
568
569The ``syslog`` destination clause directs the channel to the system log.
570Its argument is a syslog facility as described in the ``syslog`` man
571page. Known facilities are ``kern``, ``user``, ``mail``, ``daemon``,
572``auth``, ``syslog``, ``lpr``, ``news``, ``uucp``, ``cron``,
573``authpriv``, ``ftp``, ``local0``, ``local1``, ``local2``, ``local3``,
574``local4``, ``local5``, ``local6``, and ``local7``; however, not all
575facilities are supported on all operating systems. How ``syslog``
576handles messages sent to this facility is described in the
577``syslog.conf`` man page. On a system which uses a very old
578version of ``syslog``, which only uses two arguments to the ``openlog()``
579function, this clause is silently ignored.
580
581On Windows machines, syslog messages are directed to the EventViewer.
582
583The ``severity`` clause works like ``syslog``'s "priorities," except
584that they can also be used when writing straight to a file rather
585than using ``syslog``. Messages which are not at least of the severity
586level given are not selected for the channel; messages of higher
587severity levels are accepted.
588
589When using ``syslog``, the ``syslog.conf`` priorities
590also determine what eventually passes through. For example, defining a
591channel facility and severity as ``daemon`` and ``debug``, but only
592logging ``daemon.warning`` via ``syslog.conf``, causes messages of
593severity ``info`` and ``notice`` to be dropped. If the situation were
594reversed, with ``named`` writing messages of only ``warning`` or higher,
595then ``syslogd`` would print all messages it received from the channel.
596
597The ``stderr`` destination clause directs the channel to the server's
598standard error stream. This is intended for use when the server is
599running as a foreground process, as when debugging a
600configuration, for example.
601
602The server can supply extensive debugging information when it is in
603debugging mode. If the server's global debug level is greater than zero,
604debugging mode is active. The global debug level is set either
605by starting the ``named`` server with the ``-d`` flag followed by a
606positive integer, or by running ``rndc trace``. The global debug level
607can be set to zero, and debugging mode turned off, by running ``rndc
608notrace``. All debugging messages in the server have a debug level;
609higher debug levels give more detailed output. Channels that specify a
610specific debug severity, for example:
611
612::
613
614   channel specific_debug_level {
615       file "foo";
616       severity debug 3;
617   };
618
619get debugging output of level 3 or less any time the server is in
620debugging mode, regardless of the global debugging level. Channels with
621``dynamic`` severity use the server's global debug level to determine
622what messages to print.
623
624``print-time`` can be set to ``yes``, ``no``, or a time format
625specifier, which may be one of ``local``, ``iso8601``, or
626``iso8601-utc``. If set to ``no``, the date and time are not
627logged. If set to ``yes`` or ``local``, the date and time are logged in
628a human-readable format, using the local time zone. If set to
629``iso8601``, the local time is logged in ISO 8601 format. If set to
630``iso8601-utc``, the date and time are logged in ISO 8601 format,
631with time zone set to UTC. The default is ``no``.
632
633``print-time`` may be specified for a ``syslog`` channel, but it is
634usually pointless since ``syslog`` also logs the date and time.
635
636If ``print-category`` is requested, then the category of the message
637is logged as well. Finally, if ``print-severity`` is on, then the
638severity level of the message is logged. The ``print-`` options may
639be used in any combination, and are always printed in the following
640order: time, category, severity. Here is an example where all three
641``print-`` options are on:
642
643``28-Feb-2000 15:05:32.863 general: notice: running``
644
645If ``buffered`` has been turned on, the output to files is not
646flushed after each log entry. By default all log messages are flushed.
647
648There are four predefined channels that are used for ``named``'s default
649logging, as follows. If ``named`` is started with the ``-L`` option, then a fifth
650channel, ``default_logfile``, is added. How they are used is described in
651:ref:`the_category_phrase`.
652
653::
654
655   channel default_syslog {
656       // send to syslog's daemon facility
657       syslog daemon;
658       // only send priority info and higher
659       severity info;
660   };
661
662   channel default_debug {
663       // write to named.run in the working directory
664       // Note: stderr is used instead of "named.run" if
665       // the server is started with the '-g' option.
666       file "named.run";
667       // log at the server's current debug level
668       severity dynamic;
669   };
670
671   channel default_stderr {
672       // writes to stderr
673       stderr;
674       // only send priority info and higher
675       severity info;
676   };
677
678   channel null {
679      // toss anything sent to this channel
680      null;
681   };
682
683   channel default_logfile {
684       // this channel is only present if named is
685       // started with the -L option, whose argument
686       // provides the file name
687       file "...";
688       // log at the server's current debug level
689       severity dynamic;
690   };
691
692The ``default_debug`` channel has the special property that it only
693produces output when the server's debug level is non-zero. It normally
694writes to a file called ``named.run`` in the server's working directory.
695
696For security reasons, when the ``-u`` command-line option is used, the
697``named.run`` file is created only after ``named`` has changed to the
698new UID, and any debug output generated while ``named`` is starting -
699and still running as root - is discarded. To capture this
700output, run the server with the ``-L`` option to specify a
701default logfile, or the ``-g`` option to log to standard error which can
702be redirected to a file.
703
704Once a channel is defined, it cannot be redefined. The
705built-in channels cannot be altered directly, but the default logging
706can be modified by pointing categories at defined channels.
707
708.. _the_category_phrase:
709
710The ``category`` Phrase
711^^^^^^^^^^^^^^^^^^^^^^^
712
713There are many categories, so desired logs can be sent anywhere
714while unwanted logs are ignored. If
715a list of channels is not specified for a category, log messages in that
716category are sent to the ``default`` category instead. If no
717default category is specified, the following "default default" is used:
718
719::
720
721   category default { default_syslog; default_debug; };
722
723If ``named`` is started with the ``-L`` option, the default category
724is:
725
726::
727
728   category default { default_logfile; default_debug; };
729
730As an example, let's say a user wants to log security events to a file, but
731also wants to keep the default logging behavior. They would specify the
732following:
733
734::
735
736   channel my_security_channel {
737       file "my_security_file";
738       severity info;
739   };
740   category security {
741       my_security_channel;
742       default_syslog;
743       default_debug;
744   };
745
746To discard all messages in a category, specify the ``null`` channel:
747
748::
749
750   category xfer-out { null; };
751   category notify { null; };
752
753The following are the available categories and brief descriptions of the
754types of log information they contain. More categories may be added in
755future BIND releases.
756
757.. include:: logging-categories.rst
758
759.. _query_errors:
760
761The ``query-errors`` Category
762^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
763
764The ``query-errors`` category is used to indicate why and how specific queries
765resulted in responses which indicate an error.  Normally, these messages are
766logged at ``debug`` logging levels; note, however, that if query logging is
767active, some are logged at ``info``. The logging levels are described below:
768
769At ``debug`` level 1 or higher - or at ``info`` when query logging is
770active - each response with the rcode of SERVFAIL is logged as follows:
771
772``client 127.0.0.1#61502: query failed (SERVFAIL) for www.example.com/IN/AAAA at query.c:3880``
773
774This means an error resulting in SERVFAIL was detected at line 3880 of source
775file ``query.c``.  Log messages of this level are particularly helpful in identifying
776the cause of SERVFAIL for an authoritative server.
777
778At ``debug`` level 2 or higher, detailed context information about recursive
779resolutions that resulted in SERVFAIL is logged.  The log message looks
780like this:
781
782::
783
784   fetch completed at resolver.c:2970 for www.example.com/A
785   in 10.000183: timed out/success [domain:example.com,
786   referral:2,restart:7,qrysent:8,timeout:5,lame:0,quota:0,neterr:0,
787   badresp:1,adberr:0,findfail:0,valfail:0]
788
789The first part before the colon shows that a recursive resolution for
790AAAA records of www.example.com completed in 10.000183 seconds, and the
791final result that led to the SERVFAIL was determined at line 2970 of
792source file ``resolver.c``.
793
794The next part shows the detected final result and the latest result of
795DNSSEC validation.  The latter is always "success" when no validation attempt
796was made.  In this example, this query probably resulted in SERVFAIL because all
797name servers are down or unreachable, leading to a timeout in 10 seconds.
798DNSSEC validation was probably not attempted.
799
800The last part, enclosed in square brackets, shows statistics collected for this
801particular resolution attempt.  The ``domain`` field shows the deepest zone that
802the resolver reached; it is the zone where the error was finally detected.  The
803meaning of the other fields is summarized in the following list.
804
805``referral``
806    The number of referrals the resolver received throughout the resolution process. In the above ``example.com`` there are two.
807
808``restart``
809    The number of cycles that the resolver tried remote servers at the ``domain`` zone. In each cycle, the resolver sends one query (possibly resending it, depending on the response) to each known name server of the ``domain`` zone.
810
811``qrysent``
812    The number of queries the resolver sent at the ``domain`` zone.
813
814``timeout``
815    The number of timeouts the resolver received since the last response.
816
817``lame``
818    The number of lame servers the resolver detected at the ``domain`` zone. A server is detected to be lame either by an invalid response or as a result of lookup in BIND 9's address database (ADB), where lame servers are cached.
819
820``quota``
821    The number of times the resolver was unable to send a query because it had exceeded the permissible fetch quota for a server.
822
823``neterr``
824    The number of erroneous results that the resolver encountered in sending queries at the ``domain`` zone. One common case is when the remote server is unreachable and the resolver receives an "ICMP unreachable" error message.
825
826``badresp``
827    The number of unexpected responses (other than ``lame``) to queries sent by the resolver at the ``domain`` zone.
828
829``adberr``
830    Failures in finding remote server addresses of the``domain`` zone in the ADB. One common case of this is that the remote server's name does not have any address records.
831
832``findfail``
833    Failures to resolve remote server addresses. This is a total number of failures throughout the resolution process.
834
835``valfail``
836    Failures of DNSSEC validation. Validation failures are counted throughout the resolution process (not limited to the ``domain`` zone), but should only happen in ``domain``.
837
838At ``debug`` level 3 or higher, the same messages as those at
839``debug`` level 1 are logged for errors other than
840SERVFAIL. Note that negative responses such as NXDOMAIN are not errors, and are
841not logged at this debug level.
842
843At ``debug`` level 4 or higher, the detailed context information logged at
844``debug`` level 2 is logged for errors other than SERVFAIL and for negative
845responses such as NXDOMAIN.
846
847.. _parental_agents_grammar:
848
849``parental-agents`` Statement Grammar
850~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
851
852.. include:: ../misc/parental-agents.grammar.rst
853
854.. _parental_agents_statement:
855
856``parental-agents`` Statement Definition and Usage
857~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
858
859``parental-agents`` lists allow for a common set of parental agents to be easily
860used by multiple primary and secondary zones in their ``parental-agents`` lists.
861A parental agent is the entity that the zone has a relationship with to
862change its delegation information (defined in :rfc:`7344`).
863
864.. _primaries_grammar:
865
866``primaries`` Statement Grammar
867~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
868
869.. include:: ../misc/primaries.grammar.rst
870
871.. _primaries_statement:
872
873``primaries`` Statement Definition and Usage
874~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
875
876``primaries`` lists allow for a common set of primary servers  to be easily
877used by multiple stub and secondary zones in their ``primaries`` or
878``also-notify`` lists. (Note: ``primaries`` is a synonym for the original
879keyword ``masters``, which can still be used, but is no longer the
880preferred terminology.)
881
882.. _options_grammar:
883
884``options`` Statement Grammar
885~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
886
887This is the grammar of the ``options`` statement in the ``named.conf``
888file:
889
890.. include:: ../misc/options.grammar.rst
891
892.. _options:
893
894``options`` Statement Definition and Usage
895~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
896
897The ``options`` statement sets up global options to be used by BIND.
898This statement may appear only once in a configuration file. If there is
899no ``options`` statement, an options block with each option set to its
900default is used.
901
902.. _attach-cache:
903
904``attach-cache``
905   This option allows multiple views to share a single cache database. Each view has
906   its own cache database by default, but if multiple views have the
907   same operational policy for name resolution and caching, those views
908   can share a single cache to save memory, and possibly improve
909   resolution efficiency, by using this option.
910
911   The ``attach-cache`` option may also be specified in ``view``
912   statements, in which case it overrides the global ``attach-cache``
913   option.
914
915   The ``cache_name`` specifies the cache to be shared. When the ``named``
916   server configures views which are supposed to share a cache, it
917   creates a cache with the specified name for the first view of these
918   sharing views. The rest of the views simply refer to the
919   already-created cache.
920
921   One common configuration to share a cache is to allow all views
922   to share a single cache. This can be done by specifying
923   ``attach-cache`` as a global option with an arbitrary name.
924
925   Another possible operation is to allow a subset of all views to share
926   a cache while the others retain their own caches. For example, if
927   there are three views A, B, and C, and only A and B should share a
928   cache, specify the ``attach-cache`` option as a view of A (or B)'s
929   option, referring to the other view name:
930
931   ::
932
933        view "A" {
934          // this view has its own cache
935          ...
936        };
937        view "B" {
938          // this view refers to A's cache
939          attach-cache "A";
940        };
941        view "C" {
942          // this view has its own cache
943          ...
944        };
945
946   Views that share a cache must have the same policy on configurable
947   parameters that may affect caching. The current implementation
948   requires the following configurable options be consistent among these
949   views: ``check-names``, ``dnssec-accept-expired``,
950   ``dnssec-validation``, ``max-cache-ttl``, ``max-ncache-ttl``,
951   ``max-stale-ttl``, ``max-cache-size``, ``min-cache-ttl``,
952   ``min-ncache-ttl``, and ``zero-no-soa-ttl``.
953
954   Note that there may be other parameters that may cause confusion if
955   they are inconsistent for different views that share a single cache.
956   For example, if these views define different sets of forwarders that
957   can return different answers for the same question, sharing the
958   answer does not make sense or could even be harmful. It is the
959   administrator's responsibility to ensure that configuration differences in
960   different views do not cause disruption with a shared cache.
961
962``directory``
963   This sets the working directory of the server. Any non-absolute pathnames in
964   the configuration file are taken as relative to this directory.
965   The default location for most server output files (e.g.,
966   ``named.run``) is this directory. If a directory is not specified,
967   the working directory defaults to ``"."``, the directory from
968   which the server was started. The directory specified should be an
969   absolute path, and *must* be writable by the effective user ID of the
970   ``named`` process.
971
972   The option takes effect only at the time that the configuration
973   option is parsed; if other files are being included before or after specifying the
974   new ``directory``, the ``directory`` option must be listed
975   before any other directive (like ``include``) that can work with relative
976   files. The safest way to include files is to use absolute file names.
977
978``dnstap``
979   ``dnstap`` is a fast, flexible method for capturing and logging DNS
980   traffic. Developed by Robert Edmonds at Farsight Security, Inc., and
981   supported by multiple DNS implementations, ``dnstap`` uses
982   ``libfstrm`` (a lightweight high-speed framing library; see
983   https://github.com/farsightsec/fstrm) to send event payloads which
984   are encoded using Protocol Buffers (``libprotobuf-c``, a mechanism
985   for serializing structured data developed by Google, Inc.; see
986   https://developers.google.com/protocol-buffers/).
987
988   To enable ``dnstap`` at compile time, the ``fstrm`` and
989   ``protobuf-c`` libraries must be available, and BIND must be
990   configured with ``--enable-dnstap``.
991
992   The ``dnstap`` option is a bracketed list of message types to be
993   logged. These may be set differently for each view. Supported types
994   are ``client``, ``auth``, ``resolver``, ``forwarder``, and
995   ``update``. Specifying type ``all`` causes all ``dnstap``
996   messages to be logged, regardless of type.
997
998   Each type may take an additional argument to indicate whether to log
999   ``query`` messages or ``response`` messages; if not specified, both
1000   queries and responses are logged.
1001
1002   Example: To log all authoritative queries and responses, recursive
1003   client responses, and upstream queries sent by the resolver, use:
1004
1005   ::
1006
1007      dnstap {
1008        auth;
1009        client response;
1010        resolver query;
1011      };
1012
1013   Logged ``dnstap`` messages can be parsed using the ``dnstap-read``
1014   utility (see :ref:`man_dnstap-read` for details).
1015
1016   For more information on ``dnstap``, see http://dnstap.info.
1017
1018   The fstrm library has a number of tunables that are exposed in
1019   ``named.conf``, and can be modified if necessary to improve
1020   performance or prevent loss of data. These are:
1021
1022   -  ``fstrm-set-buffer-hint``: The threshold number of bytes to
1023      accumulate in the output buffer before forcing a buffer flush. The
1024      minimum is 1024, the maximum is 65536, and the default is 8192.
1025
1026   -  ``fstrm-set-flush-timeout``: The number of seconds to allow
1027      unflushed data to remain in the output buffer. The minimum is 1
1028      second, the maximum is 600 seconds (10 minutes), and the default
1029      is 1 second.
1030
1031   -  ``fstrm-set-output-notify-threshold``: The number of outstanding
1032      queue entries to allow on an input queue before waking the I/O
1033      thread. The minimum is 1 and the default is 32.
1034
1035   -  ``fstrm-set-output-queue-model``: The queuing semantics
1036      to use for queue objects. The default is ``mpsc`` (multiple
1037      producer, single consumer); the other option is ``spsc`` (single
1038      producer, single consumer).
1039
1040   -  ``fstrm-set-input-queue-size``: The number of queue entries to
1041      allocate for each input queue. This value must be a power of 2.
1042      The minimum is 2, the maximum is 16384, and the default is 512.
1043
1044   -  ``fstrm-set-output-queue-size``: The number of queue entries to
1045      allocate for each output queue. The minimum is 2, the maximum is
1046      system-dependent and based on ``IOV_MAX``, and the default is 64.
1047
1048   -  ``fstrm-set-reopen-interval``: The number of seconds to wait
1049      between attempts to reopen a closed output stream. The minimum is
1050      1 second, the maximum is 600 seconds (10 minutes), and the default
1051      is 5 seconds. For convenience, TTL-style time-unit suffixes may be
1052      used to specify the value.
1053
1054   Note that all of the above minimum, maximum, and default values are
1055   set by the ``libfstrm`` library, and may be subject to change in
1056   future versions of the library. See the ``libfstrm`` documentation
1057   for more information.
1058
1059``dnstap-output``
1060   This configures the path to which the ``dnstap`` frame stream is sent
1061   if ``dnstap`` is enabled at compile time and active.
1062
1063   The first argument is either ``file`` or ``unix``, indicating whether
1064   the destination is a file or a Unix domain socket. The second
1065   argument is the path of the file or socket. (Note: when using a
1066   socket, ``dnstap`` messages are only sent if another process such
1067   as ``fstrm_capture`` (provided with ``libfstrm``) is listening on the
1068   socket.)
1069
1070   If the first argument is ``file``, then up to three additional
1071   options can be added: ``size`` indicates the size to which a
1072   ``dnstap`` log file can grow before being rolled to a new file;
1073   ``versions`` specifies the number of rolled log files to retain; and
1074   ``suffix`` indicates whether to retain rolled log files with an
1075   incrementing counter as the suffix (``increment``) or with the
1076   current timestamp (``timestamp``). These are similar to the ``size``,
1077   ``versions``, and ``suffix`` options in a ``logging`` channel. The
1078   default is to allow ``dnstap`` log files to grow to any size without
1079   rolling.
1080
1081   ``dnstap-output`` can only be set globally in ``options``. Currently,
1082   it can only be set once while ``named`` is running; once set, it
1083   cannot be changed by ``rndc reload`` or ``rndc reconfig``.
1084
1085``dnstap-identity``
1086   This specifies an ``identity`` string to send in ``dnstap`` messages. If
1087   set to ``hostname``, which is the default, the server's hostname
1088   is sent. If set to ``none``, no identity string is sent.
1089
1090``dnstap-version``
1091   This specifies a ``version`` string to send in ``dnstap`` messages. The
1092   default is the version number of the BIND release. If set to
1093   ``none``, no version string is sent.
1094
1095``geoip-directory``
1096   When ``named`` is compiled using the MaxMind GeoIP2 geolocation API, this
1097   specifies the directory containing GeoIP database files.  By default, the
1098   option is set based on the prefix used to build the ``libmaxminddb`` module;
1099   for example, if the library is installed in ``/usr/local/lib``, then the
1100   default ``geoip-directory`` is ``/usr/local/share/GeoIP``. On Windows,
1101   the default is the ``named`` working directory.  See :ref:`acl`
1102   for details about ``geoip`` ACLs.
1103
1104``key-directory``
1105   This is the directory where the public and private DNSSEC key files should be
1106   found when performing a dynamic update of secure zones, if different
1107   than the current working directory. (Note that this option has no
1108   effect on the paths for files containing non-DNSSEC keys such as
1109   ``bind.keys``, ``rndc.key``, or ``session.key``.)
1110
1111``lmdb-mapsize``
1112   When ``named`` is built with liblmdb, this option sets a maximum size
1113   for the memory map of the new-zone database (NZD) in LMDB database
1114   format. This database is used to store configuration information for
1115   zones added using ``rndc addzone``. Note that this is not the NZD
1116   database file size, but the largest size that the database may grow
1117   to.
1118
1119   Because the database file is memory-mapped, its size is limited by
1120   the address space of the ``named`` process. The default of 32 megabytes
1121   was chosen to be usable with 32-bit ``named`` builds. The largest
1122   permitted value is 1 terabyte. Given typical zone configurations
1123   without elaborate ACLs, a 32 MB NZD file ought to be able to hold
1124   configurations of about 100,000 zones.
1125
1126``managed-keys-directory``
1127   This specifies the directory in which to store the files that track managed DNSSEC
1128   keys (i.e., those configured using the ``initial-key`` or ``initial-ds``
1129   keywords in a ``trust-anchors`` statement). By default, this is the working
1130   directory. The directory *must* be writable by the effective user ID of the
1131   ``named`` process.
1132
1133   If ``named`` is not configured to use views, managed keys for
1134   the server are tracked in a single file called
1135   ``managed-keys.bind``. Otherwise, managed keys are tracked in
1136   separate files, one file per view; each file name is the view
1137   name (or, if it contains characters that are incompatible with use as
1138   a file name, the SHA256 hash of the view name), followed by the
1139   extension ``.mkeys``.
1140
1141   (Note: in earlier releases, file names for views always used the
1142   SHA256 hash of the view name. To ensure compatibility after upgrading,
1143   if a file using the old name format is found to exist, it is
1144   used instead of the new format.)
1145
1146``max-ixfr-ratio``
1147   This sets the size threshold (expressed as a percentage of the size
1148   of the full zone) beyond which ``named`` chooses to use an AXFR
1149   response rather than IXFR when answering zone transfer requests. See
1150   :ref:`incremental_zone_transfers`.
1151
1152   The minimum value is ``1%``. The keyword ``unlimited`` disables ratio
1153   checking and allows IXFRs of any size. The default is ``unlimited``.
1154
1155``new-zones-directory``
1156   This specifies the directory in which to store the configuration
1157   parameters for zones added via ``rndc addzone``. By default, this is
1158   the working directory. If set to a relative path, it is relative
1159   to the working directory. The directory *must* be writable by the
1160   effective user ID of the ``named`` process.
1161
1162``qname-minimization``
1163   This option controls QNAME minimization behavior in the BIND
1164   resolver. When set to ``strict``, BIND follows the QNAME
1165   minimization algorithm to the letter, as specified in :rfc:`7816`.
1166   Setting this option to ``relaxed`` causes BIND to fall back to
1167   normal (non-minimized) query mode when it receives either NXDOMAIN or
1168   other unexpected responses (e.g., SERVFAIL, improper zone cut,
1169   REFUSED) to a minimized query. ``disabled`` disables QNAME
1170   minimization completely. The current default is ``relaxed``, but it
1171   may be changed to ``strict`` in a future release.
1172
1173``tkey-gssapi-keytab``
1174   This is the KRB5 keytab file to use for GSS-TSIG updates. If this option is
1175   set and tkey-gssapi-credential is not set, updates are
1176   allowed with any key matching a principal in the specified keytab.
1177
1178``tkey-gssapi-credential``
1179   This is the security credential with which the server should authenticate
1180   keys requested by the GSS-TSIG protocol. Currently only Kerberos 5
1181   authentication is available; the credential is a Kerberos
1182   principal which the server can acquire through the default system key
1183   file, normally ``/etc/krb5.keytab``. The location of the keytab file can be
1184   overridden using the ``tkey-gssapi-keytab`` option. Normally this
1185   principal is of the form ``DNS/server.domain``. To use
1186   GSS-TSIG, ``tkey-domain`` must also be set if a specific keytab is
1187   not set with ``tkey-gssapi-keytab``.
1188
1189``tkey-domain``
1190   This domain is appended to the names of all shared keys generated with
1191   ``TKEY``. When a client requests a ``TKEY`` exchange, it may or may
1192   not specify the desired name for the key. If present, the name of the
1193   shared key is ``client-specified part`` + ``tkey-domain``.
1194   Otherwise, the name of the shared key is ``random hex digits``
1195   + ``tkey-domain``. In most cases, the ``domainname``
1196   should be the server's domain name, or an otherwise nonexistent
1197   subdomain like ``_tkey.domainname``. If using GSS-TSIG,
1198   this variable must be defined, unless a specific keytab
1199   is specified using ``tkey-gssapi-keytab``.
1200
1201``tkey-dhkey``
1202   This is the Diffie-Hellman key used by the server to generate shared keys
1203   with clients using the Diffie-Hellman mode of ``TKEY``. The server
1204   must be able to load the public and private keys from files in the
1205   working directory. In most cases, the ``key_name`` should be the
1206   server's host name.
1207
1208``cache-file``
1209   This is for testing only. Do not use.
1210
1211``dump-file``
1212   This is the pathname of the file the server dumps the database to, when
1213   instructed to do so with ``rndc dumpdb``. If not specified, the
1214   default is ``named_dump.db``.
1215
1216``memstatistics-file``
1217   This is the pathname of the file the server writes memory usage statistics to
1218   on exit. If not specified, the default is ``named.memstats``.
1219
1220``lock-file``
1221   This is the pathname of a file on which ``named`` attempts to acquire a
1222   file lock when starting for the first time; if unsuccessful, the
1223   server terminates, under the assumption that another server
1224   is already running. If not specified, the default is
1225   ``none``.
1226
1227   Specifying ``lock-file none`` disables the use of a lock file.
1228   ``lock-file`` is ignored if ``named`` was run using the ``-X``
1229   option, which overrides it. Changes to ``lock-file`` are ignored if
1230   ``named`` is being reloaded or reconfigured; it is only effective
1231   when the server is first started.
1232
1233``pid-file``
1234   This is the pathname of the file the server writes its process ID in. If not
1235   specified, the default is ``/var/run/named/named.pid``. The PID file
1236   is used by programs that send signals to the running name
1237   server. Specifying ``pid-file none`` disables the use of a PID file;
1238   no file is written and any existing one is removed. Note
1239   that ``none`` is a keyword, not a filename, and therefore is not
1240   enclosed in double quotes.
1241
1242``recursing-file``
1243   This is the pathname of the file where the server dumps the queries that are
1244   currently recursing, when instructed to do so with ``rndc recursing``.
1245   If not specified, the default is ``named.recursing``.
1246
1247``statistics-file``
1248   This is the pathname of the file the server appends statistics to, when
1249   instructed to do so using ``rndc stats``. If not specified, the
1250   default is ``named.stats`` in the server's current directory. The
1251   format of the file is described in :ref:`statsfile`.
1252
1253``bindkeys-file``
1254   This is the pathname of a file to override the built-in trusted keys provided
1255   by ``named``. See the discussion of ``dnssec-validation`` for
1256   details. If not specified, the default is ``/etc/bind.keys``.
1257
1258``secroots-file``
1259   This is the pathname of the file the server dumps security roots to, when
1260   instructed to do so with ``rndc secroots``. If not specified, the
1261   default is ``named.secroots``.
1262
1263``session-keyfile``
1264   This is the pathname of the file into which to write a TSIG session key
1265   generated by ``named`` for use by ``nsupdate -l``. If not specified,
1266   the default is ``/var/run/named/session.key``. (See :ref:`dynamic_update_policies`,
1267   and in particular the discussion of the ``update-policy`` statement's
1268   ``local`` option, for more information about this feature.)
1269
1270``session-keyname``
1271   This is the key name to use for the TSIG session key. If not specified, the
1272   default is ``local-ddns``.
1273
1274``session-keyalg``
1275   This is the algorithm to use for the TSIG session key. Valid values are
1276   hmac-sha1, hmac-sha224, hmac-sha256, hmac-sha384, hmac-sha512, and
1277   hmac-md5. If not specified, the default is hmac-sha256.
1278
1279``port``
1280   This is the UDP/TCP port number the server uses to receive and send DNS
1281   protocol traffic. The default is 53. This option is mainly intended
1282   for server testing; a server using a port other than 53 is not
1283   able to communicate with the global DNS.
1284
1285``dscp``
1286   This is the global Differentiated Services Code Point (DSCP) value to
1287   classify outgoing DNS traffic, on operating systems that support DSCP.
1288   Valid values are 0 through 63. It is not configured by default.
1289
1290``random-device``
1291   This specifies a source of entropy to be used by the server; it is a
1292   device or file from which to read entropy. If it is a file,
1293   operations requiring entropy will fail when the file has been
1294   exhausted.
1295
1296   Entropy is needed for cryptographic operations such as TKEY
1297   transactions, dynamic update of signed zones, and generation of TSIG
1298   session keys. It is also used for seeding and stirring the
1299   pseudo-random number generator which is used for less critical
1300   functions requiring randomness, such as generation of DNS message
1301   transaction IDs.
1302
1303   If ``random-device`` is not specified, or if it is set to ``none``,
1304   entropy is read from the random number generation function
1305   supplied by the cryptographic library with which BIND was linked
1306   (i.e. OpenSSL or a PKCS#11 provider).
1307
1308   The ``random-device`` option takes effect during the initial
1309   configuration load at server startup time and is ignored on
1310   subsequent reloads.
1311
1312``preferred-glue``
1313   If specified, the listed type (A or AAAA) is emitted before
1314   other glue in the additional section of a query response. The default
1315   is to prefer A records when responding to queries that arrived via
1316   IPv4 and AAAA when responding to queries that arrived via IPv6.
1317
1318.. _root-delegation-only:
1319
1320``root-delegation-only``
1321   This turns on enforcement of delegation-only in TLDs (top-level domains)
1322   and root zones with an optional exclude list.
1323
1324   DS queries are expected to be made to and be answered by delegation-only
1325   zones. Such queries and responses are treated as an exception to
1326   delegation-only processing and are not converted to NXDOMAIN
1327   responses, provided a CNAME is not discovered at the query name.
1328
1329   If a delegation-only zone server also serves a child zone, it is not
1330   always possible to determine whether an answer comes from the
1331   delegation-only zone or the child zone. SOA NS and DNSKEY records are
1332   apex-only records and a matching response that contains these records
1333   or DS is treated as coming from a child zone. RRSIG records are also
1334   examined to see whether they are signed by a child zone, and the
1335   authority section is examined to see if there is evidence that
1336   the answer is from the child zone. Answers that are determined to be
1337   from a child zone are not converted to NXDOMAIN responses. Despite
1338   all these checks, there is still a possibility of false negatives when
1339   a child zone is being served.
1340
1341   Similarly, false positives can arise from empty nodes (no records at
1342   the name) in the delegation-only zone when the query type is not ``ANY``.
1343
1344   Note that some TLDs are not delegation-only; e.g., "DE", "LV", "US", and
1345   "MUSEUM". This list is not exhaustive.
1346
1347   ::
1348
1349      options {
1350          root-delegation-only exclude { "de"; "lv"; "us"; "museum"; };
1351      };
1352
1353``disable-algorithms``
1354   This disables the specified DNSSEC algorithms at and below the specified
1355   name. Multiple ``disable-algorithms`` statements are allowed. Only
1356   the best-match ``disable-algorithms`` clause is used to
1357   determine the algorithms.
1358
1359   If all supported algorithms are disabled, the zones covered by the
1360   ``disable-algorithms`` setting are treated as insecure.
1361
1362   Configured trust anchors in ``trust-anchors`` (or ``managed-keys`` or
1363   ``trusted-keys``) that match a disabled algorithm are ignored and treated
1364   as if they were not configured.
1365
1366``disable-ds-digests``
1367   This disables the specified DS digest types at and below the specified
1368   name. Multiple ``disable-ds-digests`` statements are allowed. Only
1369   the best-match ``disable-ds-digests`` clause is used to
1370   determine the digest types.
1371
1372   If all supported digest types are disabled, the zones covered by
1373   ``disable-ds-digests`` are treated as insecure.
1374
1375``dnssec-must-be-secure``
1376   This specifies hierarchies which must be or may not be secure (signed and
1377   validated). If ``yes``, then ``named`` only accepts answers if
1378   they are secure. If ``no``, then normal DNSSEC validation applies,
1379   allowing insecure answers to be accepted. The specified domain
1380   must be defined as a trust anchor, for instance in a ``trust-anchors``
1381   statement, or ``dnssec-validation auto`` must be active.
1382
1383``dns64``
1384   This directive instructs ``named`` to return mapped IPv4 addresses to
1385   AAAA queries when there are no AAAA records. It is intended to be
1386   used in conjunction with a NAT64. Each ``dns64`` defines one DNS64
1387   prefix. Multiple DNS64 prefixes can be defined.
1388
1389   Compatible IPv6 prefixes have lengths of 32, 40, 48, 56, 64, and 96, per
1390   :rfc:`6052`. Bits 64..71 inclusive must be zero, with the most significant bit
1391   of the prefix in position 0.
1392
1393   In addition, a reverse IP6.ARPA zone is created for the prefix
1394   to provide a mapping from the IP6.ARPA names to the corresponding
1395   IN-ADDR.ARPA names using synthesized CNAMEs. ``dns64-server`` and
1396   ``dns64-contact`` can be used to specify the name of the server and
1397   contact for the zones. These can be set at the view/options
1398   level but not on a per-prefix basis.
1399
1400   Each ``dns64`` supports an optional ``clients`` ACL that determines
1401   which clients are affected by this directive. If not defined, it
1402   defaults to ``any;``.
1403
1404   Each ``dns64`` supports an optional ``mapped`` ACL that selects which
1405   IPv4 addresses are to be mapped in the corresponding A RRset. If not
1406   defined, it defaults to ``any;``.
1407
1408   Normally, DNS64 does not apply to a domain name that owns one or more
1409   AAAA records; these records are simply returned. The optional
1410   ``exclude`` ACL allows specification of a list of IPv6 addresses that
1411   are ignored if they appear in a domain name's AAAA records;
1412   DNS64 is applied to any A records the domain name owns. If not
1413   defined, ``exclude`` defaults to ::ffff:0.0.0.0/96.
1414
1415   An optional ``suffix`` can also be defined to set the bits trailing
1416   the mapped IPv4 address bits. By default these bits are set to
1417   ``::``. The bits matching the prefix and mapped IPv4 address must be
1418   zero.
1419
1420   If ``recursive-only`` is set to ``yes``, the DNS64 synthesis only
1421   happens for recursive queries. The default is ``no``.
1422
1423   If ``break-dnssec`` is set to ``yes``, the DNS64 synthesis happens
1424   even if the result, if validated, would cause a DNSSEC validation
1425   failure. If this option is set to ``no`` (the default), the DO is set
1426   on the incoming query, and there are RRSIGs on the applicable
1427   records, then synthesis does not happen.
1428
1429   ::
1430
1431          acl rfc1918 { 10/8; 192.168/16; 172.16/12; };
1432
1433          dns64 64:FF9B::/96 {
1434              clients { any; };
1435              mapped { !rfc1918; any; };
1436              exclude { 64:FF9B::/96; ::ffff:0000:0000/96; };
1437              suffix ::;
1438          };
1439
1440``dnssec-loadkeys-interval``
1441   When a zone is configured with ``auto-dnssec maintain;``, its key
1442   repository must be checked periodically to see if any new keys have
1443   been added or any existing keys' timing metadata has been updated
1444   (see :ref:`man_dnssec-keygen` and :ref:`man_dnssec-settime`).
1445   The ``dnssec-loadkeys-interval`` option
1446   sets the frequency of automatic repository checks, in minutes.  The
1447   default is ``60`` (1 hour), the minimum is ``1`` (1 minute), and
1448   the maximum is ``1440`` (24 hours); any higher value is silently
1449   reduced.
1450
1451``dnssec-policy``
1452   This specifies which key and signing policy (KASP) should be used for this
1453   zone. This is a string referring to a ``dnssec-policy`` statement.  There
1454   are three built-in policies: ``default``, which uses the default policy,
1455   ``insecure``, to be used when you want to gracefully unsign your zone, and
1456   ``none``, which means no DNSSEC policy.  The default is ``none``.
1457   See :ref:`dnssec-policy Grammar <dnssec_policy_grammar>` for more details.
1458
1459``dnssec-update-mode``
1460   If this option is set to its default value of ``maintain`` in a zone
1461   of type ``primary`` which is DNSSEC-signed and configured to allow
1462   dynamic updates (see :ref:`dynamic_update_policies`), and if ``named`` has access
1463   to the private signing key(s) for the zone, then ``named``
1464   automatically signs all new or changed records and maintains signatures
1465   for the zone by regenerating RRSIG records whenever they approach
1466   their expiration date.
1467
1468   If the option is changed to ``no-resign``, then ``named`` signs
1469   all new or changed records, but scheduled maintenance of signatures
1470   is disabled.
1471
1472   With either of these settings, ``named`` rejects updates to a
1473   DNSSEC-signed zone when the signing keys are inactive or unavailable
1474   to ``named``. (A planned third option, ``external``, will disable all
1475   automatic signing and allow DNSSEC data to be submitted into a zone
1476   via dynamic update; this is not yet implemented.)
1477
1478``nta-lifetime``
1479   This specifies the default lifetime, in seconds, for
1480   negative trust anchors added via ``rndc nta``.
1481
1482   A negative trust anchor selectively disables DNSSEC validation for
1483   zones that are known to be failing because of misconfiguration, rather
1484   than an attack. When data to be validated is at or below an active
1485   NTA (and above any other configured trust anchors), ``named``
1486   aborts the DNSSEC validation process and treats the data as insecure
1487   rather than bogus. This continues until the NTA's lifetime has
1488   elapsed. NTAs persist across ``named`` restarts.
1489
1490   For convenience, TTL-style time-unit suffixes can be used to specify the NTA
1491   lifetime in seconds, minutes, or hours. It also accepts ISO 8601 duration
1492   formats.
1493
1494   ``nta-lifetime`` defaults to one hour; it cannot exceed one week.
1495
1496``nta-recheck``
1497   This specifies how often to check whether negative trust anchors added via
1498   ``rndc nta`` are still necessary.
1499
1500   A negative trust anchor is normally used when a domain has stopped
1501   validating due to operator error; it temporarily disables DNSSEC
1502   validation for that domain. In the interest of ensuring that DNSSEC
1503   validation is turned back on as soon as possible, ``named``
1504   periodically sends a query to the domain, ignoring negative trust
1505   anchors, to find out whether it can now be validated. If so, the
1506   negative trust anchor is allowed to expire early.
1507
1508   Validity checks can be disabled for an individual NTA by using
1509   ``rndc nta -f``, or for all NTAs by setting ``nta-recheck`` to zero.
1510
1511   For convenience, TTL-style time-unit suffixes can be used to specify the NTA
1512   recheck interval in seconds, minutes, or hours. It also accepts ISO 8601
1513   duration formats.
1514
1515   The default is five minutes. It cannot be longer than ``nta-lifetime``, which
1516   cannot be longer than a week.
1517
1518``max-zone-ttl``
1519
1520   This should now be configured as part of ``dnssec-policy``.
1521   Use of this option in ``options``, ``view`` and ``zone`` blocks has no
1522   effect on any zone for which a ``dnssec-policy`` has also been configured.
1523
1524   ``max-zone-ttl`` specifies a maximum permissible TTL value in seconds.
1525   For convenience, TTL-style time-unit suffixes may be used to specify the
1526   maximum value. When a zone file is loaded, any record encountered with a
1527   TTL higher than ``max-zone-ttl`` causes the zone to be rejected.
1528
1529   This is useful in DNSSEC-signed zones because when rolling to a new
1530   DNSKEY, the old key needs to remain available until RRSIG records
1531   have expired from caches. The ``max-zone-ttl`` option guarantees that
1532   the largest TTL in the zone is no higher than the set value.
1533
1534   (Note: because ``map``-format files load directly into memory, this
1535   option cannot be used with them.)
1536
1537   The default value is ``unlimited``. Setting ``max-zone-ttl`` to zero is
1538   equivalent to ``unlimited``.
1539
1540``stale-answer-ttl``
1541   This specifies the TTL to be returned on stale answers. The default is 30
1542   seconds. The minimum allowed is 1 second; a value of 0 is updated silently
1543   to 1 second.
1544
1545   For stale answers to be returned, they must be enabled, either in the
1546   configuration file using ``stale-answer-enable`` or via
1547   ``rndc serve-stale on``.
1548
1549``serial-update-method``
1550   Zones configured for dynamic DNS may use this option to set the
1551   update method to be used for the zone serial number in the SOA
1552   record.
1553
1554   With the default setting of ``serial-update-method increment;``, the
1555   SOA serial number is incremented by one each time the zone is
1556   updated.
1557
1558   When set to ``serial-update-method unixtime;``, the SOA serial number
1559   is set to the number of seconds since the Unix epoch, unless the
1560   serial number is already greater than or equal to that value, in
1561   which case it is simply incremented by one.
1562
1563   When set to ``serial-update-method date;``, the new SOA serial number
1564   is the current date in the form "YYYYMMDD", followed by two
1565   zeroes, unless the existing serial number is already greater than or
1566   equal to that value, in which case it is incremented by one.
1567
1568``zone-statistics``
1569   If ``full``, the server collects statistical data on all zones,
1570   unless specifically turned off on a per-zone basis by specifying
1571   ``zone-statistics terse`` or ``zone-statistics none`` in the ``zone``
1572   statement. The statistical data includes, for example, DNSSEC signing
1573   operations and the number of authoritative answers per query type. The
1574   default is ``terse``, providing minimal statistics on zones
1575   (including name and current serial number, but not query type
1576   counters).
1577
1578   These statistics may be accessed via the ``statistics-channel`` or
1579   using ``rndc stats``, which dumps them to the file listed in the
1580   ``statistics-file``. See also :ref:`statsfile`.
1581
1582   For backward compatibility with earlier versions of BIND 9, the
1583   ``zone-statistics`` option can also accept ``yes`` or ``no``; ``yes``
1584   has the same meaning as ``full``. As of BIND 9.10, ``no`` has the
1585   same meaning as ``none``; previously, it was the same as ``terse``.
1586
1587.. _boolean_options:
1588
1589Boolean Options
1590^^^^^^^^^^^^^^^
1591
1592``automatic-interface-scan``
1593   If ``yes`` and supported by the operating system, this automatically rescans
1594   network interfaces when the interface addresses are added or removed.  The
1595   default is ``yes``.  This configuration option does not affect the time-based
1596   ``interface-interval`` option; it is recommended to set the time-based
1597   ``interface-interval`` to 0 when the operator confirms that automatic
1598   interface scanning is supported by the operating system.
1599
1600   The ``automatic-interface-scan`` implementation uses routing sockets for the
1601   network interface discovery; therefore, the operating system must
1602   support the routing sockets for this feature to work.
1603
1604``allow-new-zones``
1605   If ``yes``, then zones can be added at runtime via ``rndc addzone``.
1606   The default is ``no``.
1607
1608   Newly added zones' configuration parameters are stored so that they
1609   can persist after the server is restarted. The configuration
1610   information is saved in a file called ``viewname.nzf`` (or, if
1611   ``named`` is compiled with liblmdb, in an LMDB database file called
1612   ``viewname.nzd``). "viewname" is the name of the view, unless the view
1613   name contains characters that are incompatible with use as a file
1614   name, in which case a cryptographic hash of the view name is used
1615   instead.
1616
1617   Configurations for zones added at runtime are stored either in
1618   a new-zone file (NZF) or a new-zone database (NZD), depending on
1619   whether ``named`` was linked with liblmdb at compile time. See
1620   :ref:`man_rndc` for further details about ``rndc addzone``.
1621
1622``auth-nxdomain``
1623   If ``yes``, then the ``AA`` bit is always set on NXDOMAIN responses,
1624   even if the server is not actually authoritative. The default is
1625   ``no``.
1626
1627``memstatistics``
1628   This writes memory statistics to the file specified by
1629   ``memstatistics-file`` at exit. The default is ``no`` unless ``-m
1630   record`` is specified on the command line, in which case it is ``yes``.
1631
1632``dialup``
1633   If ``yes``, then the server treats all zones as if they are doing
1634   zone transfers across a dial-on-demand dialup link, which can be
1635   brought up by traffic originating from this server. Although this setting has
1636   different effects according to zone type, it concentrates the zone
1637   maintenance so that everything happens quickly, once every
1638   ``heartbeat-interval``, ideally during a single call. It also
1639   suppresses some normal zone maintenance traffic. The default
1640   is ``no``.
1641
1642   If specified in the ``view`` and
1643   ``zone`` statements, the ``dialup`` option overrides the global ``dialup``
1644   option.
1645
1646   If the zone is a primary zone, the server sends out a NOTIFY
1647   request to all the secondaries (default). This should trigger the zone
1648   serial number check in the secondary (providing it supports NOTIFY),
1649   allowing the secondary to verify the zone while the connection is active.
1650   The set of servers to which NOTIFY is sent can be controlled by
1651   ``notify`` and ``also-notify``.
1652
1653   If the zone is a secondary or stub zone, the server suppresses
1654   the regular "zone up to date" (refresh) queries and only performs them
1655   when the ``heartbeat-interval`` expires, in addition to sending NOTIFY
1656   requests.
1657
1658   Finer control can be achieved by using ``notify``, which only sends
1659   NOTIFY messages; ``notify-passive``, which sends NOTIFY messages and
1660   suppresses the normal refresh queries; ``refresh``, which suppresses
1661   normal refresh processing and sends refresh queries when the
1662   ``heartbeat-interval`` expires; and ``passive``, which disables
1663   normal refresh processing.
1664
1665   +--------------------+-----------------+-----------------+-----------------+
1666   | dialup mode        | normal refresh  | heart-beat      | heart-beat      |
1667   |                    |                 | refresh         | notify          |
1668   +--------------------+-----------------+-----------------+-----------------+
1669   | ``no``             | yes             | no              | no              |
1670   | (default)          |                 |                 |                 |
1671   +--------------------+-----------------+-----------------+-----------------+
1672   | ``yes``            | no              | yes             | yes             |
1673   +--------------------+-----------------+-----------------+-----------------+
1674   | ``notify``         | yes             | no              | yes             |
1675   +--------------------+-----------------+-----------------+-----------------+
1676   | ``refresh``        | no              | yes             | no              |
1677   +--------------------+-----------------+-----------------+-----------------+
1678   | ``passive``        | no              | no              | no              |
1679   +--------------------+-----------------+-----------------+-----------------+
1680   | ``notify-passive`` | no              | no              | yes             |
1681   +--------------------+-----------------+-----------------+-----------------+
1682
1683   Note that normal NOTIFY processing is not affected by ``dialup``.
1684
1685``flush-zones-on-shutdown``
1686   When the name server exits upon receiving SIGTERM, flush or do not
1687   flush any pending zone writes. The default is
1688   ``flush-zones-on-shutdown no``.
1689
1690``geoip-use-ecs``
1691   This option was part of an experimental implementation of the EDNS
1692   CLIENT-SUBNET for authoritative servers, but is now obsolete.
1693
1694``root-key-sentinel``
1695   If ``yes``, respond to root key sentinel probes as described in
1696   draft-ietf-dnsop-kskroll-sentinel-08. The default is ``yes``.
1697
1698``reuseport``
1699   This option enables kernel load-balancing of sockets on systems which support
1700   it, including Linux (SO_REUSEPORT) and FreeBSD (SO_REUSEPORT_LB). This
1701   instructs the kernel to distribute incoming socket connections among the
1702   networking threads based on a hashing scheme. For more information, see the
1703   receive network flow classification options (``rx-flow-hash``) section in the
1704   ``ethtool`` manual page. The default is ``yes``.
1705
1706   Enabling ``reuseport`` significantly increases general throughput when
1707   incoming traffic is distributed uniformly onto the threads by the
1708   operating system. However, in cases where a worker thread is busy with a
1709   long-lasting operation, such as processing a Response Policy Zone (RPZ) or
1710   Catalog Zone update or an unusually large zone transfer, incoming traffic
1711   that hashes onto that thread may be delayed. On servers where these events
1712   occur frequently, it may be preferable to disable socket load-balancing so
1713   that other threads can pick up the traffic that would have been sent to the
1714   busy thread.
1715
1716   Note: this option can only be set when ``named`` first starts.
1717   Changes will not take effect during reconfiguration; the server
1718   must be restarted.
1719
1720``message-compression``
1721   If ``yes``, DNS name compression is used in responses to regular
1722   queries (not including AXFR or IXFR, which always use compression).
1723   Setting this option to ``no`` reduces CPU usage on servers and may
1724   improve throughput. However, it increases response size, which may
1725   cause more queries to be processed using TCP; a server with
1726   compression disabled is out of compliance with :rfc:`1123` Section
1727   6.1.3.2. The default is ``yes``.
1728
1729``minimal-responses``
1730   This option controls the addition of records to the authority and
1731   additional sections of responses. Such records may be included in
1732   responses to be helpful to clients; for example, MX records may
1733   have associated address records included in the additional section,
1734   obviating the need for a separate address lookup. However, adding
1735   these records to responses is not mandatory and requires additional
1736   database lookups, causing extra latency when marshalling responses.
1737
1738   Responses to DNSKEY, DS, CDNSKEY, and CDS requests will never have
1739   optional additional records added. Responses to NS requests will
1740   always have additional section processing.
1741
1742   ``minimal-responses`` takes one of four values:
1743
1744   -  ``no``: the server is as complete as possible when generating
1745      responses.
1746   -  ``yes``: the server only adds records to the authority and additional
1747      sections when such records are required by the DNS protocol (for
1748      example, when returning delegations or negative responses). This
1749      provides the best server performance but may result in more client
1750      queries.
1751   -  ``no-auth``: the server omits records from the authority section except
1752      when they are required, but it may still add records to the
1753      additional section.
1754   -  ``no-auth-recursive``: the same as ``no-auth`` when recursion is requested
1755      in the query (RD=1), or the same as ``no`` if recursion is not requested.
1756
1757   ``no-auth`` and ``no-auth-recursive`` are useful when answering stub
1758   clients, which usually ignore the authority section.
1759   ``no-auth-recursive`` is meant for use in mixed-mode servers that
1760   handle both authoritative and recursive queries.
1761
1762   The default is ``no-auth-recursive``.
1763
1764``glue-cache``
1765   When set to ``yes``, a cache is used to improve query performance
1766   when adding address-type (A and AAAA) glue records to the additional
1767   section of DNS response messages that delegate to a child zone.
1768
1769   The glue cache uses memory proportional to the number of delegations
1770   in the zone. The default setting is ``yes``, which improves
1771   performance at the cost of increased memory usage for the zone. To avoid
1772   this, set it to ``no``.
1773
1774``minimal-any``
1775   If set to ``yes``, the server replies with only one of
1776   the RRsets for the query name, and its covering RRSIGs if any,
1777   when generating a positive response to a query of type ANY over UDP,
1778   instead of replying with all known RRsets for the name. Similarly, a
1779   query for type RRSIG is answered with the RRSIG records covering
1780   only one type. This can reduce the impact of some kinds of attack
1781   traffic, without harming legitimate clients. (Note, however, that the
1782   RRset returned is the first one found in the database; it is not
1783   necessarily the smallest available RRset.) Additionally,
1784   ``minimal-responses`` is turned on for these queries, so no
1785   unnecessary records are added to the authority or additional
1786   sections. The default is ``no``.
1787
1788``notify``
1789   If set to ``yes`` (the default), DNS NOTIFY messages are sent when a
1790   zone the server is authoritative for changes; see :ref:`notify`.
1791   The messages are sent to the servers listed in the zone's NS records
1792   (except the primary server identified in the SOA MNAME field), and to
1793   any servers listed in the ``also-notify`` option.
1794
1795   If set to ``primary-only`` (or the older keyword ``master-only``),
1796   notifies are only sent for primary zones. If set to ``explicit``,
1797   notifies are sent only to servers explicitly listed using
1798   ``also-notify``. If set to ``no``, no notifies are sent.
1799
1800   The ``notify`` option may also be specified in the ``zone``
1801   statement, in which case it overrides the ``options notify``
1802   statement. It would only be necessary to turn off this option if it
1803   caused secondary zones to crash.
1804
1805``notify-to-soa``
1806   If ``yes``, do not check the name servers in the NS RRset against the
1807   SOA MNAME. Normally a NOTIFY message is not sent to the SOA MNAME
1808   (SOA ORIGIN), as it is supposed to contain the name of the ultimate
1809   primary server. Sometimes, however, a secondary server is listed as the SOA MNAME in
1810   hidden primary configurations; in that case, the
1811   ultimate primary should be set to still send NOTIFY messages to all the name servers
1812   listed in the NS RRset.
1813
1814``recursion``
1815   If ``yes``, and a DNS query requests recursion, then the server
1816   attempts to do all the work required to answer the query. If recursion
1817   is off and the server does not already know the answer, it
1818   returns a referral response. The default is ``yes``. Note that setting
1819   ``recursion no`` does not prevent clients from getting data from the
1820   server's cache; it only prevents new data from being cached as an
1821   effect of client queries. Caching may still occur as an effect of the
1822   server's internal operation, such as NOTIFY address lookups.
1823
1824``request-nsid``
1825   If ``yes``, then an empty EDNS(0) NSID (Name Server Identifier)
1826   option is sent with all queries to authoritative name servers during
1827   iterative resolution. If the authoritative server returns an NSID
1828   option in its response, then its contents are logged in the ``nsid``
1829   category at level ``info``. The default is ``no``.
1830
1831``request-sit``
1832   This experimental option is obsolete.
1833
1834``require-server-cookie``
1835   If ``yes``, require a valid server cookie before sending a full response to a UDP
1836   request from a cookie-aware client. BADCOOKIE is sent if there is a
1837   bad or nonexistent server cookie.
1838
1839   The default is ``no``.
1840
1841   Users wishing to test that DNS COOKIE clients correctly handle
1842   BADCOOKIE, or who are getting a lot of forged DNS requests with DNS COOKIES
1843   present, should set this to ``yes``. Setting this to ``yes`` results in a reduced amplification effect
1844   in a reflection attack, as the BADCOOKIE response is smaller than a full
1845   response, while also requiring a legitimate client to follow up with a second
1846   query with the new, valid, cookie.
1847
1848``answer-cookie``
1849   When set to the default value of ``yes``, COOKIE EDNS options are
1850   sent when applicable in replies to client queries. If set to ``no``,
1851   COOKIE EDNS options are not sent in replies. This can only be set
1852   at the global options level, not per-view.
1853
1854   ``answer-cookie no`` is intended as a temporary measure, for use when
1855   ``named`` shares an IP address with other servers that do not yet
1856   support DNS COOKIE. A mismatch between servers on the same address is
1857   not expected to cause operational problems, but the option to disable
1858   COOKIE responses so that all servers have the same behavior is
1859   provided out of an abundance of caution. DNS COOKIE is an important
1860   security mechanism, and should not be disabled unless absolutely
1861   necessary.
1862
1863``send-cookie``
1864   If ``yes``, then a COOKIE EDNS option is sent along with the query.
1865   If the resolver has previously communicated with the server, the COOKIE
1866   returned in the previous transaction is sent. This is used by the
1867   server to determine whether the resolver has talked to it before. A
1868   resolver sending the correct COOKIE is assumed not to be an off-path
1869   attacker sending a spoofed-source query; the query is therefore
1870   unlikely to be part of a reflection/amplification attack, so
1871   resolvers sending a correct COOKIE option are not subject to response
1872   rate limiting (RRL). Resolvers which do not send a correct COOKIE
1873   option may be limited to receiving smaller responses via the
1874   ``nocookie-udp-size`` option.
1875
1876   The default is ``yes``.
1877
1878``stale-answer-enable``
1879   If ``yes``, enable the returning of "stale" cached answers when the name
1880   servers for a zone are not answering and the ``stale-cache-enable`` option is
1881   also enabled. The default is not to return stale answers.
1882
1883   Stale answers can also be enabled or disabled at runtime via
1884   ``rndc serve-stale on`` or ``rndc serve-stale off``; these override
1885   the configured setting. ``rndc serve-stale reset`` restores the
1886   setting to the one specified in ``named.conf``. Note that if stale
1887   answers have been disabled by ``rndc``, they cannot be
1888   re-enabled by reloading or reconfiguring ``named``; they must be
1889   re-enabled with ``rndc serve-stale on``, or the server must be
1890   restarted.
1891
1892   Information about stale answers is logged under the ``serve-stale``
1893   log category.
1894
1895``stale-answer-client-timeout``
1896   This option defines the amount of time (in milliseconds) that ``named``
1897   waits before attempting to answer the query with a stale RRset from cache.
1898   If a stale answer is found, ``named`` continues the ongoing fetches,
1899   attempting to refresh the RRset in cache until the
1900   ``resolver-query-timeout`` interval is reached.
1901
1902   This option is off by default, which is equivalent to setting it to
1903   ``off`` or ``disabled``. It also has no effect if ``stale-answer-enable``
1904   is disabled.
1905
1906   The maximum value for this option is ``resolver-query-timeout`` minus
1907   one second. The minimum value, ``0``, causes a cached (stale) RRset to be
1908   immediately returned if it is available while still attempting to
1909   refresh the data in cache. :rfc:`8767` recommends a value of ``1800``
1910   (milliseconds).
1911
1912``stale-cache-enable``
1913   If ``yes``, enable the retaining of "stale" cached answers.  Default ``yes``.
1914
1915``stale-refresh-time``
1916   If the name servers for a given zone are not answering, this sets the time
1917   window for which ``named`` will promptly return "stale" cached answers for
1918   that RRSet being requested before a new attempt in contacting the servers
1919   is made. For convenience, TTL-style time-unit suffixes may be used to
1920   specify the value. It also accepts ISO 8601 duration formats.
1921
1922   The default ``stale-refresh-time`` is 30 seconds, as :rfc:`8767` recommends
1923   that attempts to refresh to be done no more frequently than every 30
1924   seconds. A value of zero disables the feature, meaning that normal
1925   resolution will take place first, if that fails only then ``named`` will
1926   return "stale" cached answers.
1927
1928``nocookie-udp-size``
1929   This sets the maximum size of UDP responses that are sent to queries
1930   without a valid server COOKIE. A value below 128 is silently
1931   raised to 128. The default value is 4096, but the ``max-udp-size``
1932   option may further limit the response size as the default for
1933   ``max-udp-size`` is 4096.
1934
1935``sit-secret``
1936   This experimental option is obsolete.
1937
1938``cookie-algorithm``
1939   This sets the algorithm to be used when generating the server cookie; the options are
1940   "aes" or "siphash24". The default is "siphash24". The "aes" option remains for legacy
1941   purposes.
1942
1943``cookie-secret``
1944   If set, this is a shared secret used for generating and verifying
1945   EDNS COOKIE options within an anycast cluster. If not set, the system
1946   generates a random secret at startup. The shared secret is
1947   encoded as a hex string and needs to be 128 bits for either "siphash24"
1948   or "aes".
1949
1950   If there are multiple secrets specified, the first one listed in
1951   ``named.conf`` is used to generate new server cookies. The others
1952   are only used to verify returned cookies.
1953
1954``response-padding``
1955   The EDNS Padding option is intended to improve confidentiality when
1956   DNS queries are sent over an encrypted channel, by reducing the
1957   variability in packet sizes. If a query:
1958
1959   1. contains an EDNS Padding option,
1960   2. includes a valid server cookie or uses TCP,
1961   3. is not signed using TSIG or SIG(0), and
1962   4. is from a client whose address matches the specified ACL,
1963
1964   then the response is padded with an EDNS Padding option to a multiple
1965   of ``block-size`` bytes. If these conditions are not met, the
1966   response is not padded.
1967
1968   If ``block-size`` is 0 or the ACL is ``none;``, this feature is
1969   disabled and no padding occurs; this is the default. If
1970   ``block-size`` is greater than 512, a warning is logged and the value
1971   is truncated to 512. Block sizes are ordinarily expected to be powers
1972   of two (for instance, 128), but this is not mandatory.
1973
1974``trust-anchor-telemetry``
1975   This causes ``named`` to send specially formed queries once per day to
1976   domains for which trust anchors have been configured via, e.g.,
1977   ``trust-anchors`` or ``dnssec-validation auto``.
1978
1979   The query name used for these queries has the form
1980   ``_ta-xxxx(-xxxx)(...).<domain>``, where each "xxxx" is a group of four
1981   hexadecimal digits representing the key ID of a trusted DNSSEC key.
1982   The key IDs for each domain are sorted smallest to largest prior to
1983   encoding. The query type is NULL.
1984
1985   By monitoring these queries, zone operators are able to see which
1986   resolvers have been updated to trust a new key; this may help them
1987   decide when it is safe to remove an old one.
1988
1989   The default is ``yes``.
1990
1991``use-ixfr``
1992   *This option is obsolete*. To disable IXFR to a
1993   particular server or servers, see the information on the
1994   ``provide-ixfr`` option in :ref:`server_statement_definition_and_usage`.
1995   See also :ref:`incremental_zone_transfers`.
1996
1997``provide-ixfr``
1998   See the description of ``provide-ixfr`` in :ref:`server_statement_definition_and_usage`.
1999
2000``request-ixfr``
2001   See the description of ``request-ixfr`` in :ref:`server_statement_definition_and_usage`.
2002
2003``request-expire``
2004   See the description of ``request-expire`` in :ref:`server_statement_definition_and_usage`.
2005
2006``match-mapped-addresses``
2007   If ``yes``, then an IPv4-mapped IPv6 address matches any
2008   address-match list entries that match the corresponding IPv4 address.
2009
2010   This option was introduced to work around a kernel quirk in some
2011   operating systems that causes IPv4 TCP connections, such as zone
2012   transfers, to be accepted on an IPv6 socket using mapped addresses.
2013   This caused address-match lists designed for IPv4 to fail to match.
2014   However, ``named`` now solves this problem internally. The use of
2015   this option is discouraged.
2016
2017``ixfr-from-differences``
2018   When ``yes`` and the server loads a new version of a primary zone from
2019   its zone file or receives a new version of a secondary file via zone
2020   transfer, it compares the new version to the previous one and
2021   calculates a set of differences. The differences are then logged in
2022   the zone's journal file so that the changes can be transmitted to
2023   downstream secondaries as an incremental zone transfer.
2024
2025   By allowing incremental zone transfers to be used for non-dynamic
2026   zones, this option saves bandwidth at the expense of increased CPU
2027   and memory consumption at the primary server. In particular, if the new
2028   version of a zone is completely different from the previous one, the
2029   set of differences is of a size comparable to the combined size
2030   of the old and new zone versions, and the server needs to
2031   temporarily allocate memory to hold this complete difference set.
2032
2033   ``ixfr-from-differences`` also accepts ``primary``
2034   and ``secondary`` at the view and options levels,
2035   which causes ``ixfr-from-differences`` to be enabled for all primary
2036   or secondary zones, respectively. It is off for all zones by default.
2037
2038   Note: if inline signing is enabled for a zone, the user-provided
2039   ``ixfr-from-differences`` setting is ignored for that zone.
2040
2041``multi-master``
2042   This should be set when there are multiple primary servers for a zone and the
2043   addresses refer to different machines. If ``yes``, ``named`` does not
2044   log when the serial number on the primary is less than what ``named``
2045   currently has. The default is ``no``.
2046
2047``auto-dnssec``
2048   Zones configured for dynamic DNS may use this option to allow varying
2049   levels of automatic DNSSEC key management. There are three possible
2050   settings:
2051
2052   ``auto-dnssec allow;`` permits keys to be updated and the zone fully
2053   re-signed whenever the user issues the command ``rndc sign zonename``.
2054
2055   ``auto-dnssec maintain;`` includes the above, but also
2056   automatically adjusts the zone's DNSSEC keys on a schedule, according
2057   to the keys' timing metadata (see :ref:`man_dnssec-keygen` and
2058   :ref:`man_dnssec-settime`). The command ``rndc sign zonename``
2059   causes ``named`` to load keys from the key repository and sign the
2060   zone with all keys that are active. ``rndc loadkeys zonename``
2061   causes ``named`` to load keys from the key repository and schedule
2062   key maintenance events to occur in the future, but it does not sign
2063   the full zone immediately. Note: once keys have been loaded for a
2064   zone the first time, the repository is searched for changes
2065   periodically, regardless of whether ``rndc loadkeys`` is used. The
2066   recheck interval is defined by ``dnssec-loadkeys-interval``.
2067
2068   ``auto-dnssec off;`` does not allow for DNSSEC key management.
2069   This is the default setting.
2070
2071   This option may only be activated at the zone level; if configured
2072   at the view or options level, it must be set to ``off``.
2073
2074   The DNSSEC records are written to the zone's filename set in ``file``,
2075   unless ``inline-signing`` is enabled.
2076
2077``dnssec-enable``
2078   This option is obsolete and has no effect.
2079
2080.. _dnssec-validation-option:
2081
2082``dnssec-validation``
2083   This option enables DNSSEC validation in ``named``.
2084
2085   If set to ``auto``, DNSSEC validation is enabled and a default trust
2086   anchor for the DNS root zone is used. This trust anchor is provided
2087   as part of BIND and is kept up-to-date using :ref:`rfc5011.support` key
2088   management.
2089
2090   If set to ``yes``, DNSSEC validation is enabled, but a trust anchor must be
2091   manually configured using a ``trust-anchors`` statement (or the
2092   ``managed-keys`` or ``trusted-keys`` statements, both deprecated). If
2093   there is no configured trust anchor, validation does not take place.
2094
2095   If set to ``no``, DNSSEC validation is disabled.
2096
2097   The default is ``auto``, unless BIND is built with
2098   ``configure --disable-auto-validation``, in which case the default is
2099   ``yes``.
2100
2101   The default root trust anchor is stored in the file ``bind.keys``.
2102   ``named`` loads that key at startup if ``dnssec-validation`` is
2103   set to ``auto``. A copy of the file is installed along with BIND 9,
2104   and is current as of the release date. If the root key expires, a new
2105   copy of ``bind.keys`` can be downloaded from
2106   https://www.isc.org/bind-keys.
2107
2108   (To prevent problems if ``bind.keys`` is not found, the current trust
2109   anchor is also compiled in ``named``. Relying on this is not
2110   recommended, however, as it requires ``named`` to be recompiled with
2111   a new key when the root key expires.)
2112
2113   .. note:: ``named`` loads *only* the root key from ``bind.keys``. The file
2114         cannot be used to store keys for other zones. The root key in
2115         ``bind.keys`` is ignored if ``dnssec-validation auto`` is not in
2116         use.
2117
2118         Whenever the resolver sends out queries to an EDNS-compliant
2119         server, it always sets the DO bit indicating it can support DNSSEC
2120         responses, even if ``dnssec-validation`` is off.
2121
2122``validate-except``
2123   This specifies a list of domain names at and beneath which DNSSEC
2124   validation should *not* be performed, regardless of the presence of a
2125   trust anchor at or above those names. This may be used, for example,
2126   when configuring a top-level domain intended only for local use, so
2127   that the lack of a secure delegation for that domain in the root zone
2128   does not cause validation failures. (This is similar to setting a
2129   negative trust anchor except that it is a permanent configuration,
2130   whereas negative trust anchors expire and are removed after a set
2131   period of time.)
2132
2133``dnssec-accept-expired``
2134   This accepts expired signatures when verifying DNSSEC signatures. The
2135   default is ``no``. Setting this option to ``yes`` leaves ``named``
2136   vulnerable to replay attacks.
2137
2138``querylog``
2139   Query logging provides a complete log of all incoming queries and all query
2140   errors. This provides more insight into the server's activity, but with a
2141   cost to performance which may be significant on heavily loaded servers.
2142
2143   The ``querylog`` option specifies whether query logging should be active when
2144   ``named`` first starts.  If ``querylog`` is not specified, then query logging
2145   is determined by the presence of the logging category ``queries``.  Query
2146   logging can also be activated at runtime using the command ``rndc querylog
2147   on``, or deactivated with ``rndc querylog off``.
2148
2149``check-names``
2150   This option is used to restrict the character set and syntax of
2151   certain domain names in primary files and/or DNS responses received
2152   from the network. The default varies according to usage area. For
2153   ``primary`` zones the default is ``fail``. For ``secondary`` zones the
2154   default is ``warn``. For answers received from the network
2155   (``response``), the default is ``ignore``.
2156
2157   The rules for legal hostnames and mail domains are derived from
2158   :rfc:`952` and :rfc:`821` as modified by :rfc:`1123`.
2159
2160   ``check-names`` applies to the owner names of A, AAAA, and MX records.
2161   It also applies to the domain names in the RDATA of NS, SOA, MX, and
2162   SRV records. It further applies to the RDATA of PTR records where the
2163   owner name indicates that it is a reverse lookup of a hostname (the
2164   owner name ends in IN-ADDR.ARPA, IP6.ARPA, or IP6.INT).
2165
2166``check-dup-records``
2167   This checks primary zones for records that are treated as different by
2168   DNSSEC but are semantically equal in plain DNS. The default is to
2169   ``warn``. Other possible values are ``fail`` and ``ignore``.
2170
2171``check-mx``
2172   This checks whether the MX record appears to refer to an IP address. The
2173   default is to ``warn``. Other possible values are ``fail`` and
2174   ``ignore``.
2175
2176``check-wildcard``
2177   This option is used to check for non-terminal wildcards. The use of
2178   non-terminal wildcards is almost always as a result of a lack of
2179   understanding of the wildcard matching algorithm (:rfc:`1034`). This option
2180   affects primary zones. The default (``yes``) is to check for
2181   non-terminal wildcards and issue a warning.
2182
2183``check-integrity``
2184   This performs post-load zone integrity checks on primary zones. It checks
2185   that MX and SRV records refer to address (A or AAAA) records and that
2186   glue address records exist for delegated zones. For MX and SRV
2187   records, only in-zone hostnames are checked (for out-of-zone hostnames,
2188   use ``named-checkzone``). For NS records, only names below top-of-zone
2189   are checked (for out-of-zone names and glue consistency checks, use
2190   ``named-checkzone``). The default is ``yes``.
2191
2192   The use of the SPF record to publish Sender Policy Framework is
2193   deprecated, as the migration from using TXT records to SPF records was
2194   abandoned. Enabling this option also checks that a TXT Sender Policy
2195   Framework record exists (starts with "v=spf1") if there is an SPF
2196   record. Warnings are emitted if the TXT record does not exist; they can
2197   be suppressed with ``check-spf``.
2198
2199``check-mx-cname``
2200   If ``check-integrity`` is set, then fail, warn, or ignore MX records
2201   that refer to CNAMES. The default is to ``warn``.
2202
2203``check-srv-cname``
2204   If ``check-integrity`` is set, then fail, warn, or ignore SRV records
2205   that refer to CNAMES. The default is to ``warn``.
2206
2207``check-sibling``
2208   When performing integrity checks, also check that sibling glue
2209   exists. The default is ``yes``.
2210
2211``check-spf``
2212   If ``check-integrity`` is set, check that there is a TXT Sender
2213   Policy Framework record present (starts with "v=spf1") if there is an
2214   SPF record present. The default is ``warn``.
2215
2216``zero-no-soa-ttl``
2217   If ``yes``, when returning authoritative negative responses to SOA queries, set
2218   the TTL of the SOA record returned in the authority section to zero.
2219   The default is ``yes``.
2220
2221``zero-no-soa-ttl-cache``
2222   If ``yes``, when caching a negative response to an SOA query set the TTL to zero.
2223   The default is ``no``.
2224
2225``update-check-ksk``
2226   When set to the default value of ``yes``, check the KSK bit in each
2227   key to determine how the key should be used when generating RRSIGs
2228   for a secure zone.
2229
2230   Ordinarily, zone-signing keys (that is, keys without the KSK bit set)
2231   are used to sign the entire zone, while key-signing keys (keys with
2232   the KSK bit set) are only used to sign the DNSKEY RRset at the zone
2233   apex. However, if this option is set to ``no``, then the KSK bit is
2234   ignored; KSKs are treated as if they were ZSKs and are used to sign
2235   the entire zone. This is similar to the ``dnssec-signzone -z``
2236   command-line option.
2237
2238   When this option is set to ``yes``, there must be at least two active
2239   keys for every algorithm represented in the DNSKEY RRset: at least
2240   one KSK and one ZSK per algorithm. If there is any algorithm for
2241   which this requirement is not met, this option is ignored for
2242   that algorithm.
2243
2244``dnssec-dnskey-kskonly``
2245   When this option and ``update-check-ksk`` are both set to ``yes``,
2246   only key-signing keys (that is, keys with the KSK bit set) are
2247   used to sign the DNSKEY, CDNSKEY, and CDS RRsets at the zone apex.
2248   Zone-signing keys (keys without the KSK bit set) are used to sign
2249   the remainder of the zone, but not the DNSKEY RRset. This is similar
2250   to the ``dnssec-signzone -x`` command-line option.
2251
2252   The default is ``no``. If ``update-check-ksk`` is set to ``no``, this
2253   option is ignored.
2254
2255``try-tcp-refresh``
2256   If ``yes``, try to refresh the zone using TCP if UDP queries fail. The default is
2257   ``yes``.
2258
2259``dnssec-secure-to-insecure``
2260   This allows a dynamic zone to transition from secure to insecure (i.e.,
2261   signed to unsigned) by deleting all of the DNSKEY records. The
2262   default is ``no``. If set to ``yes``, and if the DNSKEY RRset at the
2263   zone apex is deleted, all RRSIG and NSEC records are removed from
2264   the zone as well.
2265
2266   If the zone uses NSEC3, it is also necessary to delete the
2267   NSEC3PARAM RRset from the zone apex; this causes the removal of
2268   all corresponding NSEC3 records. (It is expected that this
2269   requirement will be eliminated in a future release.)
2270
2271   Note that if a zone has been configured with ``auto-dnssec maintain``
2272   and the private keys remain accessible in the key repository,
2273   the zone will be automatically signed again the next time ``named``
2274   is started.
2275
2276``synth-from-dnssec``
2277   This option synthesizes answers from cached NSEC, NSEC3, and other RRsets that have been
2278   proved to be correct using DNSSEC. The default is ``no``, but it will become
2279   ``yes`` again in future releases.
2280
2281   .. note:: DNSSEC validation must be enabled for this option to be effective.
2282      This initial implementation only covers synthesis of answers from
2283      NSEC records; synthesis from NSEC3 is planned for the future. This
2284      will also be controlled by ``synth-from-dnssec``.
2285
2286Forwarding
2287^^^^^^^^^^
2288
2289The forwarding facility can be used to create a large site-wide cache on
2290a few servers, reducing traffic over links to external name servers. It
2291can also be used to allow queries by servers that do not have direct
2292access to the Internet, but wish to look up exterior names anyway.
2293Forwarding occurs only on those queries for which the server is not
2294authoritative and does not have the answer in its cache.
2295
2296``forward``
2297   This option is only meaningful if the forwarders list is not empty. A
2298   value of ``first`` is the default and causes the server to query the
2299   forwarders first; if that does not answer the question, the
2300   server then looks for the answer itself. If ``only`` is
2301   specified, the server only queries the forwarders.
2302
2303``forwarders``
2304   This specifies a list of IP addresses to which queries are forwarded. The
2305   default is the empty list (no forwarding). Each address in the list can be
2306   associated with an optional port number and/or DSCP value, and a default port
2307   number and DSCP value can be set for the entire list.
2308
2309Forwarding can also be configured on a per-domain basis, allowing for
2310the global forwarding options to be overridden in a variety of ways.
2311Particular domains can be set to use different forwarders, or have a
2312different ``forward only/first`` behavior, or not forward at all; see
2313:ref:`zone_statement_grammar`.
2314
2315.. _dual_stack:
2316
2317Dual-stack Servers
2318^^^^^^^^^^^^^^^^^^
2319
2320Dual-stack servers are used as servers of last resort, to work around
2321problems in reachability due to the lack of support for either IPv4 or IPv6
2322on the host machine.
2323
2324``dual-stack-servers``
2325   This specifies host names or addresses of machines with access to both
2326   IPv4 and IPv6 transports. If a hostname is used, the server must be
2327   able to resolve the name using only the transport it has. If the
2328   machine is dual-stacked, the ``dual-stack-servers`` parameter has no
2329   effect unless access to a transport has been disabled on the command
2330   line (e.g., ``named -4``).
2331
2332.. _access_control:
2333
2334Access Control
2335^^^^^^^^^^^^^^
2336
2337Access to the server can be restricted based on the IP address of the
2338requesting system. See :ref:`address_match_lists`
2339for details on how to specify IP address lists.
2340
2341``allow-notify``
2342   This ACL specifies which hosts may send NOTIFY messages to inform
2343   this server of changes to zones for which it is acting as a secondary
2344   server. This is only applicable for secondary zones (i.e., type
2345   ``secondary`` or ``slave``).
2346
2347   If this option is set in ``view`` or ``options``, it is globally
2348   applied to all secondary zones. If set in the ``zone`` statement, the
2349   global value is overridden.
2350
2351   If not specified, the default is to process NOTIFY messages only from
2352   the configured ``primaries`` for the zone. ``allow-notify`` can be used
2353   to expand the list of permitted hosts, not to reduce it.
2354
2355``allow-query``
2356   This specifies which hosts are allowed to ask ordinary DNS questions.
2357   ``allow-query`` may also be specified in the ``zone`` statement, in
2358   which case it overrides the ``options allow-query`` statement. If not
2359   specified, the default is to allow queries from all hosts.
2360
2361   .. note:: ``allow-query-cache`` is used to specify access to the cache.
2362
2363``allow-query-on``
2364   This specifies which local addresses can accept ordinary DNS questions.
2365   This makes it possible, for instance, to allow queries on
2366   internal-facing interfaces but disallow them on external-facing ones,
2367   without necessarily knowing the internal network's addresses.
2368
2369   Note that ``allow-query-on`` is only checked for queries that are
2370   permitted by ``allow-query``. A query must be allowed by both ACLs,
2371   or it is refused.
2372
2373   ``allow-query-on`` may also be specified in the ``zone`` statement,
2374   in which case it overrides the ``options allow-query-on`` statement.
2375
2376   If not specified, the default is to allow queries on all addresses.
2377
2378   .. note:: ``allow-query-cache`` is used to specify access to the cache.
2379
2380``allow-query-cache``
2381   This specifies which hosts are allowed to get answers from the cache. If
2382   ``allow-recursion`` is not set, BIND checks to see if the following parameters
2383   are set, in order: ``allow-query-cache`` and ``allow-query`` (unless ``recursion no;`` is set).
2384   If neither of those parameters is set, the default (localnets; localhost;) is used.
2385
2386``allow-query-cache-on``
2387   This specifies which local addresses can send answers from the cache. If
2388   ``allow-query-cache-on`` is not set, then ``allow-recursion-on`` is
2389   used if set. Otherwise, the default is to allow cache responses to be
2390   sent from any address. Note: both ``allow-query-cache`` and
2391   ``allow-query-cache-on`` must be satisfied before a cache response
2392   can be sent; a client that is blocked by one cannot be allowed by the
2393   other.
2394
2395``allow-recursion``
2396   This specifies which hosts are allowed to make recursive queries through
2397   this server. BIND checks to see if the following parameters are set, in
2398   order: ``allow-query-cache`` and ``allow-query``. If neither of those parameters
2399   is set, the default (localnets; localhost;) is used.
2400
2401``allow-recursion-on``
2402   This specifies which local addresses can accept recursive queries. If
2403   ``allow-recursion-on`` is not set, then ``allow-query-cache-on`` is
2404   used if set; otherwise, the default is to allow recursive queries on
2405   all addresses. Any client permitted to send recursive queries can
2406   send them to any address on which ``named`` is listening. Note: both
2407   ``allow-recursion`` and ``allow-recursion-on`` must be satisfied
2408   before recursion is allowed; a client that is blocked by one cannot
2409   be allowed by the other.
2410
2411``allow-update``
2412   When set in the ``zone`` statement for a primary zone, this specifies which
2413   hosts are allowed to submit Dynamic DNS updates to that zone. The
2414   default is to deny updates from all hosts.
2415
2416   Note that allowing updates based on the requestor's IP address is
2417   insecure; see :ref:`dynamic_update_security` for details.
2418
2419   In general, this option should only be set at the ``zone`` level.
2420   While a default value can be set at the ``options`` or ``view`` level
2421   and inherited by zones, this could lead to some zones unintentionally
2422   allowing updates.
2423
2424   Updates are written to the zone's filename that is set in ``file``.
2425
2426``allow-update-forwarding``
2427   When set in the ``zone`` statement for a secondary zone, this specifies which
2428   hosts are allowed to submit Dynamic DNS updates and have them be
2429   forwarded to the primary. The default is ``{ none; }``, which means
2430   that no update forwarding is performed.
2431
2432   To enable update forwarding, specify
2433   ``allow-update-forwarding { any; };`` in the ``zone`` statement.
2434   Specifying values other than ``{ none; }`` or ``{ any; }`` is usually
2435   counterproductive; the responsibility for update access control
2436   should rest with the primary server, not the secondary.
2437
2438   Note that enabling the update forwarding feature on a secondary server
2439   may expose primary servers to attacks if they rely on insecure
2440   IP-address-based access control; see :ref:`dynamic_update_security` for more details.
2441
2442   In general this option should only be set at the ``zone`` level.
2443   While a default value can be set at the ``options`` or ``view`` level
2444   and inherited by zones, this can lead to some zones unintentionally
2445   forwarding updates.
2446
2447.. _allow-transfer-access:
2448
2449``allow-transfer``
2450   This specifies which hosts are allowed to receive zone transfers from the
2451   server. ``allow-transfer`` may also be specified in the ``zone``
2452   statement, in which case it overrides the ``allow-transfer``
2453   statement set in ``options`` or ``view``. If not specified, the
2454   default is to allow transfers to all hosts.
2455
2456``blackhole``
2457   This specifies a list of addresses which the server does not accept queries
2458   from or use to resolve a query. Queries from these addresses are not
2459   responded to. The default is ``none``.
2460
2461``keep-response-order``
2462   This specifies a list of addresses to which the server sends responses
2463   to TCP queries, in the same order in which they were received. This
2464   disables the processing of TCP queries in parallel. The default is
2465   ``none``.
2466
2467``no-case-compress``
2468   This specifies a list of addresses which require responses to use
2469   case-insensitive compression. This ACL can be used when ``named``
2470   needs to work with clients that do not comply with the requirement in
2471   :rfc:`1034` to use case-insensitive name comparisons when checking for
2472   matching domain names.
2473
2474   If left undefined, the ACL defaults to ``none``: case-insensitive
2475   compression is used for all clients. If the ACL is defined and
2476   matches a client, case is ignored when compressing domain
2477   names in DNS responses sent to that client.
2478
2479   This can result in slightly smaller responses; if a response contains
2480   the names "example.com" and "example.COM", case-insensitive
2481   compression treats the second one as a duplicate. It also
2482   ensures that the case of the query name exactly matches the case of
2483   the owner names of returned records, rather than matches the case of
2484   the records entered in the zone file. This allows responses to
2485   exactly match the query, which is required by some clients due to
2486   incorrect use of case-sensitive comparisons.
2487
2488   Case-insensitive compression is *always* used in AXFR and IXFR
2489   responses, regardless of whether the client matches this ACL.
2490
2491   There are circumstances in which ``named`` does not preserve the case
2492   of owner names of records: if a zone file defines records of
2493   different types with the same name, but the capitalization of the
2494   name is different (e.g., "www.example.com/A" and
2495   "WWW.EXAMPLE.COM/AAAA"), then all responses for that name use
2496   the *first* version of the name that was used in the zone file. This
2497   limitation may be addressed in a future release. However, domain
2498   names specified in the rdata of resource records (i.e., records of
2499   type NS, MX, CNAME, etc.) always have their case preserved unless
2500   the client matches this ACL.
2501
2502``resolver-query-timeout``
2503   This is the amount of time in milliseconds that the resolver spends
2504   attempting to resolve a recursive query before failing. The default
2505   and minimum is ``10000`` and the maximum is ``30000``. Setting it to
2506   ``0`` results in the default being used.
2507
2508   This value was originally specified in seconds. Values less than or
2509   equal to 300 are treated as seconds and converted to
2510   milliseconds before applying the above limits.
2511
2512Interfaces
2513^^^^^^^^^^
2514
2515The interfaces and ports that the server answers queries from may be
2516specified using the ``listen-on`` option. ``listen-on`` takes an
2517optional port and an ``address_match_list`` of IPv4 addresses. (IPv6
2518addresses are ignored, with a logged warning.) The server listens on
2519all interfaces allowed by the address match list. If a port is not
2520specified, port 53 is used.
2521
2522Multiple ``listen-on`` statements are allowed. For example:
2523
2524::
2525
2526   listen-on { 5.6.7.8; };
2527   listen-on port 1234 { !1.2.3.4; 1.2/16; };
2528
2529enables the name server on port 53 for the IP address 5.6.7.8, and
2530on port 1234 of an address on the machine in net 1.2 that is not
25311.2.3.4.
2532
2533If no ``listen-on`` is specified, the server listens on port 53 on
2534all IPv4 interfaces.
2535
2536The ``listen-on-v6`` option is used to specify the interfaces and the
2537ports on which the server listens for incoming queries sent using
2538IPv6. If not specified, the server listens on port 53 on all IPv6
2539interfaces.
2540
2541Multiple ``listen-on-v6`` options can be used. For example:
2542
2543::
2544
2545   listen-on-v6 { any; };
2546   listen-on-v6 port 1234 { !2001:db8::/32; any; };
2547
2548enables the name server on port 53 for any IPv6 addresses (with a
2549single wildcard socket), and on port 1234 of IPv6 addresses that are not
2550in the prefix 2001:db8::/32 (with separate sockets for each matched
2551address).
2552
2553To instruct the server not to listen on any IPv6 address, use:
2554
2555::
2556
2557   listen-on-v6 { none; };
2558
2559.. _query_address:
2560
2561Query Address
2562^^^^^^^^^^^^^
2563
2564If the server does not know the answer to a question, it queries other
2565name servers. ``query-source`` specifies the address and port used for
2566such queries. For queries sent over IPv6, there is a separate
2567``query-source-v6`` option. If ``address`` is ``*`` (asterisk) or is
2568omitted, a wildcard IP address (``INADDR_ANY``) is used.
2569
2570If ``port`` is ``*`` or is omitted, a random port number from a
2571pre-configured range is picked up and used for each query. The
2572port range(s) is specified in the ``use-v4-udp-ports`` (for IPv4)
2573and ``use-v6-udp-ports`` (for IPv6) options, excluding the ranges
2574specified in the ``avoid-v4-udp-ports`` and ``avoid-v6-udp-ports``
2575options, respectively.
2576
2577The defaults of the ``query-source`` and ``query-source-v6`` options
2578are:
2579
2580::
2581
2582   query-source address * port *;
2583   query-source-v6 address * port *;
2584
2585If ``use-v4-udp-ports`` or ``use-v6-udp-ports`` is unspecified,
2586``named`` checks whether the operating system provides a programming
2587interface to retrieve the system's default range for ephemeral ports. If
2588such an interface is available, ``named`` uses the corresponding
2589system default range; otherwise, it uses its own defaults:
2590
2591::
2592
2593   use-v4-udp-ports { range 1024 65535; };
2594   use-v6-udp-ports { range 1024 65535; };
2595
2596.. note:: Make sure the ranges are sufficiently large for security. A
2597   desirable size depends on several parameters, but we generally recommend
2598   it contain at least 16384 ports (14 bits of entropy). Note also that the
2599   system's default range when used may be too small for this purpose, and
2600   that the range may even be changed while ``named`` is running; the new
2601   range is automatically applied when ``named`` is reloaded. Explicit
2602   configuration of ``use-v4-udp-ports`` and ``use-v6-udp-ports`` is encouraged,
2603   so that the ranges are sufficiently large and are reasonably
2604   independent from the ranges used by other applications.
2605
2606.. note:: The operational configuration where ``named`` runs may prohibit
2607   the use of some ports. For example, Unix systems do not allow
2608   ``named``, if run without root privilege, to use ports less than 1024.
2609   If such ports are included in the specified (or detected) set of query
2610   ports, the corresponding query attempts will fail, resulting in
2611   resolution failures or delay. It is therefore important to configure the
2612   set of ports that can be safely used in the expected operational
2613   environment.
2614
2615The defaults of the ``avoid-v4-udp-ports`` and ``avoid-v6-udp-ports``
2616options are:
2617
2618::
2619
2620   avoid-v4-udp-ports {};
2621   avoid-v6-udp-ports {};
2622
2623.. note:: BIND 9.5.0 introduced the ``use-queryport-pool`` option to support
2624   a pool of such random ports, but this option is now obsolete because
2625   reusing the same ports in the pool may not be sufficiently secure. For
2626   the same reason, it is generally strongly discouraged to specify a
2627   particular port for the ``query-source`` or ``query-source-v6`` options;
2628   it implicitly disables the use of randomized port numbers.
2629
2630``use-queryport-pool``
2631   This option is obsolete.
2632
2633``queryport-pool-ports``
2634   This option is obsolete.
2635
2636``queryport-pool-updateinterval``
2637   This option is obsolete.
2638
2639.. note:: The address specified in the ``query-source`` option is used for both
2640   UDP and TCP queries, but the port applies only to UDP queries. TCP
2641   queries always use a random unprivileged port.
2642
2643.. warning:: Specifying a single port is discouraged, as it removes a layer of
2644   protection against spoofing errors.
2645
2646.. warning:: The configured ``port`` must not be same as the listening port.
2647
2648.. note:: See also ``transfer-source``, ``notify-source`` and ``parental-source``.
2649
2650.. _zone_transfers:
2651
2652Zone Transfers
2653^^^^^^^^^^^^^^
2654
2655BIND has mechanisms in place to facilitate zone transfers and set limits
2656on the amount of load that transfers place on the system. The following
2657options apply to zone transfers.
2658
2659``also-notify``
2660   This option defines a global list of IP addresses of name servers that are also
2661   sent NOTIFY messages whenever a fresh copy of the zone is loaded, in
2662   addition to the servers listed in the zone's NS records. This helps
2663   to ensure that copies of the zones quickly converge on stealth
2664   servers. Optionally, a port may be specified with each
2665   ``also-notify`` address to send the notify messages to a port other
2666   than the default of 53. An optional TSIG key can also be specified
2667   with each address to cause the notify messages to be signed; this can
2668   be useful when sending notifies to multiple views. In place of
2669   explicit addresses, one or more named ``primaries`` lists can be used.
2670
2671   If an ``also-notify`` list is given in a ``zone`` statement, it
2672   overrides the ``options also-notify`` statement. When a
2673   ``zone notify`` statement is set to ``no``, the IP addresses in the
2674   global ``also-notify`` list are not sent NOTIFY messages for that
2675   zone. The default is the empty list (no global notification list).
2676
2677``max-transfer-time-in``
2678   Inbound zone transfers running longer than this many minutes are
2679   terminated. The default is 120 minutes (2 hours). The maximum value
2680   is 28 days (40320 minutes).
2681
2682``max-transfer-idle-in``
2683   Inbound zone transfers making no progress in this many minutes are
2684   terminated. The default is 60 minutes (1 hour). The maximum value
2685   is 28 days (40320 minutes).
2686
2687``max-transfer-time-out``
2688   Outbound zone transfers running longer than this many minutes are
2689   terminated. The default is 120 minutes (2 hours). The maximum value
2690   is 28 days (40320 minutes).
2691
2692``max-transfer-idle-out``
2693   Outbound zone transfers making no progress in this many minutes are
2694   terminated. The default is 60 minutes (1 hour). The maximum value
2695   is 28 days (40320 minutes).
2696
2697``notify-rate``
2698   This specifies the rate at which NOTIFY requests are sent during normal zone
2699   maintenance operations. (NOTIFY requests due to initial zone loading
2700   are subject to a separate rate limit; see below.) The default is 20
2701   per second. The lowest possible rate is one per second; when set to
2702   zero, it is silently raised to one.
2703
2704``startup-notify-rate``
2705   This is the rate at which NOTIFY requests are sent when the name server
2706   is first starting up, or when zones have been newly added to the
2707   name server. The default is 20 per second. The lowest possible rate is
2708   one per second; when set to zero, it is silently raised to one.
2709
2710``serial-query-rate``
2711   Secondary servers periodically query primary servers to find out if
2712   zone serial numbers have changed. Each such query uses a minute
2713   amount of the secondary server's network bandwidth. To limit the amount
2714   of bandwidth used, BIND 9 limits the rate at which queries are sent.
2715   The value of the ``serial-query-rate`` option, an integer, is the
2716   maximum number of queries sent per second. The default is 20 per
2717   second. The lowest possible rate is one per second; when set to zero,
2718   it is silently raised to one.
2719
2720``transfer-format``
2721   Zone transfers can be sent using two different formats,
2722   ``one-answer`` and ``many-answers``. The ``transfer-format`` option
2723   is used on the primary server to determine which format it sends.
2724   ``one-answer`` uses one DNS message per resource record transferred.
2725   ``many-answers`` packs as many resource records as possible into one
2726   message. ``many-answers`` is more efficient; the default is ``many-answers``.
2727   ``transfer-format`` may be overridden on a per-server basis by using
2728   the ``server`` statement.
2729
2730``transfer-message-size``
2731   This is an upper bound on the uncompressed size of DNS messages used
2732   in zone transfers over TCP. If a message grows larger than this size,
2733   additional messages are used to complete the zone transfer.
2734   (Note, however, that this is a hint, not a hard limit; if a message
2735   contains a single resource record whose RDATA does not fit within the
2736   size limit, a larger message will be permitted so the record can be
2737   transferred.)
2738
2739   Valid values are between 512 and 65535 octets; any values outside
2740   that range are adjusted to the nearest value within it. The
2741   default is ``20480``, which was selected to improve message
2742   compression; most DNS messages of this size will compress to less
2743   than 16536 bytes. Larger messages cannot be compressed as
2744   effectively, because 16536 is the largest permissible compression
2745   offset pointer in a DNS message.
2746
2747   This option is mainly intended for server testing; there is rarely
2748   any benefit in setting a value other than the default.
2749
2750``transfers-in``
2751   This is the maximum number of inbound zone transfers that can run
2752   concurrently. The default value is ``10``. Increasing
2753   ``transfers-in`` may speed up the convergence of secondary zones, but it
2754   also may increase the load on the local system.
2755
2756``transfers-out``
2757   This is the maximum number of outbound zone transfers that can run
2758   concurrently. Zone transfer requests in excess of the limit are
2759   refused. The default value is ``10``.
2760
2761``transfers-per-ns``
2762   This is the maximum number of inbound zone transfers that can concurrently
2763   transfer from a given remote name server. The default value is
2764   ``2``. Increasing ``transfers-per-ns`` may speed up the convergence
2765   of secondary zones, but it also may increase the load on the remote name
2766   server. ``transfers-per-ns`` may be overridden on a per-server basis
2767   by using the ``transfers`` phrase of the ``server`` statement.
2768
2769``transfer-source``
2770   ``transfer-source`` determines which local address is bound to
2771   IPv4 TCP connections used to fetch zones transferred inbound by the
2772   server. It also determines the source IPv4 address, and optionally
2773   the UDP port, used for the refresh queries and forwarded dynamic
2774   updates. If not set, it defaults to a system-controlled value which
2775   is usually the address of the interface "closest to" the remote
2776   end. This address must appear in the remote end's ``allow-transfer``
2777   option for the zone being transferred, if one is specified. This
2778   statement sets the ``transfer-source`` for all zones, but can be
2779   overridden on a per-view or per-zone basis by including a
2780   ``transfer-source`` statement within the ``view`` or ``zone`` block
2781   in the configuration file.
2782
2783   .. warning:: Specifying a single port is discouraged, as it removes a layer of
2784      protection against spoofing errors.
2785
2786   .. warning:: The configured ``port`` must not be same as the listening port.
2787
2788``transfer-source-v6``
2789   This option is the same as ``transfer-source``, except zone transfers are performed
2790   using IPv6.
2791
2792``alt-transfer-source``
2793   This indicates an alternate transfer source if the one listed in ``transfer-source``
2794   fails and ``use-alt-transfer-source`` is set.
2795
2796   .. note:: To avoid using the alternate transfer source,
2797      set ``use-alt-transfer-source`` appropriately and
2798      do not depend upon getting an answer back to the first refresh
2799      query.
2800
2801``alt-transfer-source-v6``
2802   This indicates an alternate transfer source if the one listed in
2803   ``transfer-source-v6`` fails and ``use-alt-transfer-source`` is set.
2804
2805``use-alt-transfer-source``
2806   This indicates whether the alternate transfer sources should be used. If views are specified,
2807   this defaults to ``no``; otherwise, it defaults to ``yes``.
2808
2809``notify-source``
2810   ``notify-source`` determines which local source address, and
2811   optionally UDP port, is used to send NOTIFY messages. This
2812   address must appear in the secondary server's ``primaries`` zone clause or
2813   in an ``allow-notify`` clause. This statement sets the
2814   ``notify-source`` for all zones, but can be overridden on a per-zone
2815   or per-view basis by including a ``notify-source`` statement within
2816   the ``zone`` or ``view`` block in the configuration file.
2817
2818   .. warning:: Specifying a single port is discouraged, as it removes a layer of
2819      protection against spoofing errors.
2820
2821   .. warning:: The configured ``port`` must not be same as the listening port.
2822
2823``notify-source-v6``
2824   This option acts like ``notify-source``, but applies to notify messages sent to IPv6
2825   addresses.
2826
2827.. _port_lists:
2828
2829UDP Port Lists
2830^^^^^^^^^^^^^^
2831
2832``use-v4-udp-ports``, ``avoid-v4-udp-ports``, ``use-v6-udp-ports``, and
2833``avoid-v6-udp-ports`` specify a list of IPv4 and IPv6 UDP ports that
2834are or are not used as source ports for UDP messages. See
2835:ref:`query_address` about how the available ports are
2836determined. For example, with the following configuration:
2837
2838::
2839
2840   use-v6-udp-ports { range 32768 65535; };
2841   avoid-v6-udp-ports { 40000; range 50000 60000; };
2842
2843UDP ports of IPv6 messages sent from ``named`` are in one of the
2844following ranges: 32768 to 39999, 40001 to 49999, and 60001 to 65535.
2845
2846``avoid-v4-udp-ports`` and ``avoid-v6-udp-ports`` can be used to prevent
2847``named`` from choosing as its random source port a port that is blocked
2848by a firewall or a port that is used by other applications; if a
2849query went out with a source port blocked by a firewall, the answer
2850would not pass through the firewall and the name server would have to query
2851again. Note: the desired range can also be represented only with
2852``use-v4-udp-ports`` and ``use-v6-udp-ports``, and the ``avoid-``
2853options are redundant in that sense; they are provided for backward
2854compatibility and to possibly simplify the port specification.
2855
2856.. _resource_limits:
2857
2858Operating System Resource Limits
2859^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2860
2861The server's usage of many system resources can be limited. Scaled
2862values are allowed when specifying resource limits. For example, ``1G``
2863can be used instead of ``1073741824`` to specify a limit of one
2864gigabyte. ``unlimited`` requests unlimited use, or the maximum available
2865amount. ``default`` uses the limit that was in force when the server was
2866started. See the description of ``size_spec`` in :ref:`configuration_file_elements`.
2867
2868The following options set operating system resource limits for the name
2869server process. Some operating systems do not support some or any of the
2870limits; on such systems, a warning is issued if an unsupported
2871limit is used.
2872
2873``coresize``
2874   This sets the maximum size of a core dump. The default is ``default``.
2875
2876``datasize``
2877   This sets the maximum amount of data memory the server may use. The default is
2878   ``default``. This is a hard limit on server memory usage; if the
2879   server attempts to allocate memory in excess of this limit, the
2880   allocation will fail, which may in turn leave the server unable to
2881   perform DNS service. Therefore, this option is rarely useful as a way
2882   to limit the amount of memory used by the server, but it can be
2883   used to raise an operating system data size limit that is too small
2884   by default. To limit the amount of memory used by the
2885   server, use the ``max-cache-size`` and ``recursive-clients`` options
2886   instead.
2887
2888``files``
2889   This sets the maximum number of files the server may have open concurrently.
2890   The default is ``unlimited``.
2891
2892``stacksize``
2893   This sets the maximum amount of stack memory the server may use. The default is
2894   ``default``.
2895
2896.. _server_resource_limits:
2897
2898Server Resource Limits
2899^^^^^^^^^^^^^^^^^^^^^^
2900
2901The following options set limits on the server's resource consumption
2902that are enforced internally by the server rather than by the operating
2903system.
2904
2905``max-journal-size``
2906   This sets a maximum size for each journal file (see :ref:`journal`),
2907   expressed in bytes or, if followed by an
2908   optional unit suffix ('k', 'm', or 'g'), in kilobytes, megabytes, or
2909   gigabytes. When the journal file approaches the specified size, some
2910   of the oldest transactions in the journal are automatically
2911   removed. The largest permitted value is 2 gigabytes. Very small
2912   values are rounded up to 4096 bytes. It is possible to specify ``unlimited``,
2913   which also means 2 gigabytes. If the limit is set to ``default`` or
2914   left unset, the journal is allowed to grow up to twice as large
2915   as the zone. (There is little benefit in storing larger journals.)
2916
2917   This option may also be set on a per-zone basis.
2918
2919``max-records``
2920   This sets the maximum number of records permitted in a zone. The default is
2921   zero, which means the maximum is unlimited.
2922
2923``recursive-clients``
2924   This sets the maximum number (a "hard quota") of simultaneous recursive lookups
2925   the server performs on behalf of clients. The default is
2926   ``1000``. Because each recursing client uses a fair bit of memory (on
2927   the order of 20 kilobytes), the value of the ``recursive-clients``
2928   option may have to be decreased on hosts with limited memory.
2929
2930   ``recursive-clients`` defines a "hard quota" limit for pending
2931   recursive clients; when more clients than this are pending, new
2932   incoming requests are not accepted, and for each incoming request
2933   a previous pending request is dropped.
2934
2935   A "soft quota" is also set. When this lower quota is exceeded,
2936   incoming requests are accepted, but for each one, a pending request
2937   is dropped. If ``recursive-clients`` is greater than 1000, the
2938   soft quota is set to ``recursive-clients`` minus 100; otherwise it is
2939   set to 90% of ``recursive-clients``.
2940
2941``tcp-clients``
2942   This is the maximum number of simultaneous client TCP connections that the
2943   server accepts. The default is ``150``.
2944
2945.. _clients-per-query:
2946
2947``clients-per-query``; ``max-clients-per-query``
2948   These set the initial value (minimum) and maximum number of recursive
2949   simultaneous clients for any given query (<qname,qtype,qclass>) that
2950   the server accepts before dropping additional clients. ``named``
2951   attempts to self-tune this value and changes are logged. The
2952   default values are 10 and 100.
2953
2954   This value should reflect how many queries come in for a given name
2955   in the time it takes to resolve that name. If the number of queries
2956   exceeds this value, ``named`` assumes that it is dealing with a
2957   non-responsive zone and drops additional queries. If it gets a
2958   response after dropping queries, it raises the estimate. The
2959   estimate is then lowered in 20 minutes if it has remained
2960   unchanged.
2961
2962   If ``clients-per-query`` is set to zero, there is no limit on
2963   the number of clients per query and no queries are dropped.
2964
2965   If ``max-clients-per-query`` is set to zero, there is no upper
2966   bound other than that imposed by ``recursive-clients``.
2967
2968``fetches-per-zone``
2969   This sets the maximum number of simultaneous iterative queries to any one
2970   domain that the server permits before blocking new queries for
2971   data in or beneath that zone. This value should reflect how many
2972   fetches would normally be sent to any one zone in the time it would
2973   take to resolve them. It should be smaller than
2974   ``recursive-clients``.
2975
2976   When many clients simultaneously query for the same name and type,
2977   the clients are all attached to the same fetch, up to the
2978   ``max-clients-per-query`` limit, and only one iterative query is
2979   sent. However, when clients are simultaneously querying for
2980   *different* names or types, multiple queries are sent and
2981   ``max-clients-per-query`` is not effective as a limit.
2982
2983   Optionally, this value may be followed by the keyword ``drop`` or
2984   ``fail``, indicating whether queries which exceed the fetch quota for
2985   a zone are dropped with no response, or answered with SERVFAIL.
2986   The default is ``drop``.
2987
2988   If ``fetches-per-zone`` is set to zero, there is no limit on the
2989   number of fetches per query and no queries are dropped. The
2990   default is zero.
2991
2992   The current list of active fetches can be dumped by running
2993   ``rndc recursing``. The list includes the number of active fetches
2994   for each domain and the number of queries that have been passed
2995   (allowed) or dropped (spilled) as a result of the ``fetches-per-zone``
2996   limit. (Note: these counters are not cumulative over time;
2997   whenever the number of active fetches for a domain drops to zero,
2998   the counter for that domain is deleted, and the next time a fetch
2999   is sent to that domain, it is recreated with the counters set
3000   to zero.)
3001
3002``fetches-per-server``
3003   This sets the maximum number of simultaneous iterative queries that the server
3004   allows to be sent to a single upstream name server before
3005   blocking additional queries. This value should reflect how many
3006   fetches would normally be sent to any one server in the time it would
3007   take to resolve them. It should be smaller than
3008   ``recursive-clients``.
3009
3010   Optionally, this value may be followed by the keyword ``drop`` or
3011   ``fail``, indicating whether queries are dropped with no
3012   response or answered with SERVFAIL, when all of the servers
3013   authoritative for a zone are found to have exceeded the per-server
3014   quota. The default is ``fail``.
3015
3016   If ``fetches-per-server`` is set to zero, there is no limit on
3017   the number of fetches per query and no queries are dropped. The
3018   default is zero.
3019
3020   The ``fetches-per-server`` quota is dynamically adjusted in response
3021   to detected congestion. As queries are sent to a server and either are
3022   answered or time out, an exponentially weighted moving average
3023   is calculated of the ratio of timeouts to responses. If the current
3024   average timeout ratio rises above a "high" threshold, then
3025   ``fetches-per-server`` is reduced for that server. If the timeout
3026   ratio drops below a "low" threshold, then ``fetches-per-server`` is
3027   increased. The ``fetch-quota-params`` options can be used to adjust
3028   the parameters for this calculation.
3029
3030``fetch-quota-params``
3031   This sets the parameters to use for dynamic resizing of the
3032   ``fetches-per-server`` quota in response to detected congestion.
3033
3034   The first argument is an integer value indicating how frequently to
3035   recalculate the moving average of the ratio of timeouts to responses
3036   for each server. The default is 100, meaning that BIND recalculates the
3037   average ratio after every 100 queries have either been answered or
3038   timed out.
3039
3040   The remaining three arguments represent the "low" threshold
3041   (defaulting to a timeout ratio of 0.1), the "high" threshold
3042   (defaulting to a timeout ratio of 0.3), and the discount rate for the
3043   moving average (defaulting to 0.7). A higher discount rate causes
3044   recent events to weigh more heavily when calculating the moving
3045   average; a lower discount rate causes past events to weigh more
3046   heavily, smoothing out short-term blips in the timeout ratio. These
3047   arguments are all fixed-point numbers with precision of 1/100; at
3048   most two places after the decimal point are significant.
3049
3050``reserved-sockets``
3051   This sets the number of file descriptors reserved for TCP, stdio, etc. This
3052   needs to be big enough to cover the number of interfaces ``named``
3053   listens on plus ``tcp-clients``, as well as to provide room for
3054   outgoing TCP queries and incoming zone transfers. The default is
3055   ``512``. The minimum value is ``128`` and the maximum value is
3056   ``128`` fewer than maxsockets (-S). This option may be removed in the
3057   future.
3058
3059   This option has little effect on Windows.
3060
3061``max-cache-size``
3062   This sets the maximum amount of memory to use for an individual cache
3063   database and its associated metadata, in bytes or percentage of total
3064   physical memory. By default, each view has its own separate cache,
3065   which means the total amount of memory required for cache data is the
3066   sum of the cache database sizes for all views (unless the
3067   :ref:`attach-cache <attach-cache>` option is used).
3068
3069   When the amount of data in a cache database reaches the configured
3070   limit, ``named`` starts purging non-expired records (following an
3071   LRU-based strategy).
3072
3073   The default size limit for each individual cache is:
3074
3075     - 90% of physical memory for views with ``recursion`` set to
3076       ``yes`` (the default), or
3077
3078     - 2 MB for views with ``recursion`` set to ``no``.
3079
3080   Any positive value smaller than 2 MB is ignored and reset to 2 MB.
3081   The keyword ``unlimited``, or the value ``0``, places no limit on the
3082   cache size; records are then purged from the cache only when they
3083   expire (according to their TTLs).
3084
3085   .. note::
3086
3087       For configurations which define multiple views with separate
3088       caches and recursion enabled, it is recommended to set
3089       ``max-cache-size`` appropriately for each view, as using the
3090       default value of that option (90% of physical memory for each
3091       individual cache) may lead to memory exhaustion over time.
3092
3093   Upon startup and reconfiguration, caches with a limited size
3094   preallocate a small amount of memory (less than 1% of
3095   ``max-cache-size`` for a given view). This preallocation serves as an
3096   optimization to eliminate extra latency introduced by resizing
3097   internal cache structures.
3098
3099   On systems where detection of the amount of physical memory is not
3100   supported, percentage-based values fall back to ``unlimited``. Note
3101   that the amount of physical memory available is only detected on
3102   startup, so ``named`` does not adjust the cache size limits if the
3103   amount of physical memory is changed at runtime.
3104
3105``tcp-listen-queue``
3106   This sets the listen-queue depth. The default and minimum is 10. If the kernel
3107   supports the accept filter "dataready", this also controls how many
3108   TCP connections are queued in kernel space waiting for some
3109   data before being passed to accept. Non-zero values less than 10 are
3110   silently raised. A value of 0 may also be used; on most platforms
3111   this sets the listen-queue length to a system-defined default value.
3112
3113``tcp-initial-timeout``
3114   This sets the amount of time (in units of 100 milliseconds) that the server waits on
3115   a new TCP connection for the first message from the client. The
3116   default is 300 (30 seconds), the minimum is 25 (2.5 seconds), and the
3117   maximum is 1200 (two minutes). Values above the maximum or below the
3118   minimum are adjusted with a logged warning. (Note: this value
3119   must be greater than the expected round-trip delay time; otherwise, no
3120   client will ever have enough time to submit a message.) This value
3121   can be updated at runtime by using ``rndc tcp-timeouts``.
3122
3123``tcp-idle-timeout``
3124   This sets the amount of time (in units of 100 milliseconds) that the server waits on
3125   an idle TCP connection before closing it, when the client is not using
3126   the EDNS TCP keepalive option. The default is 300 (30 seconds), the
3127   maximum is 1200 (two minutes), and the minimum is 1 (one-tenth of a
3128   second). Values above the maximum or below the minimum are
3129   adjusted with a logged warning. See ``tcp-keepalive-timeout`` for
3130   clients using the EDNS TCP keepalive option. This value can be
3131   updated at runtime by using ``rndc tcp-timeouts``.
3132
3133``tcp-keepalive-timeout``
3134   This sets the amount of time (in units of 100 milliseconds) that the server waits on
3135   an idle TCP connection before closing it, when the client is using the
3136   EDNS TCP keepalive option. The default is 300 (30 seconds), the
3137   maximum is 65535 (about 1.8 hours), and the minimum is 1 (one-tenth
3138   of a second). Values above the maximum or below the minimum are
3139   adjusted with a logged warning. This value may be greater than
3140   ``tcp-idle-timeout`` because clients using the EDNS TCP keepalive
3141   option are expected to use TCP connections for more than one message.
3142   This value can be updated at runtime by using ``rndc tcp-timeouts``.
3143
3144``tcp-advertised-timeout``
3145   This sets the timeout value (in units of 100 milliseconds) that the server sends
3146   in responses containing the EDNS TCP keepalive option, which informs a
3147   client of the amount of time it may keep the session open. The
3148   default is 300 (30 seconds), the maximum is 65535 (about 1.8 hours),
3149   and the minimum is 0, which signals that the clients must close TCP
3150   connections immediately. Ordinarily this should be set to the same
3151   value as ``tcp-keepalive-timeout``. This value can be updated at
3152   runtime by using ``rndc tcp-timeouts``.
3153
3154``update-quota``
3155   This is the maximum number of simultaneous DNS UPDATE messages that
3156   the server will accept for updating local authoritiative zones or
3157   forwarding to a primary server. The default is ``100``.
3158
3159.. _intervals:
3160
3161Periodic Task Intervals
3162^^^^^^^^^^^^^^^^^^^^^^^
3163
3164``cleaning-interval``
3165   This option is obsolete.
3166
3167``heartbeat-interval``
3168   The server performs zone maintenance tasks for all zones marked
3169   as ``dialup`` whenever this interval expires. The default is 60
3170   minutes. Reasonable values are up to 1 day (1440 minutes). The
3171   maximum value is 28 days (40320 minutes). If set to 0, no zone
3172   maintenance for these zones occurs.
3173
3174``interface-interval``
3175   The server scans the network interface list every ``interface-interval``
3176   minutes. The default is 60 minutes; the maximum value is 28 days (40320
3177   minutes). If set to 0, interface scanning only occurs when the configuration
3178   file is loaded, or when ``automatic-interface-scan`` is enabled and supported
3179   by the operating system. After the scan, the server begins listening for
3180   queries on any newly discovered interfaces (provided they are allowed by the
3181   ``listen-on`` configuration), and stops listening on interfaces that have
3182   gone away. For convenience, TTL-style time-unit suffixes may be used to
3183   specify the value. It also accepts ISO 8601 duration formats.
3184
3185.. _the_sortlist_statement:
3186
3187The ``sortlist`` Statement
3188^^^^^^^^^^^^^^^^^^^^^^^^^^
3189
3190The response to a DNS query may consist of multiple resource records
3191(RRs) forming a resource record set (RRset). The name server
3192normally returns the RRs within the RRset in an indeterminate order (but
3193see the ``rrset-order`` statement in :ref:`rrset_ordering`). The client resolver code should
3194rearrange the RRs as appropriate: that is, using any addresses on the
3195local net in preference to other addresses. However, not all resolvers
3196can do this or are correctly configured. When a client is using a local
3197server, the sorting can be performed in the server, based on the
3198client's address. This only requires configuring the name servers, not
3199all the clients.
3200
3201The ``sortlist`` statement (see below) takes an ``address_match_list`` and
3202interprets it in a special way. Each top-level statement in the ``sortlist``
3203must itself be an explicit ``address_match_list`` with one or two elements. The
3204first element (which may be an IP address, an IP prefix, an ACL name, or a nested
3205``address_match_list``) of each top-level list is checked against the source
3206address of the query until a match is found. When the addresses in the first
3207element overlap, the first rule to match is selected.
3208
3209Once the source address of the query has been matched, if the top-level
3210statement contains only one element, the actual primitive element that
3211matched the source address is used to select the address in the response
3212to move to the beginning of the response. If the statement is a list of
3213two elements, then the second element is interpreted as a topology
3214preference list. Each top-level element is assigned a distance, and the
3215address in the response with the minimum distance is moved to the
3216beginning of the response.
3217
3218In the following example, any queries received from any of the addresses
3219of the host itself get responses preferring addresses on any of the
3220locally connected networks. Next most preferred are addresses on the
3221192.168.1/24 network, and after that either the 192.168.2/24 or
3222192.168.3/24 network, with no preference shown between these two
3223networks. Queries received from a host on the 192.168.1/24 network
3224prefer other addresses on that network to the 192.168.2/24 and
3225192.168.3/24 networks. Queries received from a host on the 192.168.4/24
3226or the 192.168.5/24 network only prefer other addresses on their
3227directly connected networks.
3228
3229::
3230
3231   sortlist {
3232       // IF the local host
3233       // THEN first fit on the following nets
3234       { localhost;
3235       { localnets;
3236           192.168.1/24;
3237           { 192.168.2/24; 192.168.3/24; }; }; };
3238       // IF on class C 192.168.1 THEN use .1, or .2 or .3
3239       { 192.168.1/24;
3240       { 192.168.1/24;
3241           { 192.168.2/24; 192.168.3/24; }; }; };
3242       // IF on class C 192.168.2 THEN use .2, or .1 or .3
3243       { 192.168.2/24;
3244       { 192.168.2/24;
3245           { 192.168.1/24; 192.168.3/24; }; }; };
3246       // IF on class C 192.168.3 THEN use .3, or .1 or .2
3247       { 192.168.3/24;
3248       { 192.168.3/24;
3249           { 192.168.1/24; 192.168.2/24; }; }; };
3250       // IF .4 or .5 THEN prefer that net
3251       { { 192.168.4/24; 192.168.5/24; };
3252       };
3253   };
3254
3255The following example illlustrates reasonable behavior for the local host
3256and hosts on directly connected networks. Responses sent to queries from the
3257local host favor any of the directly connected networks. Responses
3258sent to queries from any other hosts on a directly connected network
3259prefer addresses on that same network. Responses to other queries
3260are not sorted.
3261
3262::
3263
3264   sortlist {
3265          { localhost; localnets; };
3266          { localnets; };
3267   };
3268
3269.. _rrset_ordering:
3270
3271RRset Ordering
3272^^^^^^^^^^^^^^
3273
3274.. note::
3275
3276    While alternating the order of records in a DNS response between
3277    subsequent queries is a known load distribution technique, certain
3278    caveats apply (mostly stemming from caching) which usually make it a
3279    suboptimal choice for load balancing purposes when used on its own.
3280
3281The ``rrset-order`` statement permits configuration of the ordering of
3282the records in a multiple-record response. See also:
3283:ref:`the_sortlist_statement`.
3284
3285Each rule in an ``rrset-order`` statement is defined as follows:
3286
3287::
3288
3289    [class <class_name>] [type <type_name>] [name "<domain_name>"] order <ordering>
3290
3291The default qualifiers for each rule are:
3292
3293  - If no ``class`` is specified, the default is ``ANY``.
3294  - If no ``type`` is specified, the default is ``ANY``.
3295  - If no ``name`` is specified, the default is ``*`` (asterisk).
3296
3297``<domain_name>`` only matches the name itself, not any of its
3298subdomains.  To make a rule match all subdomains of a given name, a
3299wildcard name (``*.<domain_name>``) must be used.  Note that
3300``*.<domain_name>`` does *not* match ``<domain_name>`` itself; to
3301specify RRset ordering for a name and all of its subdomains, two
3302separate rules must be defined: one for ``<domain_name>`` and one for
3303``*.<domain_name>``.
3304
3305The legal values for ``<ordering>`` are:
3306
3307``fixed``
3308    Records are returned in the order they are defined in the zone file.
3309
3310.. note::
3311
3312    The ``fixed`` option is only available if BIND is configured with
3313    ``--enable-fixed-rrset`` at compile time.
3314
3315``random``
3316    Records are returned in a random order.
3317
3318``cyclic``
3319    Records are returned in a cyclic round-robin order, rotating by one
3320    record per query.
3321
3322``none``
3323    Records are returned in the order they were retrieved from the
3324    database. This order is indeterminate, but remains consistent as
3325    long as the database is not modified.
3326
3327The default RRset order used depends on whether any ``rrset-order``
3328statements are present in the configuration file used by ``named``:
3329
3330  - If no ``rrset-order`` statement is present in the configuration
3331    file, the implicit default is to return all records in ``random``
3332    order.
3333
3334  - If any ``rrset-order`` statements are present in the configuration
3335    file, but no ordering rule specified in these statements matches a
3336    given RRset, the default order for that RRset is ``none``.
3337
3338Note that if multiple ``rrset-order`` statements are present in the
3339configuration file (at both the ``options`` and ``view`` levels), they
3340are *not* combined; instead, the more-specific one (``view``) replaces
3341the less-specific one (``options``).
3342
3343If multiple rules within a single ``rrset-order`` statement match a
3344given RRset, the first matching rule is applied.
3345
3346Example:
3347
3348::
3349
3350    rrset-order {
3351        type A name "foo.isc.org" order random;
3352        type AAAA name "foo.isc.org" order cyclic;
3353        name "bar.isc.org" order fixed;
3354        name "*.bar.isc.org" order random;
3355        name "*.baz.isc.org" order cyclic;
3356    };
3357
3358With the above configuration, the following RRset ordering is used:
3359
3360===================    ========    ===========
3361QNAME                  QTYPE       RRset Order
3362===================    ========    ===========
3363``foo.isc.org``        ``A``       ``random``
3364``foo.isc.org``        ``AAAA``    ``cyclic``
3365``foo.isc.org``        ``TXT``     ``none``
3366``sub.foo.isc.org``    all         ``none``
3367``bar.isc.org``        all         ``fixed``
3368``sub.bar.isc.org``    all         ``random``
3369``baz.isc.org``        all         ``none``
3370``sub.baz.isc.org``    all         ``cyclic``
3371===================    ========    ===========
3372
3373.. _tuning:
3374
3375Tuning
3376^^^^^^
3377
3378``lame-ttl``
3379   This is always set to 0. More information is available in the
3380   `security advisory for CVE-2021-25219
3381   <https://kb.isc.org/docs/cve-2021-25219>`_.
3382
3383``servfail-ttl``
3384   This sets the number of seconds to cache a SERVFAIL response due to DNSSEC
3385   validation failure or other general server failure. If set to ``0``,
3386   SERVFAIL caching is disabled. The SERVFAIL cache is not consulted if
3387   a query has the CD (Checking Disabled) bit set; this allows a query
3388   that failed due to DNSSEC validation to be retried without waiting
3389   for the SERVFAIL TTL to expire.
3390
3391   The maximum value is ``30`` seconds; any higher value is
3392   silently reduced. The default is ``1`` second.
3393
3394``min-ncache-ttl``
3395   To reduce network traffic and increase performance, the server stores
3396   negative answers. ``min-ncache-ttl`` is used to set a minimum
3397   retention time for these answers in the server, in seconds. For
3398   convenience, TTL-style time-unit suffixes may be used to specify the
3399   value. It also accepts ISO 8601 duration formats.
3400
3401   The default ``min-ncache-ttl`` is ``0`` seconds. ``min-ncache-ttl`` cannot
3402   exceed 90 seconds and is truncated to 90 seconds if set to a greater
3403   value.
3404
3405``min-cache-ttl``
3406   This sets the minimum time for which the server caches ordinary (positive)
3407   answers, in seconds. For convenience, TTL-style time-unit suffixes may be used
3408   to specify the value. It also accepts ISO 8601 duration formats.
3409
3410   The default ``min-cache-ttl`` is ``0`` seconds. ``min-cache-ttl`` cannot
3411   exceed 90 seconds and is truncated to 90 seconds if set to a greater
3412   value.
3413
3414``max-ncache-ttl``
3415   To reduce network traffic and increase performance, the server stores
3416   negative answers. ``max-ncache-ttl`` is used to set a maximum retention time
3417   for these answers in the server, in seconds. For convenience, TTL-style
3418   time-unit suffixes may be used to specify the value.  It also accepts ISO 8601
3419   duration formats.
3420
3421   The default ``max-ncache-ttl`` is 10800 seconds (3 hours). ``max-ncache-ttl``
3422   cannot exceed 7 days and is silently truncated to 7 days if set to a
3423   greater value.
3424
3425``max-cache-ttl``
3426   This sets the maximum time for which the server caches ordinary (positive)
3427   answers, in seconds. For convenience, TTL-style time-unit suffixes may be used
3428   to specify the value. It also accepts ISO 8601 duration formats.
3429
3430   The default ``max-cache-ttl`` is 604800 (one week). A value of zero may cause
3431   all queries to return SERVFAIL, because of lost caches of intermediate RRsets
3432   (such as NS and glue AAAA/A records) in the resolution process.
3433
3434``max-stale-ttl``
3435   If retaining stale RRsets in cache is enabled, and returning of stale cached
3436   answers is also enabled, ``max-stale-ttl`` sets the maximum time for which
3437   the server retains records past their normal expiry to return them as stale
3438   records, when the servers for those records are not reachable. The default
3439   is 1 day. The minimum allowed is 1 second; a value of 0 is updated silently
3440   to 1 second.
3441
3442   For stale answers to be returned, the retaining of them in cache must be
3443   enabled via the configuration option ``stale-cache-enable``, and returning
3444   cached answers must be enabled, either in the configuration file using the
3445   ``stale-answer-enable`` option or by calling ``rndc serve-stale on``.
3446
3447   When ``stale-cache-enable`` is set to ``no``, setting the ``max-stale-ttl``
3448   has no effect, the value of ``max-cache-ttl`` will be ``0`` in such case.
3449
3450``resolver-nonbackoff-tries``
3451   This specifies how many retries occur before exponential backoff kicks in. The
3452   default is ``3``.
3453
3454``resolver-retry-interval``
3455   This sets the base retry interval in milliseconds. The default is ``800``.
3456
3457``sig-validity-interval``
3458   this specifies the upper bound of the number of days that RRSIGs
3459   generated by ``named`` are valid; the default is ``30`` days,
3460   with a maximum of 3660 days (10 years). The optional second value
3461   specifies the minimum bound on those RRSIGs and also determines
3462   how long before expiry ``named`` starts regenerating those RRSIGs.
3463   The default value for the lower bound is 1/4 of the upper bound;
3464   it is expressed in days if the upper bound is greater than 7,
3465   and hours if it is less than or equal to 7 days.
3466
3467   When new RRSIGs are generated, the length of time is randomly
3468   chosen between these two limits, to spread out the re-signing
3469   load. When RRSIGs are re-generated, the upper bound is used, with
3470   a small amount of jitter added. New RRSIGs are generated by a
3471   number of processes, including the processing of UPDATE requests
3472   (ref:`dynamic_update`), the addition and removal of records via
3473   in-line signing, and the initial signing of a zone.
3474
3475   The signature inception time is unconditionally set to one hour
3476   before the current time, to allow for a limited amount of clock skew.
3477
3478   The ``sig-validity-interval`` can be overridden for DNSKEY records by
3479   setting ``dnskey-sig-validity``.
3480
3481   The ``sig-validity-interval`` should be at least several multiples
3482   of the SOA expire interval, to allow for reasonable interaction
3483   between the various timer and expiry dates.
3484
3485``dnskey-sig-validity``
3486   This specifies the number of days into the future when DNSSEC signatures
3487   that are automatically generated for DNSKEY RRsets as a result of
3488   dynamic updates (:ref:`dynamic_update`) will expire.
3489   If set to a non-zero value, this overrides the value set by
3490   ``sig-validity-interval``. The default is zero, meaning
3491   ``sig-validity-interval`` is used. The maximum value is 3660 days (10
3492   years), and higher values are rejected.
3493
3494``sig-signing-nodes``
3495   This specifies the maximum number of nodes to be examined in each quantum,
3496   when signing a zone with a new DNSKEY. The default is ``100``.
3497
3498``sig-signing-signatures``
3499   This specifies a threshold number of signatures that terminates
3500   processing a quantum, when signing a zone with a new DNSKEY. The
3501   default is ``10``.
3502
3503``sig-signing-type``
3504   This specifies a private RDATA type to be used when generating signing-state
3505   records. The default is ``65534``.
3506
3507   This parameter may be removed in a future version,
3508   once there is a standard type.
3509
3510   Signing-state records are used internally by ``named`` to track
3511   the current state of a zone-signing process, i.e., whether it is
3512   still active or has been completed. The records can be inspected
3513   using the command ``rndc signing -list zone``. Once ``named`` has
3514   finished signing a zone with a particular key, the signing-state
3515   record associated with that key can be removed from the zone by
3516   running ``rndc signing -clear keyid/algorithm zone``. To clear all of
3517   the completed signing-state records for a zone, use
3518   ``rndc signing -clear all zone``.
3519
3520``min-refresh-time``; ``max-refresh-time``; ``min-retry-time``; ``max-retry-time``
3521   These options control the server's behavior on refreshing a zone
3522   (querying for SOA changes) or retrying failed transfers. Usually the
3523   SOA values for the zone are used, up to a hard-coded maximum expiry
3524   of 24 weeks. However, these values are set by the primary, giving
3525   secondary server administrators little control over their contents.
3526
3527   These options allow the administrator to set a minimum and maximum
3528   refresh and retry time in seconds per-zone, per-view, or globally.
3529   These options are valid for secondary and stub zones, and clamp the SOA
3530   refresh and retry times to the specified values.
3531
3532   The following defaults apply: ``min-refresh-time`` 300 seconds,
3533   ``max-refresh-time`` 2419200 seconds (4 weeks), ``min-retry-time``
3534   500 seconds, and ``max-retry-time`` 1209600 seconds (2 weeks).
3535
3536``edns-udp-size``
3537   This sets the maximum advertised EDNS UDP buffer size, in bytes, to control
3538   the size of packets received from authoritative servers in response
3539   to recursive queries. Valid values are 512 to 4096; values outside
3540   this range are silently adjusted to the nearest value within it.
3541   The default value is 1232.
3542
3543   The usual reason for setting ``edns-udp-size`` to a non-default value
3544   is to get UDP answers to pass through broken firewalls that block
3545   fragmented packets and/or block UDP DNS packets that are greater than
3546   512 bytes.
3547
3548   When ``named`` first queries a remote server, it advertises a UDP
3549   buffer size of 512, as this has the greatest chance of success on the
3550   first try.
3551
3552   If the initial query is successful with EDNS advertising a buffer size of
3553   512, then ``named`` will advertise progressively larger buffer sizes on
3554   successive queries, until responses begin timing out or ``edns-udp-size`` is
3555   reached.
3556
3557   The default buffer sizes used by ``named`` are 512, 1232, 1432, and
3558   4096, but never exceeding ``edns-udp-size``. (The values 1232 and
3559   1432 are chosen to allow for an IPv4-/IPv6-encapsulated UDP message
3560   to be sent without fragmentation at the minimum MTU sizes for
3561   Ethernet and IPv6 networks.)
3562
3563   The ``named`` now sets the DON'T FRAGMENT flag on outgoing UDP packets.
3564   According to the measurements done by multiple parties this should not be
3565   causing any operational problems as most of the Internet "core" is able to
3566   cope with IP message sizes between 1400-1500 bytes, the 1232 size was picked
3567   as a conservative minimal number that could be changed by the DNS operator to
3568   a estimated path MTU minus the estimated header space. In practice, the
3569   smallest MTU witnessed in the operational DNS community is 1500 octets, the
3570   Ethernet maximum payload size, so a a useful default for maximum DNS/UDP
3571   payload size on **reliable** networks would be 1432.
3572
3573   Any server-specific ``edns-udp-size`` setting has precedence over all
3574   the above rules.
3575
3576``max-udp-size``
3577   This sets the maximum EDNS UDP message size that ``named`` sends, in bytes.
3578   Valid values are 512 to 4096; values outside this range are
3579   silently adjusted to the nearest value within it. The default value
3580   is 1232.
3581
3582   This value applies to responses sent by a server; to set the
3583   advertised buffer size in queries, see ``edns-udp-size``.
3584
3585   The usual reason for setting ``max-udp-size`` to a non-default value
3586   is to allow UDP answers to pass through broken firewalls that block
3587   fragmented packets and/or block UDP packets that are greater than 512
3588   bytes. This is independent of the advertised receive buffer
3589   (``edns-udp-size``).
3590
3591   Setting this to a low value encourages additional TCP traffic to
3592   the name server.
3593
3594``masterfile-format``
3595   This specifies the file format of zone files (see :ref:`zonefile_format`
3596   for details).  The default value is ``text``, which is the standard
3597   textual representation, except for secondary zones, in which the default
3598   value is ``raw``. Files in formats other than ``text`` are typically
3599   expected to be generated by the ``named-compilezone`` tool, or dumped by
3600   ``named``.
3601
3602   Note that when a zone file in a format other than ``text`` is loaded,
3603   ``named`` may omit some of the checks which are performed for a file in
3604   ``text`` format. For example, ``check-names`` only applies when loading
3605   zones in ``text`` format, and ``max-zone-ttl`` only applies to ``text``
3606   and ``raw``.  Zone files in binary formats should be generated with the
3607   same check level as that specified in the ``named`` configuration file.
3608
3609   ``map`` format files are loaded directly into memory via memory mapping,
3610   with only minimal validity checking. Because they are not guaranteed to
3611   be compatible from one version of BIND 9 to another, and are not
3612   compatible from one system architecture to another, they should be used
3613   with caution. See :ref:`zonefile_format` for further discussion.
3614
3615   When configured in ``options``, this statement sets the
3616   ``masterfile-format`` for all zones, but it can be overridden on a
3617   per-zone or per-view basis by including a ``masterfile-format``
3618   statement within the ``zone`` or ``view`` block in the configuration
3619   file.
3620
3621``masterfile-style``
3622   This specifies the formatting of zone files during dump, when the
3623   ``masterfile-format`` is ``text``. This option is ignored with any
3624   other ``masterfile-format``.
3625
3626   When set to ``relative``, records are printed in a multi-line format,
3627   with owner names expressed relative to a shared origin. When set to
3628   ``full``, records are printed in a single-line format with absolute
3629   owner names. The ``full`` format is most suitable when a zone file
3630   needs to be processed automatically by a script. The ``relative``
3631   format is more human-readable, and is thus suitable when a zone is to
3632   be edited by hand. The default is ``relative``.
3633
3634``max-recursion-depth``
3635   This sets the maximum number of levels of recursion that are permitted at
3636   any one time while servicing a recursive query. Resolving a name may
3637   require looking up a name server address, which in turn requires
3638   resolving another name, etc.; if the number of recursions exceeds
3639   this value, the recursive query is terminated and returns SERVFAIL.
3640   The default is 7.
3641
3642``max-recursion-queries``
3643   This sets the maximum number of iterative queries that may be sent while
3644   servicing a recursive query. If more queries are sent, the recursive
3645   query is terminated and returns SERVFAIL. The default is 100.
3646
3647``notify-delay``
3648   This sets the delay, in seconds, between sending sets of NOTIFY messages
3649   for a zone. Whenever a NOTIFY message is sent for a zone, a timer will
3650   be set for this duration. If the zone is updated again before the timer
3651   expires, the NOTIFY for that update will be postponed. The default is 5
3652   seconds.
3653
3654   The overall rate at which NOTIFY messages are sent for all zones is
3655   controlled by ``notify-rate``.
3656
3657``max-rsa-exponent-size``
3658   This sets the maximum RSA exponent size, in bits, that is accepted when
3659   validating. Valid values are 35 to 4096 bits. The default, zero, is
3660   also accepted and is equivalent to 4096.
3661
3662``prefetch``
3663   When a query is received for cached data which is to expire shortly,
3664   ``named`` can refresh the data from the authoritative server
3665   immediately, ensuring that the cache always has an answer available.
3666
3667   ``prefetch`` specifies the "trigger" TTL value at which prefetch
3668   of the current query takes place; when a cache record with a
3669   lower or equal TTL value is encountered during query processing, it is
3670   refreshed. Valid trigger TTL values are 1 to 10 seconds. Values
3671   larger than 10 seconds are silently reduced to 10. Setting a
3672   trigger TTL to zero causes prefetch to be disabled. The default
3673   trigger TTL is ``2``.
3674
3675   An optional second argument specifies the "eligibility" TTL: the
3676   smallest *original* TTL value that is accepted for a record to
3677   be eligible for prefetching. The eligibility TTL must be at least six
3678   seconds longer than the trigger TTL; if not, ``named``
3679   silently adjusts it upward. The default eligibility TTL is ``9``.
3680
3681``v6-bias``
3682   When determining the next name server to try, this indicates by how many
3683   milliseconds to prefer IPv6 name servers. The default is ``50``
3684   milliseconds.
3685
3686.. _builtin:
3687
3688Built-in Server Information Zones
3689^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3690
3691The server provides some helpful diagnostic information through a number
3692of built-in zones under the pseudo-top-level-domain ``bind`` in the
3693``CHAOS`` class. These zones are part of a built-in view
3694(see :ref:`view_statement_grammar`) of class ``CHAOS``, which is
3695separate from the default view of class ``IN``. Most global
3696configuration options (``allow-query``, etc.) apply to this view,
3697but some are locally overridden: ``notify``, ``recursion``, and
3698``allow-new-zones`` are always set to ``no``, and ``rate-limit`` is set
3699to allow three responses per second.
3700
3701To disable these zones, use the options below or hide the
3702built-in ``CHAOS`` view by defining an explicit view of class ``CHAOS``
3703that matches all clients.
3704
3705``version``
3706   This is the version the server should report via a query of the name
3707   ``version.bind`` with type ``TXT`` and class ``CHAOS``. The default is
3708   the real version number of this server. Specifying ``version none``
3709   disables processing of the queries.
3710
3711   Setting ``version`` to any value (including ``none``) also disables
3712   queries for ``authors.bind TXT CH``.
3713
3714``hostname``
3715   This is the hostname the server should report via a query of the name
3716   ``hostname.bind`` with type ``TXT`` and class ``CHAOS``. This defaults
3717   to the hostname of the machine hosting the name server, as found by
3718   the ``gethostname()`` function. The primary purpose of such queries is to
3719   identify which of a group of anycast servers is actually answering
3720   the queries. Specifying ``hostname none;`` disables processing of
3721   the queries.
3722
3723``server-id``
3724   This is the ID the server should report when receiving a Name Server
3725   Identifier (NSID) query, or a query of the name ``ID.SERVER`` with
3726   type ``TXT`` and class ``CHAOS``. The primary purpose of such queries is
3727   to identify which of a group of anycast servers is actually answering
3728   the queries. Specifying ``server-id none;`` disables processing of
3729   the queries. Specifying ``server-id hostname;`` causes ``named``
3730   to use the hostname as found by the ``gethostname()`` function. The
3731   default ``server-id`` is ``none``.
3732
3733.. _empty:
3734
3735Built-in Empty Zones
3736^^^^^^^^^^^^^^^^^^^^
3737
3738The ``named`` server has some built-in empty zones, for SOA and NS records
3739only. These are for zones that should normally be answered locally and for
3740which queries should not be sent to the Internet's root servers. The
3741official servers that cover these namespaces return NXDOMAIN responses
3742to these queries. In particular, these cover the reverse namespaces for
3743addresses from :rfc:`1918`, :rfc:`4193`, :rfc:`5737`, and :rfc:`6598`. They also
3744include the reverse namespace for the IPv6 local address (locally assigned),
3745IPv6 link local addresses, the IPv6 loopback address, and the IPv6
3746unknown address.
3747
3748The server attempts to determine if a built-in zone already exists
3749or is active (covered by a forward-only forwarding declaration) and does
3750not create an empty zone if either is true.
3751
3752The current list of empty zones is:
3753
3754-  10.IN-ADDR.ARPA
3755-  16.172.IN-ADDR.ARPA
3756-  17.172.IN-ADDR.ARPA
3757-  18.172.IN-ADDR.ARPA
3758-  19.172.IN-ADDR.ARPA
3759-  20.172.IN-ADDR.ARPA
3760-  21.172.IN-ADDR.ARPA
3761-  22.172.IN-ADDR.ARPA
3762-  23.172.IN-ADDR.ARPA
3763-  24.172.IN-ADDR.ARPA
3764-  25.172.IN-ADDR.ARPA
3765-  26.172.IN-ADDR.ARPA
3766-  27.172.IN-ADDR.ARPA
3767-  28.172.IN-ADDR.ARPA
3768-  29.172.IN-ADDR.ARPA
3769-  30.172.IN-ADDR.ARPA
3770-  31.172.IN-ADDR.ARPA
3771-  168.192.IN-ADDR.ARPA
3772-  64.100.IN-ADDR.ARPA
3773-  65.100.IN-ADDR.ARPA
3774-  66.100.IN-ADDR.ARPA
3775-  67.100.IN-ADDR.ARPA
3776-  68.100.IN-ADDR.ARPA
3777-  69.100.IN-ADDR.ARPA
3778-  70.100.IN-ADDR.ARPA
3779-  71.100.IN-ADDR.ARPA
3780-  72.100.IN-ADDR.ARPA
3781-  73.100.IN-ADDR.ARPA
3782-  74.100.IN-ADDR.ARPA
3783-  75.100.IN-ADDR.ARPA
3784-  76.100.IN-ADDR.ARPA
3785-  77.100.IN-ADDR.ARPA
3786-  78.100.IN-ADDR.ARPA
3787-  79.100.IN-ADDR.ARPA
3788-  80.100.IN-ADDR.ARPA
3789-  81.100.IN-ADDR.ARPA
3790-  82.100.IN-ADDR.ARPA
3791-  83.100.IN-ADDR.ARPA
3792-  84.100.IN-ADDR.ARPA
3793-  85.100.IN-ADDR.ARPA
3794-  86.100.IN-ADDR.ARPA
3795-  87.100.IN-ADDR.ARPA
3796-  88.100.IN-ADDR.ARPA
3797-  89.100.IN-ADDR.ARPA
3798-  90.100.IN-ADDR.ARPA
3799-  91.100.IN-ADDR.ARPA
3800-  92.100.IN-ADDR.ARPA
3801-  93.100.IN-ADDR.ARPA
3802-  94.100.IN-ADDR.ARPA
3803-  95.100.IN-ADDR.ARPA
3804-  96.100.IN-ADDR.ARPA
3805-  97.100.IN-ADDR.ARPA
3806-  98.100.IN-ADDR.ARPA
3807-  99.100.IN-ADDR.ARPA
3808-  100.100.IN-ADDR.ARPA
3809-  101.100.IN-ADDR.ARPA
3810-  102.100.IN-ADDR.ARPA
3811-  103.100.IN-ADDR.ARPA
3812-  104.100.IN-ADDR.ARPA
3813-  105.100.IN-ADDR.ARPA
3814-  106.100.IN-ADDR.ARPA
3815-  107.100.IN-ADDR.ARPA
3816-  108.100.IN-ADDR.ARPA
3817-  109.100.IN-ADDR.ARPA
3818-  110.100.IN-ADDR.ARPA
3819-  111.100.IN-ADDR.ARPA
3820-  112.100.IN-ADDR.ARPA
3821-  113.100.IN-ADDR.ARPA
3822-  114.100.IN-ADDR.ARPA
3823-  115.100.IN-ADDR.ARPA
3824-  116.100.IN-ADDR.ARPA
3825-  117.100.IN-ADDR.ARPA
3826-  118.100.IN-ADDR.ARPA
3827-  119.100.IN-ADDR.ARPA
3828-  120.100.IN-ADDR.ARPA
3829-  121.100.IN-ADDR.ARPA
3830-  122.100.IN-ADDR.ARPA
3831-  123.100.IN-ADDR.ARPA
3832-  124.100.IN-ADDR.ARPA
3833-  125.100.IN-ADDR.ARPA
3834-  126.100.IN-ADDR.ARPA
3835-  127.100.IN-ADDR.ARPA
3836-  0.IN-ADDR.ARPA
3837-  127.IN-ADDR.ARPA
3838-  254.169.IN-ADDR.ARPA
3839-  2.0.192.IN-ADDR.ARPA
3840-  100.51.198.IN-ADDR.ARPA
3841-  113.0.203.IN-ADDR.ARPA
3842-  255.255.255.255.IN-ADDR.ARPA
3843-  0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.IP6.ARPA
3844-  1.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.IP6.ARPA
3845-  8.B.D.0.1.0.0.2.IP6.ARPA
3846-  D.F.IP6.ARPA
3847-  8.E.F.IP6.ARPA
3848-  9.E.F.IP6.ARPA
3849-  A.E.F.IP6.ARPA
3850-  B.E.F.IP6.ARPA
3851-  EMPTY.AS112.ARPA
3852-  HOME.ARPA
3853
3854Empty zones can be set at the view level and only apply to views of
3855class IN. Disabled empty zones are only inherited from options if there
3856are no disabled empty zones specified at the view level. To override the
3857options list of disabled zones, disable the root zone at the
3858view level. For example:
3859
3860::
3861
3862           disable-empty-zone ".";
3863
3864If using the address ranges covered here,
3865reverse zones covering the addresses should already be in place. In practice this
3866appears to not be the case, with many queries being made to the
3867infrastructure servers for names in these spaces. So many, in fact, that
3868sacrificial servers had to be deployed to channel the query load
3869away from the infrastructure servers.
3870
3871.. note::
3872
3873   The real parent servers for these zones should disable all empty zones
3874   under the parent zone they serve. For the real root servers, this is
3875   all built-in empty zones. This enables them to return referrals
3876   to deeper in the tree.
3877
3878``empty-server``
3879   This specifies the server name that appears in the returned SOA record for
3880   empty zones. If none is specified, the zone's name is used.
3881
3882``empty-contact``
3883   This specifies the contact name that appears in the returned SOA record for
3884   empty zones. If none is specified, "." is used.
3885
3886``empty-zones-enable``
3887   This enables or disables all empty zones. By default, they are enabled.
3888
3889``disable-empty-zone``
3890   This disables individual empty zones. By default, none are disabled. This
3891   option can be specified multiple times.
3892
3893.. _content_filtering:
3894
3895Content Filtering
3896^^^^^^^^^^^^^^^^^
3897
3898BIND 9 provides the ability to filter out responses from external
3899DNS servers containing certain types of data in the answer section.
3900Specifically, it can reject address (A or AAAA) records if the
3901corresponding IPv4 or IPv6 addresses match the given
3902``address_match_list`` of the ``deny-answer-addresses`` option. It can
3903also reject CNAME or DNAME records if the "alias" name (i.e., the CNAME
3904alias or the substituted query name due to DNAME) matches the given
3905``namelist`` of the ``deny-answer-aliases`` option, where "match" means
3906the alias name is a subdomain of one of the ``name_list`` elements. If
3907the optional ``namelist`` is specified with ``except-from``, records
3908whose query name matches the list are accepted regardless of the
3909filter setting. Likewise, if the alias name is a subdomain of the
3910corresponding zone, the ``deny-answer-aliases`` filter does not apply;
3911for example, even if "example.com" is specified for
3912``deny-answer-aliases``,
3913
3914::
3915
3916   www.example.com. CNAME xxx.example.com.
3917
3918returned by an "example.com" server is accepted.
3919
3920In the ``address_match_list`` of the ``deny-answer-addresses`` option,
3921only ``ip_addr`` and ``ip_prefix`` are meaningful; any ``key_id`` is
3922silently ignored.
3923
3924If a response message is rejected due to the filtering, the entire
3925message is discarded without being cached, and a SERVFAIL error is
3926returned to the client.
3927
3928This filtering is intended to prevent "DNS rebinding attacks," in which
3929an attacker, in response to a query for a domain name the attacker
3930controls, returns an IP address within the user's own network or an alias name
3931within the user's own domain. A naive web browser or script could then serve
3932as an unintended proxy, allowing the attacker to get access to an
3933internal node of the local network that could not be externally accessed
3934otherwise. See the paper available at
3935https://dl.acm.org/doi/10.1145/1315245.1315298 for more details
3936about these attacks.
3937
3938For example, with a domain named "example.net" and an internal
3939network using an IPv4 prefix 192.0.2.0/24, an administrator might specify the
3940following rules:
3941
3942::
3943
3944   deny-answer-addresses { 192.0.2.0/24; } except-from { "example.net"; };
3945   deny-answer-aliases { "example.net"; };
3946
3947If an external attacker let a web browser in the local network look up
3948an IPv4 address of "attacker.example.com", the attacker's DNS server
3949would return a response like this:
3950
3951::
3952
3953   attacker.example.com. A 192.0.2.1
3954
3955in the answer section. Since the rdata of this record (the IPv4 address)
3956matches the specified prefix 192.0.2.0/24, this response would be
3957ignored.
3958
3959On the other hand, if the browser looked up a legitimate internal web
3960server "www.example.net" and the following response were returned to the
3961BIND 9 server:
3962
3963::
3964
3965   www.example.net. A 192.0.2.2
3966
3967it would be accepted, since the owner name "www.example.net" matches the
3968``except-from`` element, "example.net".
3969
3970Note that this is not really an attack on the DNS per se. In fact, there
3971is nothing wrong with having an "external" name mapped to an "internal"
3972IP address or domain name from the DNS point of view; it might actually
3973be provided for a legitimate purpose, such as for debugging. As long as
3974the mapping is provided by the correct owner, it either is not possible or does
3975not make sense to detect whether the intent of the mapping is legitimate
3976within the DNS. The "rebinding" attack must primarily be
3977protected at the application that uses the DNS. For a large site,
3978however, it may be difficult to protect all possible applications at
3979once. This filtering feature is provided only to help such an
3980operational environment; turning it on is generally discouraged
3981unless there is no other choice and the attack is a
3982real threat to applications.
3983
3984Care should be particularly taken if using this option for
3985addresses within 127.0.0.0/8. These addresses are obviously "internal,"
3986but many applications conventionally rely on a DNS mapping from some
3987name to such an address. Filtering out DNS records containing this
3988address spuriously can break such applications.
3989
3990.. _rpz:
3991
3992Response Policy Zone (RPZ) Rewriting
3993^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3994
3995BIND 9 includes a limited mechanism to modify DNS responses for requests
3996analogous to email anti-spam DNS rejection lists. Responses can be changed to
3997deny the existence of domains (NXDOMAIN), deny the existence of IP
3998addresses for domains (NODATA), or contain other IP addresses or data.
3999
4000Response policy zones are named in the ``response-policy`` option for
4001the view, or among the global options if there is no ``response-policy``
4002option for the view. Response policy zones are ordinary DNS zones
4003containing RRsets that can be queried normally if allowed. It is usually
4004best to restrict those queries with something like
4005``allow-query { localhost; };``. Note that zones using
4006``masterfile-format map`` cannot be used as policy zones.
4007
4008A ``response-policy`` option can support multiple policy zones. To
4009maximize performance, a radix tree is used to quickly identify response
4010policy zones containing triggers that match the current query. This
4011imposes an upper limit of 64 on the number of policy zones in a single
4012``response-policy`` option; more than that is a configuration error.
4013
4014Rules encoded in response policy zones are processed after those defined in
4015:ref:`access_control`. All queries from clients which are not permitted access
4016to the resolver are answered with a status code of REFUSED, regardless of
4017configured RPZ rules.
4018
4019Five policy triggers can be encoded in RPZ records.
4020
4021``RPZ-CLIENT-IP``
4022   IP records are triggered by the IP address of the DNS client. Client
4023   IP address triggers are encoded in records that have owner names that
4024   are subdomains of ``rpz-client-ip``, relativized to the policy zone
4025   origin name, and that encode an address or address block. IPv4 addresses
4026   are represented as ``prefixlength.B4.B3.B2.B1.rpz-client-ip``. The
4027   IPv4 prefix length must be between 1 and 32. All four bytes - B4, B3,
4028   B2, and B1 - must be present. B4 is the decimal value of the least
4029   significant byte of the IPv4 address as in IN-ADDR.ARPA.
4030
4031   IPv6 addresses are encoded in a format similar to the standard IPv6
4032   text representation,
4033   ``prefixlength.W8.W7.W6.W5.W4.W3.W2.W1.rpz-client-ip``. Each of
4034   W8,...,W1 is a one- to four-digit hexadecimal number representing 16
4035   bits of the IPv6 address as in the standard text representation of
4036   IPv6 addresses, but reversed as in IP6.ARPA. (Note that this
4037   representation of IPv6 addresses is different from IP6.ARPA, where each
4038   hex digit occupies a label.) All 8 words must be present except when
4039   one set of consecutive zero words is replaced with ``.zz.``, analogous
4040   to double colons (::) in standard IPv6 text encodings. The IPv6
4041   prefix length must be between 1 and 128.
4042
4043``QNAME``
4044   QNAME policy records are triggered by query names of requests and
4045   targets of CNAME records resolved to generate the response. The owner
4046   name of a QNAME policy record is the query name relativized to the
4047   policy zone.
4048
4049``RPZ-IP``
4050   IP triggers are IP addresses in an A or AAAA record in the ANSWER
4051   section of a response. They are encoded like client-IP triggers,
4052   except as subdomains of ``rpz-ip``.
4053
4054``RPZ-NSDNAME``
4055   NSDNAME triggers match names of authoritative servers for the query name, a
4056   parent of the query name, a CNAME for the query name, or a parent of a CNAME.
4057   They are encoded as subdomains of ``rpz-nsdname``, relativized
4058   to the RPZ origin name.  NSIP triggers match IP addresses in A and AAAA
4059   RRsets for domains that can be checked against NSDNAME policy records. The
4060   ``nsdname-enable`` phrase turns NSDNAME triggers off or on for a single
4061   policy zone or for all zones.
4062
4063   If authoritative name servers for the query name are not yet known, ``named``
4064   recursively looks up the authoritative servers for the query name before
4065   applying an RPZ-NSDNAME rule, which can cause a processing delay.
4066
4067``RPZ-NSIP``
4068   NSIP triggers match the IP addresses of authoritative servers. They
4069   are encoded like IP triggers, except as subdomains of ``rpz-nsip``.
4070   NSDNAME and NSIP triggers are checked only for names with at least
4071   ``min-ns-dots`` dots. The default value of ``min-ns-dots`` is 1, to
4072   exclude top-level domains. The ``nsip-enable`` phrase turns NSIP
4073   triggers off or on for a single policy zone or for all zones.
4074
4075   If a name server's IP address is not yet known, ``named``
4076   recursively looks up the IP address before applying an RPZ-NSIP rule,
4077   which can cause a processing delay. To speed up processing at the cost
4078   of precision, the ``nsip-wait-recurse`` option can be used; when set
4079   to ``no``, RPZ-NSIP rules are only applied when a name server's
4080   IP address has already been looked up and cached. If a server's IP
4081   address is not in the cache, the RPZ-NSIP rule is ignored,
4082   but the address is looked up in the background and the rule
4083   is applied to subsequent queries. The default is ``yes``,
4084   meaning RPZ-NSIP rules are always applied, even if an address
4085   needs to be looked up first.
4086
4087The query response is checked against all response policy zones, so two
4088or more policy records can be triggered by a response. Because DNS
4089responses are rewritten according to at most one policy record, a single
4090record encoding an action (other than ``DISABLED`` actions) must be
4091chosen. Triggers, or the records that encode them, are chosen for
4092rewriting in the following order:
4093
40941. Choose the triggered record in the zone that appears first in the
4095   response-policy option.
40962. Prefer CLIENT-IP to QNAME to IP to NSDNAME to NSIP triggers in a
4097   single zone.
40983. Among NSDNAME triggers, prefer the trigger that matches the smallest
4099   name under the DNSSEC ordering.
41004. Among IP or NSIP triggers, prefer the trigger with the longest
4101   prefix.
41025. Among triggers with the same prefix length, prefer the IP or NSIP
4103   trigger that matches the smallest IP address.
4104
4105When the processing of a response is restarted to resolve DNAME or CNAME
4106records and a policy record set has not been triggered, all response
4107policy zones are again consulted for the DNAME or CNAME names and
4108addresses.
4109
4110RPZ record sets are any types of DNS record, except DNAME or DNSSEC, that
4111encode actions or responses to individual queries. Any of the policies
4112can be used with any of the triggers. For example, while the
4113``TCP-only`` policy is commonly used with ``client-IP`` triggers, it can
4114be used with any type of trigger to force the use of TCP for responses
4115with owner names in a zone.
4116
4117``PASSTHRU``
4118   The auto-acceptance policy is specified by a CNAME whose target is
4119   ``rpz-passthru``. It causes the response to not be rewritten and is
4120   most often used to "poke holes" in policies for CIDR blocks.
4121
4122``DROP``
4123   The auto-rejection policy is specified by a CNAME whose target is
4124   ``rpz-drop``. It causes the response to be discarded. Nothing is sent
4125   to the DNS client.
4126
4127``TCP-Only``
4128   The "slip" policy is specified by a CNAME whose target is
4129   ``rpz-tcp-only``. It changes UDP responses to short, truncated DNS
4130   responses that require the DNS client to try again with TCP. It is
4131   used to mitigate distributed DNS reflection attacks.
4132
4133``NXDOMAIN``
4134   The "domain undefined" response is encoded by a CNAME whose target is
4135   the root domain (.).
4136
4137``NODATA``
4138   The empty set of resource records is specified by a CNAME whose target
4139   is the wildcard top-level domain (``*.``). It rewrites the response to
4140   NODATA or ANCOUNT=0.
4141
4142``Local Data``
4143   A set of ordinary DNS records can be used to answer queries. Queries
4144   for record types not in the set are answered with NODATA.
4145
4146   A special form of local data is a CNAME whose target is a wildcard
4147   such as \*.example.com. It is used as if an ordinary CNAME after
4148   the asterisk (\*) has been replaced with the query name.
4149   This special form is useful for query logging in the walled garden's
4150   authoritative DNS server.
4151
4152All of the actions specified in all of the individual records in a
4153policy zone can be overridden with a ``policy`` clause in the
4154``response-policy`` option. An organization using a policy zone provided
4155by another organization might use this mechanism to redirect domains to
4156its own walled garden.
4157
4158``GIVEN``
4159   The placeholder policy says "do not override but perform the action
4160   specified in the zone."
4161
4162``DISABLED``
4163   The testing override policy causes policy zone records to do nothing
4164   but log what they would have done if the policy zone were not
4165   disabled. The response to the DNS query is written (or not)
4166   according to any triggered policy records that are not disabled.
4167   Disabled policy zones should appear first, because they are often
4168   not logged if a higher-precedence trigger is found first.
4169
4170``PASSTHRU``; ``DROP``; ``TCP-Only``; ``NXDOMAIN``; ``NODATA``
4171   These settings each override the corresponding per-record policy.
4172
4173``CNAME domain``
4174   This causes all RPZ policy records to act as if they were "cname domain"
4175   records.
4176
4177By default, the actions encoded in a response policy zone are applied
4178only to queries that ask for recursion (RD=1). That default can be
4179changed for a single policy zone, or for all response policy zones in a view,
4180with a ``recursive-only no`` clause. This feature is useful for serving
4181the same zone files both inside and outside an :rfc:`1918` cloud and using
4182RPZ to delete answers that would otherwise contain :rfc:`1918` values on
4183the externally visible name server or view.
4184
4185Also by default, RPZ actions are applied only to DNS requests that
4186either do not request DNSSEC metadata (DO=0) or when no DNSSEC records
4187are available for the requested name in the original zone (not the response
4188policy zone). This default can be changed for all response policy zones
4189in a view with a ``break-dnssec yes`` clause. In that case, RPZ actions
4190are applied regardless of DNSSEC. The name of the clause option reflects
4191the fact that results rewritten by RPZ actions cannot verify.
4192
4193No DNS records are needed for a QNAME or Client-IP trigger; the name or
4194IP address itself is sufficient, so in principle the query name need not
4195be recursively resolved. However, not resolving the requested name can
4196leak the fact that response policy rewriting is in use, and that the name
4197is listed in a policy zone, to operators of servers for listed names. To
4198prevent that information leak, by default any recursion needed for a
4199request is done before any policy triggers are considered. Because
4200listed domains often have slow authoritative servers, this behavior can
4201cost significant time. The ``qname-wait-recurse no`` option overrides
4202the default and enables that behavior when recursion cannot change a
4203non-error response. The option does not affect QNAME or client-IP
4204triggers in policy zones listed after other zones containing IP, NSIP,
4205and NSDNAME triggers, because those may depend on the A, AAAA, and NS
4206records that would be found during recursive resolution. It also does
4207not affect DNSSEC requests (DO=1) unless ``break-dnssec yes`` is in use,
4208because the response would depend on whether RRSIG records were
4209found during resolution. Using this option can cause error responses
4210such as SERVFAIL to appear to be rewritten, since no recursion is being
4211done to discover problems at the authoritative server.
4212
4213The ``dnsrps-enable yes`` option turns on the DNS Response Policy Service
4214(DNSRPS) interface, if it has been compiled in ``named`` using
4215``configure --enable-dnsrps``.
4216
4217The ``dnsrps-options`` block provides additional RPZ configuration
4218settings, which are passed through to the DNSRPS provider library.
4219Multiple DNSRPS settings in an ``dnsrps-options`` string should be
4220separated with semi-colons (;). The DNSRPS provider, librpz, is passed a
4221configuration string consisting of the ``dnsrps-options`` text,
4222concatenated with settings derived from the ``response-policy``
4223statement.
4224
4225Note: the ``dnsrps-options`` text should only include configuration
4226settings that are specific to the DNSRPS provider. For example, the
4227DNSRPS provider from Farsight Security takes options such as
4228``dnsrpzd-conf``, ``dnsrpzd-sock``, and ``dnzrpzd-args`` (for details of
4229these options, see the ``librpz`` documentation). Other RPZ
4230configuration settings could be included in ``dnsrps-options`` as well,
4231but if ``named`` were switched back to traditional RPZ by setting
4232``dnsrps-enable`` to "no", those options would be ignored.
4233
4234The TTL of a record modified by RPZ policies is set from the TTL of the
4235relevant record in the policy zone. It is then limited to a maximum value.
4236The ``max-policy-ttl`` clause changes the maximum number of seconds from its
4237default of 5. For convenience, TTL-style time-unit suffixes may be used
4238to specify the value. It also accepts ISO 8601 duration formats.
4239
4240For example, an administrator might use this option statement:
4241
4242::
4243
4244       response-policy { zone "badlist"; };
4245
4246and this zone statement:
4247
4248::
4249
4250       zone "badlist" {type primary; file "primary/badlist"; allow-query {none;}; };
4251
4252with this zone file:
4253
4254::
4255
4256   $TTL 1H
4257   @                       SOA LOCALHOST. named-mgr.example.com (1 1h 15m 30d 2h)
4258               NS  LOCALHOST.
4259
4260   ; QNAME policy records.  There are no periods (.) after the owner names.
4261   nxdomain.domain.com     CNAME   .               ; NXDOMAIN policy
4262   *.nxdomain.domain.com   CNAME   .               ; NXDOMAIN policy
4263   nodata.domain.com       CNAME   *.              ; NODATA policy
4264   *.nodata.domain.com     CNAME   *.              ; NODATA policy
4265   bad.domain.com          A       10.0.0.1        ; redirect to a walled garden
4266               AAAA    2001:2::1
4267   bzone.domain.com        CNAME   garden.example.com.
4268
4269   ; do not rewrite (PASSTHRU) OK.DOMAIN.COM
4270   ok.domain.com           CNAME   rpz-passthru.
4271
4272   ; redirect x.bzone.domain.com to x.bzone.domain.com.garden.example.com
4273   *.bzone.domain.com      CNAME   *.garden.example.com.
4274
4275   ; IP policy records that rewrite all responses containing A records in 127/8
4276   ;       except 127.0.0.1
4277   8.0.0.0.127.rpz-ip      CNAME   .
4278   32.1.0.0.127.rpz-ip     CNAME   rpz-passthru.
4279
4280   ; NSDNAME and NSIP policy records
4281   ns.domain.com.rpz-nsdname   CNAME   .
4282   48.zz.2.2001.rpz-nsip       CNAME   .
4283
4284   ; auto-reject and auto-accept some DNS clients
4285   112.zz.2001.rpz-client-ip    CNAME   rpz-drop.
4286   8.0.0.0.127.rpz-client-ip    CNAME   rpz-drop.
4287
4288   ; force some DNS clients and responses in the example.com zone to TCP
4289   16.0.0.1.10.rpz-client-ip   CNAME   rpz-tcp-only.
4290   example.com                 CNAME   rpz-tcp-only.
4291   *.example.com               CNAME   rpz-tcp-only.
4292
4293RPZ can affect server performance. Each configured response policy zone
4294requires the server to perform one to four additional database lookups
4295before a query can be answered. For example, a DNS server with four
4296policy zones, each with all four kinds of response triggers (QNAME, IP,
4297NSIP, and NSDNAME), requires a total of 17 times as many database lookups
4298as a similar DNS server with no response policy zones. A BIND 9 server
4299with adequate memory and one response policy zone with QNAME and IP
4300triggers might achieve a maximum queries-per-second (QPS) rate about 20%
4301lower. A server with four response policy zones with QNAME and IP
4302triggers might have a maximum QPS rate about 50% lower.
4303
4304Responses rewritten by RPZ are counted in the ``RPZRewrites``
4305statistics.
4306
4307The ``log`` clause can be used to optionally turn off rewrite logging
4308for a particular response policy zone. By default, all rewrites are
4309logged.
4310
4311The ``add-soa`` option controls whether the RPZ's SOA record is added to
4312the section for traceback of changes from this zone.
4313This can be set at the individual policy zone level or at the
4314response-policy level. The default is ``yes``.
4315
4316Updates to RPZ zones are processed asynchronously; if there is more than
4317one update pending they are bundled together. If an update to a RPZ zone
4318(for example, via IXFR) happens less than ``min-update-interval``
4319seconds after the most recent update, the changes are not
4320carried out until this interval has elapsed. The default is ``60``
4321seconds. For convenience, TTL-style time-unit suffixes may be used to
4322specify the value. It also accepts ISO 8601 duration formats.
4323
4324.. _rrl:
4325
4326Response Rate Limiting
4327^^^^^^^^^^^^^^^^^^^^^^
4328
4329Excessive, almost-identical UDP *responses* can be controlled by
4330configuring a ``rate-limit`` clause in an ``options`` or ``view``
4331statement. This mechanism keeps authoritative BIND 9 from being used to
4332amplify reflection denial-of-service (DoS) attacks. Short BADCOOKIE errors or
4333truncated (TC=1) responses can be sent to provide rate-limited responses to
4334legitimate clients within a range of forged, attacked IP addresses.
4335Legitimate clients react to dropped responses by retrying,
4336to BADCOOKIE errors by including a server cookie when retrying,
4337and to truncated responses by switching to TCP.
4338
4339This mechanism is intended for authoritative DNS servers. It can be used
4340on recursive servers, but can slow applications such as SMTP servers
4341(mail receivers) and HTTP clients (web browsers) that repeatedly request
4342the same domains. When possible, closing "open" recursive servers is
4343better.
4344
4345Response rate limiting uses a "credit" or "token bucket" scheme. Each
4346combination of identical response and client has a conceptual "account"
4347that earns a specified number of credits every second. A prospective
4348response debits its account by one. Responses are dropped or truncated
4349while the account is negative. Responses are tracked within a rolling
4350window of time which defaults to 15 seconds, but which can be configured with
4351the ``window`` option to any value from 1 to 3600 seconds (1 hour). The
4352account cannot become more positive than the per-second limit or more
4353negative than ``window`` times the per-second limit. When the specified
4354number of credits for a class of responses is set to 0, those responses
4355are not rate-limited.
4356
4357The notions of "identical response" and "DNS client" for rate limiting
4358are not simplistic. All responses to an address block are counted as if
4359to a single client. The prefix lengths of address blocks are specified
4360with ``ipv4-prefix-length`` (default 24) and ``ipv6-prefix-length``
4361(default 56).
4362
4363All non-empty responses for a valid domain name (qname) and record type
4364(qtype) are identical and have a limit specified with
4365``responses-per-second`` (default 0 or no limit). All valid wildcard
4366domain names are interpreted as the zone's origin name concatenated to
4367the "*" name. All empty (NODATA) responses for a valid domain,
4368regardless of query type, are identical.  Responses in the NODATA class
4369are limited by ``nodata-per-second`` (default ``responses-per-second``).
4370Requests for any and all undefined subdomains of a given valid domain
4371result in NXDOMAIN errors, and are identical regardless of query type.
4372They are limited by ``nxdomains-per-second`` (default
4373``responses-per-second``). This controls some attacks using random
4374names, but can be relaxed or turned off (set to 0) on servers that
4375expect many legitimate NXDOMAIN responses, such as from anti-spam
4376rejection lists. Referrals or delegations to the server of a given
4377domain are identical and are limited by ``referrals-per-second``
4378(default ``responses-per-second``).
4379
4380Responses generated from local wildcards are counted and limited as if
4381they were for the parent domain name. This controls flooding using
4382random.wild.example.com.
4383
4384All requests that result in DNS errors other than NXDOMAIN, such as
4385SERVFAIL and FORMERR, are identical regardless of requested name (qname)
4386or record type (qtype). This controls attacks using invalid requests or
4387distant, broken authoritative servers. By default the limit on errors is
4388the same as the ``responses-per-second`` value, but it can be set
4389separately with ``errors-per-second``.
4390
4391Many attacks using DNS involve UDP requests with forged source
4392addresses. Rate limiting prevents the use of BIND 9 to flood a network
4393with responses to requests with forged source addresses, but could let a
4394third party block responses to legitimate requests. There is a mechanism
4395that can answer some legitimate requests from a client whose address is
4396being forged in a flood. Setting ``slip`` to 2 (its default) causes
4397every other UDP request without a valid server cookie to be answered with
4398a small response. The small size and reduced frequency, and resulting lack of
4399amplification, of "slipped" responses make them unattractive for
4400reflection DoS attacks. ``slip`` must be between 0 and 10. A value of 0
4401does not "slip"; no small responses are sent due to rate limiting. Rather,
4402all responses are dropped. A value of 1 causes every response to slip;
4403values between 2 and 10 cause every nth response to slip.
4404
4405If the request included a client cookie, then a "slipped" response is
4406a BADCOOKIE error with a server cookie, which allows a legitimate client
4407to include the server cookie to be exempted from the rate limiting
4408when it retries the request.
4409If the request did not include a cookie, then a "slipped" response is
4410a truncated (TC=1) response, which prompts a legitimate client to
4411switch to TCP and thus be exempted from the rate limiting. Some error
4412responses, including REFUSED and SERVFAIL, cannot be replaced with
4413truncated responses and are instead leaked at the ``slip`` rate.
4414
4415(Note: dropped responses from an authoritative server may reduce the
4416difficulty of a third party successfully forging a response to a
4417recursive resolver. The best security against forged responses is for
4418authoritative operators to sign their zones using DNSSEC and for
4419resolver operators to validate the responses. When this is not an
4420option, operators who are more concerned with response integrity than
4421with flood mitigation may consider setting ``slip`` to 1, causing all
4422rate-limited responses to be truncated rather than dropped. This reduces
4423the effectiveness of rate-limiting against reflection attacks.)
4424
4425When the approximate query-per-second rate exceeds the ``qps-scale``
4426value, the ``responses-per-second``, ``errors-per-second``,
4427``nxdomains-per-second``, and ``all-per-second`` values are reduced by
4428the ratio of the current rate to the ``qps-scale`` value. This feature
4429can tighten defenses during attacks. For example, with
4430``qps-scale 250; responses-per-second 20;`` and a total query rate of
44311000 queries/second for all queries from all DNS clients including via
4432TCP, then the effective responses/second limit changes to (250/1000)*20,
4433or 5. Responses to requests that included a valid server cookie,
4434and responses sent via TCP, are not limited but are counted to compute
4435the query-per-second rate.
4436
4437Communities of DNS clients can be given their own parameters or no
4438rate limiting by putting ``rate-limit`` statements in ``view`` statements
4439instead of in the global ``option`` statement. A ``rate-limit`` statement
4440in a view replaces, rather than supplements, a ``rate-limit``
4441statement among the main options. DNS clients within a view can be
4442exempted from rate limits with the ``exempt-clients`` clause.
4443
4444UDP responses of all kinds can be limited with the ``all-per-second``
4445phrase. This rate limiting is unlike the rate limiting provided by
4446``responses-per-second``, ``errors-per-second``, and
4447``nxdomains-per-second`` on a DNS server, which are often invisible to
4448the victim of a DNS reflection attack. Unless the forged requests of the
4449attack are the same as the legitimate requests of the victim, the
4450victim's requests are not affected. Responses affected by an
4451``all-per-second`` limit are always dropped; the ``slip`` value has no
4452effect. An ``all-per-second`` limit should be at least 4 times as large
4453as the other limits, because single DNS clients often send bursts of
4454legitimate requests. For example, the receipt of a single mail message
4455can prompt requests from an SMTP server for NS, PTR, A, and AAAA records
4456as the incoming SMTP/TCP/IP connection is considered. The SMTP server
4457can need additional NS, A, AAAA, MX, TXT, and SPF records as it
4458considers the SMTP ``Mail From`` command. Web browsers often repeatedly
4459resolve the same names that are duplicated in HTML <IMG> tags in a page.
4460``all-per-second`` is similar to the rate limiting offered by firewalls
4461but is often inferior. Attacks that justify ignoring the contents of DNS
4462responses are likely to be attacks on the DNS server itself. They
4463usually should be discarded before the DNS server spends resources making
4464TCP connections or parsing DNS requests, but that rate limiting must be
4465done before the DNS server sees the requests.
4466
4467The maximum size of the table used to track requests and rate-limit
4468responses is set with ``max-table-size``. Each entry in the table is
4469between 40 and 80 bytes. The table needs approximately as many entries
4470as the number of requests received per second. The default is 20,000. To
4471reduce the cold start of growing the table, ``min-table-size`` (default 500)
4472can set the minimum table size. Enable ``rate-limit`` category
4473logging to monitor expansions of the table and inform choices for the
4474initial and maximum table size.
4475
4476Use ``log-only yes`` to test rate-limiting parameters without actually
4477dropping any requests.
4478
4479Responses dropped by rate limits are included in the ``RateDropped`` and
4480``QryDropped`` statistics. Responses that are truncated by rate limits are
4481included in ``RateSlipped`` and ``RespTruncated``.
4482
4483NXDOMAIN Redirection
4484^^^^^^^^^^^^^^^^^^^^
4485
4486``named`` supports NXDOMAIN redirection via two methods:
4487
4488-  Redirect zone (:ref:`zone_statement_grammar`)
4489-  Redirect namespace
4490
4491With either method, when ``named`` gets an NXDOMAIN response it examines a
4492separate namespace to see if the NXDOMAIN response should be replaced
4493with an alternative response.
4494
4495With a redirect zone (``zone "." { type redirect; };``), the data used
4496to replace the NXDOMAIN is held in a single zone which is not part of
4497the normal namespace. All the redirect information is contained in the
4498zone; there are no delegations.
4499
4500With a redirect namespace (``option { nxdomain-redirect <suffix> };``),
4501the data used to replace the NXDOMAIN is part of the normal namespace
4502and is looked up by appending the specified suffix to the original
4503query name. This roughly doubles the cache required to process
4504NXDOMAIN responses, as both the original NXDOMAIN response and the
4505replacement data (or an NXDOMAIN indicating that there is no
4506replacement) must be stored.
4507
4508If both a redirect zone and a redirect namespace are configured, the
4509redirect zone is tried first.
4510
4511.. _server_statement_grammar:
4512
4513``server`` Statement Grammar
4514~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4515
4516.. include:: ../misc/server.grammar.rst
4517
4518.. _server_statement_definition_and_usage:
4519
4520``server`` Statement Definition and Usage
4521~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4522
4523The ``server`` statement defines characteristics to be associated with a
4524remote name server. If a prefix length is specified, then a range of
4525servers is covered. Only the most specific server clause applies,
4526regardless of the order in ``named.conf``.
4527
4528The ``server`` statement can occur at the top level of the configuration
4529file or inside a ``view`` statement. If a ``view`` statement contains
4530one or more ``server`` statements, only those apply to the view and any
4531top-level ones are ignored. If a view contains no ``server`` statements,
4532any top-level ``server`` statements are used as defaults.
4533
4534If a remote server is giving out bad data, marking it
4535as bogus prevents further queries to it. The default value of
4536``bogus`` is ``no``.
4537
4538The ``provide-ixfr`` clause determines whether the local server, acting
4539as primary, responds with an incremental zone transfer when the given
4540remote server, a secondary, requests it. If set to ``yes``, incremental
4541transfer is provided whenever possible. If set to ``no``, all
4542transfers to the remote server are non-incremental. If not set, the
4543value of the ``provide-ixfr`` option in the view or global options block
4544is used as a default.
4545
4546The ``request-ixfr`` clause determines whether the local server, acting
4547as a secondary, requests incremental zone transfers from the given
4548remote server, a primary. If not set, the value of the ``request-ixfr``
4549option in the view or global options block is used as a default. It may
4550also be set in the zone block; if set there, it overrides the
4551global or view setting for that zone.
4552
4553IXFR requests to servers that do not support IXFR automatically
4554fall back to AXFR. Therefore, there is no need to manually list which
4555servers support IXFR and which ones do not; the global default of
4556``yes`` should always work. The purpose of the ``provide-ixfr`` and
4557``request-ixfr`` clauses is to make it possible to disable the use of
4558IXFR even when both primary and secondary claim to support it: for example, if
4559one of the servers is buggy and crashes or corrupts data when IXFR is
4560used.
4561
4562The ``request-expire`` clause determines whether the local server, when
4563acting as a secondary, requests the EDNS EXPIRE value. The EDNS EXPIRE
4564value indicates the remaining time before the zone data expires and
4565needs to be refreshed. This is used when a secondary server transfers
4566a zone from another secondary server; when transferring from the
4567primary, the expiration timer is set from the EXPIRE field of the SOA
4568record instead. The default is ``yes``.
4569
4570The ``edns`` clause determines whether the local server attempts to
4571use EDNS when communicating with the remote server. The default is
4572``yes``.
4573
4574The ``edns-udp-size`` option sets the EDNS UDP size that is advertised
4575by ``named`` when querying the remote server. Valid values are 512 to
45764096 bytes; values outside this range are silently adjusted to the
4577nearest value within it. This option is useful when
4578advertising a different value to this server than the value advertised
4579globally: for example, when there is a firewall at the remote site that
4580is blocking large replies. Note: currently, this sets a single UDP size
4581for all packets sent to the server; ``named`` does not deviate from this
4582value. This differs from the behavior of ``edns-udp-size`` in
4583``options`` or ``view`` statements, where it specifies a maximum value.
4584The ``server`` statement behavior may be brought into conformance with
4585the ``options``/``view`` behavior in future releases.
4586
4587The ``edns-version`` option sets the maximum EDNS VERSION that is
4588sent to the server(s) by the resolver. The actual EDNS version sent is
4589still subject to normal EDNS version-negotiation rules (see :rfc:`6891`),
4590the maximum EDNS version supported by the server, and any other
4591heuristics that indicate that a lower version should be sent. This
4592option is intended to be used when a remote server reacts badly to a
4593given EDNS version or higher; it should be set to the highest version
4594the remote server is known to support. Valid values are 0 to 255; higher
4595values are silently adjusted. This option is not needed until
4596higher EDNS versions than 0 are in use.
4597
4598The ``max-udp-size`` option sets the maximum EDNS UDP message size
4599``named`` sends. Valid values are 512 to 4096 bytes; values outside
4600this range are silently adjusted. This option is useful when
4601there is a firewall that is blocking large replies from
4602``named``.
4603
4604The ``padding`` option adds EDNS Padding options to outgoing messages,
4605increasing the packet size to a multiple of the specified block size.
4606Valid block sizes range from 0 (the default, which disables the use of
4607EDNS Padding) to 512 bytes. Larger values are reduced to 512, with a
4608logged warning. Note: this option is not currently compatible with no
4609TSIG or SIG(0), as the EDNS OPT record containing the padding would have
4610to be added to the packet after it had already been signed.
4611
4612The ``tcp-only`` option sets the transport protocol to TCP. The default
4613is to use the UDP transport and to fallback on TCP only when a truncated
4614response is received.
4615
4616The ``tcp-keepalive`` option adds EDNS TCP keepalive to messages sent
4617over TCP. Note that currently idle timeouts in responses are ignored.
4618
4619The server supports two zone transfer methods. The first,
4620``one-answer``, uses one DNS message per resource record transferred.
4621``many-answers`` packs as many resource records as possible into a single
4622message, which is more efficient.
4623It is possible to specify which method to use for a server via the
4624``transfer-format`` option; if not set there, the
4625``transfer-format`` specified by the ``options`` statement is used.
4626
4627``transfers`` is used to limit the number of concurrent inbound zone
4628transfers from the specified server. If no ``transfers`` clause is
4629specified, the limit is set according to the ``transfers-per-ns``
4630option.
4631
4632The ``keys`` clause identifies a ``key_id`` defined by the ``key``
4633statement, to be used for transaction security (see :ref:`tsig`)
4634when talking to the remote server. When a request is sent to the remote
4635server, a request signature is generated using the key specified
4636here and appended to the message. A request originating from the remote
4637server is not required to be signed by this key.
4638
4639Only a single key per server is currently supported.
4640
4641The ``transfer-source`` and ``transfer-source-v6`` clauses specify the
4642IPv4 and IPv6 source address, respectively, to be used for zone transfer with the
4643remote server. For an IPv4 remote server, only
4644``transfer-source`` can be specified. Similarly, for an IPv6 remote
4645server, only ``transfer-source-v6`` can be specified. For more details,
4646see the description of ``transfer-source`` and ``transfer-source-v6`` in
4647:ref:`zone_transfers`.
4648
4649The ``notify-source`` and ``notify-source-v6`` clauses specify the IPv4
4650and IPv6 source address, respectively, to be used for notify messages sent to remote
4651servers. For an IPv4 remote server, only ``notify-source``
4652can be specified. Similarly, for an IPv6 remote server, only
4653``notify-source-v6`` can be specified.
4654
4655The ``query-source`` and ``query-source-v6`` clauses specify the IPv4
4656and IPv6 source address, respectively, to be used for queries sent to remote servers.
4657For an IPv4 remote server, only ``query-source`` can be
4658specified. Similarly, for an IPv6 remote server, only
4659``query-source-v6`` can be specified.
4660
4661The ``request-nsid`` clause determines whether the local server adds
4662an NSID EDNS option to requests sent to the server. This overrides
4663``request-nsid`` set at the view or option level.
4664
4665The ``send-cookie`` clause determines whether the local server adds
4666a COOKIE EDNS option to requests sent to the server. This overrides
4667``send-cookie`` set at the view or option level. The ``named`` server
4668may determine that COOKIE is not supported by the remote server and not
4669add a COOKIE EDNS option to requests.
4670
4671.. _statschannels:
4672
4673``statistics-channels`` Statement Grammar
4674~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4675
4676.. include:: ../misc/statistics-channels.grammar.rst
4677
4678.. _statistics_channels:
4679
4680``statistics-channels`` Statement Definition and Usage
4681~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4682
4683The ``statistics-channels`` statement declares communication channels to
4684be used by system administrators to get access to statistics information
4685on the name server.
4686
4687This statement is intended to be flexible to support multiple communication
4688protocols in the future, but currently only HTTP access is supported. It
4689requires that BIND 9 be compiled with libxml2 and/or json-c (also known
4690as libjson0); the ``statistics-channels`` statement is still accepted
4691even if it is built without the library, but any HTTP access fails
4692with an error.
4693
4694An ``inet`` control channel is a TCP socket listening at the specified
4695``ip_port`` on the specified ``ip_addr``, which can be an IPv4 or IPv6
4696address. An ``ip_addr`` of ``*`` (asterisk) is interpreted as the IPv4
4697wildcard address; connections are accepted on any of the system's
4698IPv4 addresses. To listen on the IPv6 wildcard address, use an
4699``ip_addr`` of ``::``.
4700
4701If no port is specified, port 80 is used for HTTP channels. The asterisk
4702(``*``) cannot be used for ``ip_port``.
4703
4704Attempts to open a statistics channel are restricted by the
4705optional ``allow`` clause. Connections to the statistics channel are
4706permitted based on the ``address_match_list``. If no ``allow`` clause is
4707present, ``named`` accepts connection attempts from any address. Since
4708the statistics may contain sensitive internal information, the source of
4709connection requests must be restricted appropriately so that only
4710trusted parties can access the statistics channel.
4711
4712Gathering data exposed by the statistics channel locks various subsystems in
4713``named``, which could slow down query processing if statistics data is
4714requested too often.
4715
4716An issue in the statistics channel would be considered a security issue
4717only if it could be exploited by unprivileged users circumventing the access
4718control list. In other words, any issue in the statistics channel that could be
4719used to access information unavailable otherwise, or to crash ``named``, is
4720not considered a security issue if it can be avoided through the
4721use of a secure configuration.
4722
4723If no ``statistics-channels`` statement is present, ``named`` does not
4724open any communication channels.
4725
4726The statistics are available in various formats and views, depending on
4727the URI used to access them. For example, if the statistics channel is
4728configured to listen on 127.0.0.1 port 8888, then the statistics are
4729accessible in XML format at http://127.0.0.1:8888/ or
4730http://127.0.0.1:8888/xml. A CSS file is included, which can format the
4731XML statistics into tables when viewed with a stylesheet-capable
4732browser, and into charts and graphs using the Google Charts API when
4733using a JavaScript-capable browser.
4734
4735Broken-out subsets of the statistics can be viewed at
4736http://127.0.0.1:8888/xml/v3/status (server uptime and last
4737reconfiguration time), http://127.0.0.1:8888/xml/v3/server (server and
4738resolver statistics), http://127.0.0.1:8888/xml/v3/zones (zone
4739statistics), http://127.0.0.1:8888/xml/v3/net (network status and socket
4740statistics), http://127.0.0.1:8888/xml/v3/mem (memory manager
4741statistics), http://127.0.0.1:8888/xml/v3/tasks (task manager
4742statistics), and http://127.0.0.1:8888/xml/v3/traffic (traffic sizes).
4743
4744The full set of statistics can also be read in JSON format at
4745http://127.0.0.1:8888/json, with the broken-out subsets at
4746http://127.0.0.1:8888/json/v1/status (server uptime and last
4747reconfiguration time), http://127.0.0.1:8888/json/v1/server (server and
4748resolver statistics), http://127.0.0.1:8888/json/v1/zones (zone
4749statistics), http://127.0.0.1:8888/json/v1/net (network status and
4750socket statistics), http://127.0.0.1:8888/json/v1/mem (memory manager
4751statistics), http://127.0.0.1:8888/json/v1/tasks (task manager
4752statistics), and http://127.0.0.1:8888/json/v1/traffic (traffic sizes).
4753
4754.. _trust_anchors:
4755
4756``trust-anchors`` Statement Grammar
4757~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4758
4759.. include:: ../misc/trust-anchors.grammar.rst
4760
4761.. _trust-anchors:
4762
4763``trust-anchors`` Statement Definition and Usage
4764~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4765
4766The ``trust-anchors`` statement defines DNSSEC trust anchors. DNSSEC is
4767described in :ref:`DNSSEC`.
4768
4769A trust anchor is defined when the public key or public key digest for a non-authoritative
4770zone is known but cannot be securely obtained through DNS, either
4771because it is the DNS root zone or because its parent zone is unsigned.
4772Once a key or digest has been configured as a trust anchor, it is treated as if it
4773has been validated and proven secure.
4774
4775The resolver attempts DNSSEC validation on all DNS data in subdomains of
4776configured trust anchors. Validation below specified names can be
4777temporarily disabled by using ``rndc nta``, or permanently disabled with
4778the ``validate-except`` option.
4779
4780All keys listed in ``trust-anchors``, and their corresponding zones, are
4781deemed to exist regardless of what parent zones say. Only keys
4782configured as trust anchors are used to validate the DNSKEY RRset for
4783the corresponding name. The parent's DS RRset is not used.
4784
4785``trust-anchors`` may be set at the top level of ``named.conf`` or within
4786a view. If it is set in both places, the configurations are additive;
4787keys defined at the top level are inherited by all views, but keys
4788defined in a view are only used within that view.
4789
4790The ``trust-anchors`` statement can contain
4791multiple trust-anchor entries, each consisting of a
4792domain name, followed by an "anchor type" keyword indicating
4793the trust anchor's format, followed by the key or digest data.
4794
4795If the anchor type is ``static-key`` or
4796``initial-key``, then it is followed with the
4797key's flags, protocol, and algorithm, plus the Base64 representation
4798of the public key data. This is identical to the text
4799representation of a DNSKEY record.  Spaces, tabs, newlines, and
4800carriage returns are ignored in the key data, so the
4801configuration may be split into multiple lines.
4802
4803If the anchor type is ``static-ds`` or
4804``initial-ds``, it is followed with the
4805key tag, algorithm, digest type, and the hexadecimal
4806representation of the key digest. This is identical to the
4807text representation of a DS record.  Spaces, tabs, newlines,
4808and carriage returns are ignored.
4809
4810Trust anchors configured with the
4811``static-key`` or ``static-ds``
4812anchor types are immutable, while keys configured with
4813``initial-key`` or ``initial-ds``
4814can be kept up-to-date automatically, without intervention from the resolver operator.
4815(``static-key`` keys are identical to keys configured using the
4816deprecated ``trusted-keys`` statement.)
4817
4818Suppose, for example, that a zone's key-signing key was compromised, and
4819the zone owner had to revoke and replace the key. A resolver which had
4820the original key
4821configured using ``static-key`` or
4822``static-ds`` would be unable to validate
4823this zone any longer; it would reply with a SERVFAIL response
4824code.  This would continue until the resolver operator had
4825updated the ``trust-anchors`` statement with
4826the new key.
4827
4828If, however, the trust anchor had been configured using
4829``initial-key`` or ``initial-ds``
4830instead, the zone owner could add a "stand-by" key to
4831the zone in advance. ``named`` would store
4832the stand-by key, and when the original key was revoked,
4833``named`` would be able to transition smoothly
4834to the new key.  It would also recognize that the old key had
4835been revoked and cease using that key to validate answers,
4836minimizing the damage that the compromised key could do.
4837This is the process used to keep the ICANN root DNSSEC key
4838up-to-date.
4839
4840Whereas ``static-key`` and
4841``static-ds`` trust anchors continue
4842to be trusted until they are removed from
4843``named.conf``, an
4844``initial-key`` or ``initial-ds``
4845is only trusted *once*: for as long as it
4846takes to load the managed key database and start the
4847:rfc:`5011` key maintenance process.
4848
4849It is not possible to mix static with initial trust anchors
4850for the same domain name.
4851
4852The first time ``named`` runs with an
4853``initial-key`` or ``initial-ds``
4854configured in ``named.conf``, it fetches the
4855DNSKEY RRset directly from the zone apex,
4856and validates it
4857using the trust anchor specified in ``trust-anchors``.
4858If the DNSKEY RRset is validly signed by a key matching
4859the trust anchor, then it is used as the basis for a new
4860managed-keys database.
4861
4862From that point on, whenever ``named`` runs, it sees the ``initial-key`` or ``initial-ds``
4863listed in ``trust-anchors``, checks to make sure :rfc:`5011` key maintenance
4864has already been initialized for the specified domain, and if so,
4865simply moves on. The key specified in the ``trust-anchors`` statement is
4866not used to validate answers; it is superseded by the key or keys stored
4867in the managed-keys database.
4868
4869The next time ``named`` runs after an ``initial-key`` or ``initial-ds`` has been *removed*
4870from the ``trust-anchors`` statement (or changed to a ``static-key`` or ``static-ds``), the
4871corresponding zone is removed from the managed-keys database, and
4872:rfc:`5011` key maintenance is no longer used for that domain.
4873
4874In the current implementation, the managed-keys database is stored as a
4875master-format zone file.
4876
4877On servers which do not use views, this file is named
4878``managed-keys.bind``. When views are in use, there is a separate
4879managed-keys database for each view; the filename is the view name
4880(or, if a view name contains characters which would make it illegal as a
4881filename, a hash of the view name), followed by the suffix ``.mkeys``.
4882
4883When the key database is changed, the zone is updated. As with any other
4884dynamic zone, changes are written into a journal file, e.g.,
4885``managed-keys.bind.jnl`` or ``internal.mkeys.jnl``. Changes are
4886committed to the primary file as soon as possible afterward,
4887usually within 30 seconds. Whenever ``named`` is using
4888automatic key maintenance, the zone file and journal file can be
4889expected to exist in the working directory. (For this reason, among
4890others, the working directory should be always be writable by
4891``named``.)
4892
4893If the ``dnssec-validation`` option is set to ``auto``, ``named``
4894automatically initializes an ``initial-key`` for the root zone. The key
4895that is used to initialize the key-maintenance process is stored in
4896``bind.keys``; the location of this file can be overridden with the
4897``bindkeys-file`` option. As a fallback in the event no ``bind.keys``
4898can be found, the initializing key is also compiled directly into
4899``named``.
4900
4901.. _dnssec_policy_grammar:
4902
4903``dnssec-policy`` Statement Grammar
4904~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4905
4906.. include:: ../misc/dnssec-policy.grammar.rst
4907
4908.. _dnssec_policy:
4909
4910``dnssec-policy`` Statement Definition and Usage
4911~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4912
4913The ``dnssec-policy`` statement defines a key and signing policy (KASP)
4914for zones.
4915
4916A KASP determines how one or more zones are signed with DNSSEC.  For
4917example, it specifies how often keys should roll, which cryptographic
4918algorithms to use, and how often RRSIG records need to be refreshed.
4919
4920Keys are not shared among zones, which means that one set of keys per
4921zone is generated even if they have the same policy.  If multiple views
4922are configured with different versions of the same zone, each separate
4923version uses the same set of signing keys.
4924
4925Multiple key and signing policies can be configured.  To attach a policy
4926to a zone, add a ``dnssec-policy`` option to the ``zone`` statement,
4927specifying the name of the policy that should be used.
4928
4929The ``dnssec-policy`` statement requires dynamic DNS to be set up, or
4930``inline-signing`` to be enabled.
4931
4932If ``inline-signing`` is enabled, this means that a signed version of the
4933zone is maintained separately and is written out to a different file on disk
4934(the zone's filename plus a ``.signed`` extension).
4935
4936If the zone is dynamic because it is configured with an ``update-policy`` or
4937``allow-update``, the DNSSEC records are written to the filename set in the
4938original zone's ``file``, unless ``inline-signing`` is explicitly set.
4939
4940Key rollover timing is computed for each key according to the key
4941lifetime defined in the KASP.  The lifetime may be modified by zone TTLs
4942and propagation delays, to prevent validation failures.  When a key
4943reaches the end of its lifetime, ``named`` generates and publishes a new
4944key automatically, then deactivates the old key and activates the new
4945one; finally, the old key is retired according to a computed schedule.
4946
4947Zone-signing key (ZSK) rollovers require no operator input.  Key-signing
4948key (KSK) and combined-signing key (CSK) rollovers require action to be
4949taken to submit a DS record to the parent.  Rollover timing for KSKs and
4950CSKs is adjusted to take into account delays in processing and
4951propagating DS updates.
4952
4953There are two predefined ``dnssec-policy`` names: ``none`` and
4954``default``.  Setting a zone's policy to ``none`` is the same as not
4955setting ``dnssec-policy`` at all; the zone is not signed.  Policy
4956``default`` causes the zone to be signed with a single combined-signing
4957key (CSK) using algorithm ECDSAP256SHA256; this key has an unlimited
4958lifetime.  (A verbose copy of this policy may be found in the source
4959tree, in the file ``doc/misc/dnssec-policy.default.conf``.)
4960
4961.. note:: The default signing policy may change in future releases.
4962   This could require changes to a signing policy when upgrading to a
4963   new version of BIND.  Check the release notes carefully when
4964   upgrading to be informed of such changes.  To prevent policy changes
4965   on upgrade, use an explicitly defined ``dnssec-policy``, rather than
4966   ``default``.
4967
4968If a ``dnssec-policy`` statement is modified and the server restarted or
4969reconfigured, ``named`` attempts to change the policy smoothly from the
4970old one to the new.  For example, if the key algorithm is changed, then
4971a new key is generated with the new algorithm, and the old algorithm is
4972retired when the existing key's lifetime ends.
4973
4974.. note:: Rolling to a new policy while another key rollover is already
4975   in progress is not yet supported, and may result in unexpected
4976   behavior.
4977
4978The following options can be specified in a ``dnssec-policy`` statement:
4979
4980  ``dnskey-ttl``
4981    This indicates the TTL to use when generating DNSKEY resource
4982    records.  The default is 1 hour (3600 seconds).
4983
4984  ``keys``
4985    This is a list specifying the algorithms and roles to use when
4986    generating keys and signing the zone.  Entries in this list do not
4987    represent specific DNSSEC keys, which may be changed on a regular
4988    basis, but the roles that keys play in the signing policy.  For
4989    example, configuring a KSK of algorithm RSASHA256 ensures that the
4990    DNSKEY RRset always includes a key-signing key for that algorithm.
4991
4992    Here is an example (for illustration purposes only) of some possible
4993    entries in a ``keys`` list:
4994
4995    ::
4996
4997        keys {
4998            ksk key-directory lifetime unlimited algorithm rsasha256 2048;
4999            zsk lifetime P30D algorithm 8;
5000            csk lifetime P6MT12H3M15S algorithm ecdsa256;
5001        };
5002
5003    This example specifies that three keys should be used in the zone.
5004    The first token determines which role the key plays in signing
5005    RRsets.  If set to ``ksk``, then this is a key-signing key; it has
5006    the KSK flag set and is only used to sign DNSKEY, CDS, and CDNSKEY
5007    RRsets.  If set to ``zsk``, this is a zone-signing key; the KSK flag
5008    is unset, and the key signs all RRsets *except* DNSKEY, CDS, and
5009    CDNSKEY.  If set to ``csk``, the key has the KSK flag set and is
5010    used to sign all RRsets.
5011
5012    An optional second token determines where the key is stored.
5013    Currently, keys can only be stored in the configured
5014    ``key-directory``.  This token may be used in the future to store
5015    keys in hardware security modules or separate directories.
5016
5017    The ``lifetime`` parameter specifies how long a key may be used
5018    before rolling over.  In the example above, the first key has an
5019    unlimited lifetime, the second key may be used for 30 days, and the
5020    third key has a rather peculiar lifetime of 6 months, 12 hours, 3
5021    minutes, and 15 seconds.  A lifetime of 0 seconds is the same as
5022    ``unlimited``.
5023
5024    Note that the lifetime of a key may be extended if retiring it too
5025    soon would cause validation failures.  For example, if the key were
5026    configured to roll more frequently than its own TTL, its lifetime
5027    would automatically be extended to account for this.
5028
5029    The ``algorithm`` parameter specifies the key's algorithm, expressed
5030    either as a string ("rsasha256", "ecdsa384", etc.) or as a decimal
5031    number.  An optional second parameter specifies the key's size in
5032    bits.  If it is omitted, as shown in the example for the second and
5033    third keys, an appropriate default size for the algorithm is used.
5034    Each KSK/ZSK pair must have the same algorithm. A CSK combines the
5035    functionality of a ZSK and a KSK.
5036
5037  ``purge-keys``
5038    This is the time after when DNSSEC keys that have been deleted from
5039    the zone can be removed from disk. If a key still determined to have
5040    presence (for example in some resolver cache), ``named`` will not
5041    remove the key files.
5042
5043    The default is ``P90D`` (90 days). Set this option to ``0`` to never
5044    purge deleted keys.
5045
5046  ``publish-safety``
5047    This is a margin that is added to the pre-publication interval in
5048    rollover timing calculations, to give some extra time to cover
5049    unforeseen events.  This increases the time between when keys are
5050    published and when they become active.  The default is ``PT1H`` (1
5051    hour).
5052
5053  ``retire-safety``
5054    This is a margin that is added to the post-publication interval in
5055    rollover timing calculations, to give some extra time to cover
5056    unforeseen events.  This increases the time a key remains published
5057    after it is no longer active.  The default is ``PT1H`` (1 hour).
5058
5059  ``signatures-refresh``
5060    This determines how frequently an RRSIG record needs to be
5061    refreshed.  The signature is renewed when the time until the
5062    expiration time is less than the specified interval.  The default is
5063    ``P5D`` (5 days), meaning signatures that expire in 5 days or sooner
5064    are refreshed.
5065
5066  ``signatures-validity``
5067    This indicates the validity period of an RRSIG record (subject to
5068    inception offset and jitter).  The default is ``P2W`` (2 weeks).
5069
5070  ``signatures-validity-dnskey``
5071    This is similar to ``signatures-validity``, but for DNSKEY records.
5072    The default is ``P2W`` (2 weeks).
5073
5074  ``max-zone-ttl``
5075
5076   This specifies the maximum permissible TTL value for the zone.  When
5077   a zone file is loaded, any record encountered with a TTL higher than
5078   ``max-zone-ttl`` causes the zone to be rejected.
5079
5080   This ensures that when rolling to a new DNSKEY, the old key will remain
5081   available until RRSIG records have expired from caches. The
5082   ``max-zone-ttl`` option guarantees that the largest TTL in the
5083   zone is no higher than a known and predictable value.
5084
5085    .. note:: Because ``map``-format files load directly into memory,
5086       this option cannot be used with them.
5087
5088   The default value ``PT24H`` (24 hours).  A value of zero is treated
5089   as if the default value were in use.
5090
5091
5092  ``nsec3param``
5093    Use NSEC3 instead of NSEC, and optionally set the NSEC3 parameters.
5094
5095    Here is an example of an ``nsec3`` configuration:
5096
5097    ::
5098
5099        nsec3param iterations 5 optout no salt-length 8;
5100
5101    The default is to use NSEC. The ``iterations``, ``optout``, and
5102    ``salt-length`` parts are optional, but if not set, the values in
5103    the example above are the default NSEC3 parameters. Note that the
5104    specific salt string is not specified by the user; :iscman:`named` creates a salt
5105    of the indicated length.
5106
5107    .. warning::
5108       Do not use extra :term:`iterations <Iterations>`, :term:`salt <Salt>`, and
5109       :term:`opt-out <Opt-out>` unless their implications are fully understood.
5110       A higher number of iterations causes interoperability problems and opens
5111       servers to CPU-exhausting DoS attacks.
5112
5113  ``zone-propagation-delay``
5114    This is the expected propagation delay from the time when a zone is
5115    first updated to the time when the new version of the zone is served
5116    by all secondary servers.  The default is ``PT5M`` (5 minutes).
5117
5118  ``parent-ds-ttl``
5119    This is the TTL of the DS RRset that the parent zone uses.  The
5120    default is ``P1D`` (1 day).
5121
5122  ``parent-propagation-delay``
5123    This is the expected propagation delay from the time when the parent
5124    zone is updated to the time when the new version is served by all of
5125    the parent zone's name servers.  The default is ``PT1H`` (1 hour).
5126
5127Automated KSK Rollovers
5128^^^^^^^^^^^^^^^^^^^^^^^
5129
5130BIND has mechanisms in place to facilitate automated KSK rollovers. It
5131publishes CDS and CDNSKEY records that can be used by the parent zone to
5132publish or withdraw the zone's DS records. BIND will query the parental
5133agents to see if the new DS is actually published before withdrawing the
5134old DNSSEC key.
5135
5136   .. note::
5137      The DS response is not validated so it is recommended to set up a
5138      trust relationship with the parental agent. For example, use TSIG to
5139      authenticate the parental agent, or point to a validating resolver.
5140
5141The following options apply to DS queries sent to ``parental-agents``:
5142
5143``parental-source``
5144   ``parental-source`` determines which local source address, and optionally
5145   UDP port, is used to send parental DS queries. This statement sets the
5146   ``parental-source`` for all zones, but can be overridden on a per-zone or
5147   per-view basis by including a ``parental-source`` statement within the
5148   ``zone`` or ``view`` block in the configuration file.
5149
5150   .. warning:: Specifying a single port is discouraged, as it removes a layer of
5151      protection against spoofing errors.
5152
5153   .. warning:: The configured ``port`` must not be same as the listening port.
5154
5155``parental-source-v6``
5156   This option acts like ``parental-source``, but applies to parental DS
5157   queries sent to IPv6 addresses.
5158
5159.. _managed-keys:
5160
5161``managed-keys`` Statement Grammar
5162~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
5163
5164.. include:: ../misc/managed-keys.grammar.rst
5165
5166.. _managed_keys:
5167
5168``managed-keys`` Statement Definition and Usage
5169~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
5170
5171The ``managed-keys`` statement has been
5172deprecated in favor of :ref:`trust_anchors`
5173with the ``initial-key`` keyword.
5174
5175.. _trusted-keys:
5176
5177``trusted-keys`` Statement Grammar
5178~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
5179
5180.. include:: ../misc/trusted-keys.grammar.rst
5181
5182.. _trusted_keys:
5183
5184``trusted-keys`` Statement Definition and Usage
5185~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
5186
5187The ``trusted-keys`` statement has been deprecated in favor of
5188:ref:`trust_anchors` with the ``static-key`` keyword.
5189
5190.. _view_statement_grammar:
5191
5192``view`` Statement Grammar
5193~~~~~~~~~~~~~~~~~~~~~~~~~~
5194
5195::
5196
5197   view view_name [ class ] {
5198       match-clients { address_match_list } ;
5199       match-destinations { address_match_list } ;
5200       match-recursive-only yes_or_no ;
5201     [ view_option ; ... ]
5202     [ zone_statement ; ... ]
5203   } ;
5204
5205.. _view_statement:
5206
5207``view`` Statement Definition and Usage
5208~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
5209
5210The ``view`` statement is a powerful feature of BIND 9 that lets a name
5211server answer a DNS query differently depending on who is asking. It is
5212particularly useful for implementing split DNS setups without having to
5213run multiple servers.
5214
5215Each ``view`` statement defines a view of the DNS namespace that is
5216seen by a subset of clients. A client matches a view if its source IP
5217address matches the ``address_match_list`` of the view's
5218``match-clients`` clause, and its destination IP address matches the
5219``address_match_list`` of the view's ``match-destinations`` clause. If
5220not specified, both ``match-clients`` and ``match-destinations`` default
5221to matching all addresses. In addition to checking IP addresses,
5222``match-clients`` and ``match-destinations`` can also take ``keys``
5223which provide an mechanism for the client to select the view. A view can
5224also be specified as ``match-recursive-only``, which means that only
5225recursive requests from matching clients match that view. The order
5226of the ``view`` statements is significant; a client request is
5227resolved in the context of the first ``view`` that it matches.
5228
5229Zones defined within a ``view`` statement are only accessible to
5230clients that match the ``view``. By defining a zone of the same name in
5231multiple views, different zone data can be given to different clients:
5232for example, "internal" and "external" clients in a split DNS setup.
5233
5234Many of the options given in the ``options`` statement can also be used
5235within a ``view`` statement, and then apply only when resolving queries
5236with that view. When no view-specific value is given, the value in the
5237``options`` statement is used as a default. Also, zone options can have
5238default values specified in the ``view`` statement; these view-specific
5239defaults take precedence over those in the ``options`` statement.
5240
5241Views are class-specific. If no class is given, class IN is assumed.
5242Note that all non-IN views must contain a hint zone, since only the IN
5243class has compiled-in default hints.
5244
5245If there are no ``view`` statements in the config file, a default view
5246that matches any client is automatically created in class IN. Any
5247``zone`` statements specified on the top level of the configuration file
5248are considered to be part of this default view, and the ``options``
5249statement applies to the default view. If any explicit ``view``
5250statements are present, all ``zone`` statements must occur inside
5251``view`` statements.
5252
5253Here is an example of a typical split DNS setup implemented using
5254``view`` statements:
5255
5256::
5257
5258   view "internal" {
5259         // This should match our internal networks.
5260         match-clients { 10.0.0.0/8; };
5261
5262         // Provide recursive service to internal
5263         // clients only.
5264         recursion yes;
5265
5266         // Provide a complete view of the example.com
5267         // zone including addresses of internal hosts.
5268         zone "example.com" {
5269           type primary;
5270           file "example-internal.db";
5271         };
5272   };
5273
5274   view "external" {
5275         // Match all clients not matched by the
5276         // previous view.
5277         match-clients { any; };
5278
5279         // Refuse recursive service to external clients.
5280         recursion no;
5281
5282         // Provide a restricted view of the example.com
5283         // zone containing only publicly accessible hosts.
5284         zone "example.com" {
5285          type primary;
5286          file "example-external.db";
5287         };
5288   };
5289
5290.. _zone_statement_grammar:
5291
5292``zone`` Statement Grammar
5293~~~~~~~~~~~~~~~~~~~~~~~~~~
5294
5295.. include:: ../misc/master.zoneopt.rst
5296.. include:: ../misc/slave.zoneopt.rst
5297.. include:: ../misc/mirror.zoneopt.rst
5298.. include:: ../misc/hint.zoneopt.rst
5299.. include:: ../misc/stub.zoneopt.rst
5300.. include:: ../misc/static-stub.zoneopt.rst
5301.. include:: ../misc/forward.zoneopt.rst
5302.. include:: ../misc/redirect.zoneopt.rst
5303.. include:: ../misc/delegation-only.zoneopt.rst
5304.. include:: ../misc/in-view.zoneopt.rst
5305
5306.. _zone_statement:
5307
5308``zone`` Statement Definition and Usage
5309~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
5310
5311.. _zone_types:
5312
5313Zone Types
5314^^^^^^^^^^
5315
5316The ``type`` keyword is required for the ``zone`` configuration unless
5317it is an ``in-view`` configuration. Its acceptable values are:
5318``primary`` (or ``master``), ``secondary`` (or ``slave``), ``mirror``,
5319``hint``, ``stub``, ``static-stub``, ``forward``, ``redirect``,
5320or ``delegation-only``.
5321
5322``primary``
5323   A primary zone has a master copy of the data for the zone and is able
5324   to provide authoritative answers for it. Type ``master`` is a synonym
5325   for ``primary``.
5326
5327``secondary``
5328    A secondary zone is a replica of a primary zone. Type ``slave`` is a
5329    synonym for ``secondary``. The ``primaries`` list specifies one or more IP
5330    addresses of primary servers that the secondary contacts to update
5331    its copy of the zone.  Primaries list elements can
5332    also be names of other primaries lists.  By default,
5333    transfers are made from port 53 on the servers;
5334    this can be changed for all servers by specifying
5335    a port number before the list of IP addresses,
5336    or on a per-server basis after the IP address.
5337    Authentication to the primary can also be done with
5338    per-server TSIG keys.  If a file is specified, then the
5339    replica is written to this file
5340    whenever the zone
5341    is changed, and reloaded from this file on a server
5342    restart. Use of a file is recommended, since it
5343    often speeds server startup and eliminates a
5344    needless waste of bandwidth. Note that for large
5345    numbers (in the tens or hundreds of thousands) of
5346    zones per server, it is best to use a two-level
5347    naming scheme for zone filenames. For example,
5348    a secondary server for the zone
5349    ``example.com`` might place
5350    the zone contents into a file called
5351    ``ex/example.com``, where
5352    ``ex/`` is just the first two
5353    letters of the zone name. (Most operating systems
5354    behave very slowly if there are 100000 files in a single directory.)
5355
5356``mirror``
5357   A mirror zone is similar to a zone of type ``secondary``, except its
5358   data is subject to DNSSEC validation before being used in answers.
5359   Validation is applied to the entire zone during the zone transfer
5360   process, and again when the zone file is loaded from disk upon
5361   restarting ``named``. If validation of a new version of a mirror zone
5362   fails, a retransfer is scheduled; in the meantime, the most recent
5363   correctly validated version of that zone is used until it either
5364   expires or a newer version validates correctly. If no usable zone
5365   data is available for a mirror zone, due to either transfer failure
5366   or expiration, traditional DNS recursion is used to look up the
5367   answers instead. Mirror zones cannot be used in a view that does not
5368   have recursion enabled.
5369
5370   Answers coming from a mirror zone look almost exactly like answers
5371   from a zone of type ``secondary``, with the notable exceptions that
5372   the AA bit ("authoritative answer") is not set, and the AD bit
5373   ("authenticated data") is.
5374
5375   Mirror zones are intended to be used to set up a fast local copy of
5376   the root zone (see :rfc:`8806`). A default list of primary servers
5377   for the IANA root zone is built into ``named``, so its mirroring can
5378   be enabled using the following configuration:
5379
5380   ::
5381
5382       zone "." {
5383           type mirror;
5384       };
5385
5386   Mirror zone validation always happens for the entire zone contents.
5387   This ensures that each version of the zone used by the resolver is
5388   fully self-consistent with respect to DNSSEC. For incoming mirror
5389   zone IXFRs, every revision of the zone contained in the IXFR sequence
5390   is validated independently, in the order in which the zone revisions
5391   appear on the wire. For this reason, it might be useful to force use
5392   of AXFR for mirror zones by setting ``request-ixfr no;`` for the
5393   relevant zone (or view). Other, more efficient zone verification
5394   methods may be added in the future.
5395
5396   To make mirror zone contents persist between ``named`` restarts, use
5397   the :ref:`file <file-option>` option.
5398
5399   Mirroring a zone other than root requires an explicit list of primary
5400   servers to be provided using the ``primaries`` option (see
5401   :ref:`primaries_grammar` for details), and a key-signing key (KSK)
5402   for the specified zone to be explicitly configured as a trust anchor
5403   (see :ref:`trust-anchors`).
5404
5405   When configuring NOTIFY for a mirror zone, only ``notify no;`` and
5406   ``notify explicit;`` can be used at the zone level; any other
5407   ``notify`` setting at the zone level is a configuration error. Using
5408   any other ``notify`` setting at the ``options`` or ``view`` level
5409   causes that setting to be overridden with ``notify explicit;`` for
5410   the mirror zone. The global default for the ``notify`` option is
5411   ``yes``, so mirror zones are by default configured with ``notify
5412   explicit;``.
5413
5414   Outgoing transfers of mirror zones are disabled by default but may be
5415   enabled using :ref:`allow-transfer <allow-transfer-access>`.
5416
5417   .. note::
5418      Use of this zone type with any zone other than the root should be
5419      considered *experimental* and may cause performance issues,
5420      especially for zones that are large and/or frequently updated.
5421
5422``hint``
5423   The initial set of root name servers is specified using a hint zone.
5424   When the server starts, it uses the root hints to find a root name
5425   server and get the most recent list of root name servers. If no hint zone
5426   is specified for class IN, the server uses a compiled-in default set of
5427   root servers hints. Classes other than IN have no built-in default hints.
5428
5429``stub``
5430   A stub zone is similar to a secondary zone, except that it replicates only
5431   the NS records of a primary zone instead of the entire zone. Stub zones
5432   are not a standard part of the DNS; they are a feature specific to the
5433   BIND implementation.
5434
5435   Stub zones can be used to eliminate the need for a glue NS record in a parent
5436   zone, at the expense of maintaining a stub zone entry and a set of name
5437   server addresses in ``named.conf``. This usage is not recommended for
5438   new configurations, and BIND 9 supports it only in a limited way. If a BIND 9
5439   primary, serving a parent zone, has child stub
5440   zones configured, all the secondary servers for the parent zone also need to
5441   have the same child stub zones configured.
5442
5443   Stub zones can also be used as a way to force the resolution of a given
5444   domain to use a particular set of authoritative servers. For example, the
5445   caching name servers on a private network using :rfc:`1918` addressing may be
5446   configured with stub zones for ``10.in-addr.arpa`` to use a set of
5447   internal name servers as the authoritative servers for that domain.
5448
5449``static-stub``
5450   A static-stub zone is similar to a stub zone, with the following
5451   exceptions: the zone data is statically configured, rather than
5452   transferred from a primary server; and when recursion is necessary for a query
5453   that matches a static-stub zone, the locally configured data (name server
5454   names and glue addresses) is always used, even if different authoritative
5455   information is cached.
5456
5457   Zone data is configured via the ``server-addresses`` and ``server-names``
5458   zone options.
5459
5460   The zone data is maintained in the form of NS and (if necessary) glue A or
5461   AAAA RRs internally, which can be seen by dumping zone databases with
5462   ``rndc dumpdb -all``. The configured RRs are considered local configuration
5463   parameters rather than public data. Non-recursive queries (i.e., those
5464   with the RD bit off) to a static-stub zone are therefore prohibited and
5465   are responded to with REFUSED.
5466
5467   Since the data is statically configured, no zone maintenance action takes
5468   place for a static-stub zone. For example, there is no periodic refresh
5469   attempt, and an incoming notify message is rejected with an rcode
5470   of NOTAUTH.
5471
5472   Each static-stub zone is configured with internally generated NS and (if
5473   necessary) glue A or AAAA RRs.
5474
5475``forward``
5476   A forward zone is a way to configure forwarding on a per-domain basis.
5477   A ``zone`` statement of type ``forward`` can contain a ``forward`` and/or
5478   ``forwarders`` statement, which applies to queries within the domain
5479   given by the zone name. If no ``forwarders`` statement is present, or an
5480   empty list for ``forwarders`` is given, then no forwarding is done
5481   for the domain, canceling the effects of any forwarders in the ``options``
5482   statement. Thus, to use this type of zone to change the
5483   behavior of the global ``forward`` option (that is, "forward first" to,
5484   then "forward only", or vice versa), but use the same servers as set
5485   globally, re-specify the global forwarders.
5486
5487``redirect``
5488   Redirect zones are used to provide answers to queries when normal
5489   resolution would result in NXDOMAIN being returned. Only one redirect zone
5490   is supported per view. ``allow-query`` can be used to restrict which
5491   clients see these answers.
5492
5493   If the client has requested DNSSEC records (DO=1) and the NXDOMAIN response
5494   is signed, no substitution occurs.
5495
5496   To redirect all NXDOMAIN responses to 100.100.100.2 and
5497   2001:ffff:ffff::100.100.100.2, configure a type ``redirect`` zone
5498   named ".", with the zone file containing wildcard records that point to
5499   the desired addresses: ``*. IN A 100.100.100.2`` and
5500   ``*. IN AAAA 2001:ffff:ffff::100.100.100.2``.
5501
5502   As another example, to redirect all Spanish names (under .ES), use similar entries
5503   but with the names ``*.ES.`` instead of ``*.``. To redirect all commercial
5504   Spanish names (under COM.ES), use wildcard entries
5505   called ``*.COM.ES.``.
5506
5507   Note that the redirect zone supports all possible types; it is not
5508   limited to A and AAAA records.
5509
5510   If a redirect zone is configured with a ``primaries`` option, then it is
5511   transferred in as if it were a secondary zone. Otherwise, it is loaded from a
5512   file as if it were a primary zone.
5513
5514   Because redirect zones are not referenced directly by name, they are not
5515   kept in the zone lookup table with normal primary and secondary zones. To reload
5516   a redirect zone, use ``rndc reload -redirect``; to retransfer a
5517   redirect zone configured as a secondary, use ``rndc retransfer -redirect``.
5518   When using ``rndc reload`` without specifying a zone name, redirect
5519   zones are reloaded along with other zones.
5520
5521``delegation-only``
5522   This zone type is used to enforce the delegation-only status of infrastructure
5523   zones (e.g., COM, NET, ORG). Any answer that is received without an
5524   explicit or implicit delegation in the authority section is treated
5525   as NXDOMAIN. This does not apply to the zone apex, and should not be
5526   applied to leaf zones.
5527
5528   ``delegation-only`` has no effect on answers received from forwarders.
5529
5530   See caveats in :ref:`root-delegation-only <root-delegation-only>`.
5531
5532``in-view``
5533   When using multiple views, a ``primary`` or ``secondary`` zone configured
5534   in one view can be referenced in a subsequent view. This allows both views
5535   to use the same zone without the overhead of loading it more than once. This
5536   is configured using a ``zone`` statement, with an ``in-view`` option
5537   specifying the view in which the zone is defined. A ``zone`` statement
5538   containing ``in-view`` does not need to specify a type, since that is part
5539   of the zone definition in the other view.
5540
5541   See :ref:`multiple_views` for more information.
5542
5543Class
5544^^^^^
5545
5546The zone's name may optionally be followed by a class. If a class is not
5547specified, class ``IN`` (for ``Internet``) is assumed. This is correct
5548for the vast majority of cases.
5549
5550The ``hesiod`` class is named for an information service from MIT's
5551Project Athena. It was used to share information about various systems
5552databases, such as users, groups, printers, and so on. The keyword ``HS``
5553is a synonym for hesiod.
5554
5555Another MIT development is Chaosnet, a LAN protocol created in the
5556mid-1970s. Zone data for it can be specified with the ``CHAOS`` class.
5557
5558.. _zone_options:
5559
5560Zone Options
5561^^^^^^^^^^^^
5562
5563``allow-notify``
5564   See the description of ``allow-notify`` in :ref:`access_control`.
5565
5566``allow-query``
5567   See the description of ``allow-query`` in :ref:`access_control`.
5568
5569``allow-query-on``
5570   See the description of ``allow-query-on`` in :ref:`access_control`.
5571
5572``allow-transfer``
5573   See the description of ``allow-transfer`` in :ref:`access_control`.
5574
5575``allow-update``
5576   See the description of ``allow-update`` in :ref:`access_control`.
5577
5578``update-policy``
5579   This specifies a "Simple Secure Update" policy. See :ref:`dynamic_update_policies`.
5580
5581``allow-update-forwarding``
5582   See the description of ``allow-update-forwarding`` in :ref:`access_control`.
5583
5584``also-notify``
5585   This option is only meaningful if ``notify`` is active for this zone. The set of
5586   machines that receive a ``DNS NOTIFY`` message for this zone is
5587   made up of all the listed name servers (other than the primary)
5588   for the zone, plus any IP addresses specified with
5589   ``also-notify``. A port may be specified with each ``also-notify``
5590   address to send the notify messages to a port other than the default
5591   of 53. A TSIG key may also be specified to cause the ``NOTIFY`` to be
5592   signed by the given key. ``also-notify`` is not meaningful for stub
5593   zones. The default is the empty list.
5594
5595``check-names``
5596   This option is used to restrict the character set and syntax of
5597   certain domain names in primary files and/or DNS responses received
5598   from the network. The default varies according to zone type. For
5599   ``primary`` zones the default is ``fail``; for ``secondary`` zones the
5600   default is ``warn``. It is not implemented for ``hint`` zones.
5601
5602``check-mx``
5603   See the description of ``check-mx`` in :ref:`boolean_options`.
5604
5605``check-spf``
5606   See the description of ``check-spf`` in :ref:`boolean_options`.
5607
5608``check-wildcard``
5609   See the description of ``check-wildcard`` in :ref:`boolean_options`.
5610
5611``check-integrity``
5612   See the description of ``check-integrity`` in :ref:`boolean_options`.
5613
5614``check-sibling``
5615   See the description of ``check-sibling`` in :ref:`boolean_options`.
5616
5617``zero-no-soa-ttl``
5618   See the description of ``zero-no-soa-ttl`` in :ref:`boolean_options`.
5619
5620``update-check-ksk``
5621   See the description of ``update-check-ksk`` in :ref:`boolean_options`.
5622
5623``dnssec-loadkeys-interval``
5624   See the description of ``dnssec-loadkeys-interval`` in :ref:`options`.
5625
5626``dnssec-update-mode``
5627   See the description of ``dnssec-update-mode`` in :ref:`options`.
5628
5629``dnssec-dnskey-kskonly``
5630   See the description of ``dnssec-dnskey-kskonly`` in :ref:`boolean_options`.
5631
5632``try-tcp-refresh``
5633   See the description of ``try-tcp-refresh`` in :ref:`boolean_options`.
5634
5635``database``
5636   This specifies the type of database to be used to store the zone data.
5637   The string following the ``database`` keyword is interpreted as a
5638   list of whitespace-delimited words. The first word identifies the
5639   database type, and any subsequent words are passed as arguments to
5640   the database to be interpreted in a way specific to the database
5641   type.
5642
5643   The default is ``rbt``, BIND 9's native in-memory red-black tree
5644   database. This database does not take arguments.
5645
5646   Other values are possible if additional database drivers have been
5647   linked into the server. Some sample drivers are included with the
5648   distribution but none are linked in by default.
5649
5650``dialup``
5651   See the description of ``dialup`` in :ref:`boolean_options`.
5652
5653``delegation-only``
5654   This flag only applies to forward, hint, and stub zones. If set to
5655   ``yes``, then the zone is treated as if it is also a
5656   delegation-only type zone.
5657
5658   See caveats in :ref:`root-delegation-only <root-delegation-only>`.
5659
5660.. _file-option:
5661
5662``file``
5663   This sets the zone's filename. In ``primary``, ``hint``, and ``redirect``
5664   zones which do not have ``primaries`` defined, zone data is loaded from
5665   this file. In ``secondary``, ``mirror``, ``stub``, and ``redirect`` zones
5666   which do have ``primaries`` defined, zone data is retrieved from
5667   another server and saved in this file. This option is not applicable
5668   to other zone types.
5669
5670``forward``
5671   This option is only meaningful if the zone has a forwarders list. The ``only`` value
5672   causes the lookup to fail after trying the forwarders and getting no
5673   answer, while ``first`` allows a normal lookup to be tried.
5674
5675``forwarders``
5676   This is used to override the list of global forwarders. If it is not
5677   specified in a zone of type ``forward``, no forwarding is done for
5678   the zone and the global options are not used.
5679
5680``journal``
5681   This allows the default journal's filename to be overridden. The default is
5682   the zone's filename with "``.jnl``" appended. This is applicable to
5683   ``primary`` and ``secondary`` zones.
5684
5685``max-ixfr-ratio``
5686   See the description of ``max-ixfr-ratio`` in :ref:`options`.
5687
5688``max-journal-size``
5689   See the description of ``max-journal-size`` in :ref:`server_resource_limits`.
5690
5691``max-records``
5692   See the description of ``max-records`` in :ref:`server_resource_limits`.
5693
5694``max-transfer-time-in``
5695   See the description of ``max-transfer-time-in`` in :ref:`zone_transfers`.
5696
5697``max-transfer-idle-in``
5698   See the description of ``max-transfer-idle-in`` in :ref:`zone_transfers`.
5699
5700``max-transfer-time-out``
5701   See the description of ``max-transfer-time-out`` in :ref:`zone_transfers`.
5702
5703``max-transfer-idle-out``
5704   See the description of ``max-transfer-idle-out`` in :ref:`zone_transfers`.
5705
5706``notify``
5707   See the description of ``notify`` in :ref:`boolean_options`.
5708
5709``notify-delay``
5710   See the description of ``notify-delay`` in :ref:`tuning`.
5711
5712``notify-to-soa``
5713   See the description of ``notify-to-soa`` in :ref:`boolean_options`.
5714
5715``zone-statistics``
5716   See the description of ``zone-statistics`` in :ref:`options`.
5717
5718``server-addresses``
5719   This option is only meaningful for static-stub zones. This is a list of IP addresses
5720   to which queries should be sent in recursive resolution for the zone.
5721   A non-empty list for this option internally configures the apex
5722   NS RR with associated glue A or AAAA RRs.
5723
5724   For example, if "example.com" is configured as a static-stub zone
5725   with 192.0.2.1 and 2001:db8::1234 in a ``server-addresses`` option,
5726   the following RRs are internally configured:
5727
5728   ::
5729
5730      example.com. NS example.com.
5731      example.com. A 192.0.2.1
5732      example.com. AAAA 2001:db8::1234
5733
5734   These records are used internally to resolve names under the
5735   static-stub zone. For instance, if the server receives a query for
5736   "www.example.com" with the RD bit on, the server initiates
5737   recursive resolution and sends queries to 192.0.2.1 and/or
5738   2001:db8::1234.
5739
5740``server-names``
5741   This option is only meaningful for static-stub zones. This is a list of domain names
5742   of name servers that act as authoritative servers of the static-stub
5743   zone. These names are resolved to IP addresses when ``named``
5744   needs to send queries to these servers. For this supplemental
5745   resolution to be successful, these names must not be a subdomain of the
5746   origin name of the static-stub zone. That is, when "example.net" is the
5747   origin of a static-stub zone, "ns.example" and "master.example.com"
5748   can be specified in the ``server-names`` option, but "ns.example.net"
5749   cannot; it is rejected by the configuration parser.
5750
5751   A non-empty list for this option internally configures the apex
5752   NS RR with the specified names. For example, if "example.com" is
5753   configured as a static-stub zone with "ns1.example.net" and
5754   "ns2.example.net" in a ``server-names`` option, the following RRs
5755   are internally configured:
5756
5757   ::
5758
5759      example.com. NS ns1.example.net.
5760      example.com. NS ns2.example.net.
5761
5762   These records are used internally to resolve names under the
5763   static-stub zone. For instance, if the server receives a query for
5764   "www.example.com" with the RD bit on, the server initiates recursive
5765   resolution, resolves "ns1.example.net" and/or "ns2.example.net" to IP
5766   addresses, and then sends queries to one or more of these addresses.
5767
5768``sig-validity-interval``
5769   See the description of ``sig-validity-interval`` in :ref:`tuning`.
5770
5771``sig-signing-nodes``
5772   See the description of ``sig-signing-nodes`` in :ref:`tuning`.
5773
5774``sig-signing-signatures``
5775   See the description of ``sig-signing-signatures`` in
5776   :ref:`tuning`.
5777
5778``sig-signing-type``
5779   See the description of ``sig-signing-type`` in :ref:`tuning`.
5780
5781``transfer-source``
5782   See the description of ``transfer-source`` in :ref:`zone_transfers`.
5783
5784``transfer-source-v6``
5785   See the description of ``transfer-source-v6`` in :ref:`zone_transfers`.
5786
5787``alt-transfer-source``
5788   See the description of ``alt-transfer-source`` in :ref:`zone_transfers`.
5789
5790``alt-transfer-source-v6``
5791   See the description of ``alt-transfer-source-v6`` in :ref:`zone_transfers`.
5792
5793``use-alt-transfer-source``
5794   See the description of ``use-alt-transfer-source`` in :ref:`zone_transfers`.
5795
5796``notify-source``
5797   See the description of ``notify-source`` in :ref:`zone_transfers`.
5798
5799``notify-source-v6``
5800   See the description of ``notify-source-v6`` in :ref:`zone_transfers`.
5801
5802``min-refresh-time``; ``max-refresh-time``; ``min-retry-time``; ``max-retry-time``
5803   See the descriptions in :ref:`tuning`.
5804
5805``ixfr-from-differences``
5806   See the description of ``ixfr-from-differences`` in :ref:`boolean_options`.
5807   (Note that the ``ixfr-from-differences`` choices of ``primary`` and ``secondary``
5808   are not available at the zone level.)
5809
5810``key-directory``
5811   See the description of ``key-directory`` in :ref:`options`.
5812
5813``auto-dnssec``
5814   See the description of ``auto-dnssec`` in :ref:`options`.
5815
5816``serial-update-method``
5817   See the description of ``serial-update-method`` in :ref:`options`.
5818
5819``inline-signing``
5820   If ``yes``, BIND 9 maintains a separate signed version of the zone.
5821   An unsigned zone is transferred in or loaded from disk and the signed
5822   version of the zone is served with, possibly, a different serial
5823   number. The signed version of the zone is stored in a file that is
5824   the zone's filename (set in ``file``) with a ``.signed`` extension.
5825   This behavior is disabled by default.
5826
5827``multi-master``
5828   See the description of ``multi-master`` in :ref:`boolean_options`.
5829
5830``masterfile-format``
5831   See the description of ``masterfile-format`` in :ref:`tuning`.
5832
5833``max-zone-ttl``
5834   See the description of ``max-zone-ttl`` in :ref:`options`.
5835
5836``dnssec-secure-to-insecure``
5837   See the description of ``dnssec-secure-to-insecure`` in :ref:`boolean_options`.
5838
5839.. _dynamic_update_policies:
5840
5841Dynamic Update Policies
5842^^^^^^^^^^^^^^^^^^^^^^^
5843
5844BIND 9 supports two methods of granting clients the right to
5845perform dynamic updates to a zone:
5846
5847- ``allow-update`` - a simple access control list
5848- ``update-policy`` - fine-grained access control
5849
5850In both cases, BIND 9 writes the updates to the zone's filename
5851set in ``file``.
5852
5853In the case of a DNSSEC zone, DNSSEC records are also written to
5854the zone's filename, unless ``inline-signing`` is enabled.
5855
5856   .. note:: The zone file can no longer be manually updated while ``named``
5857      is running; it is now necessary to perform :option:`rndc freeze`, edit,
5858      and then perform :option:`rndc thaw`. Comments and formatting
5859      in the zone file are lost when dynamic updates occur.
5860
5861The ``allow-update`` clause is a simple access control list. Any client
5862that matches the ACL is granted permission to update any record in the
5863zone.
5864
5865The ``update-policy`` clause allows more fine-grained control over which
5866updates are allowed. It specifies a set of rules, in which each rule
5867either grants or denies permission for one or more names in the zone to
5868be updated by one or more identities. Identity is determined by the key
5869that signed the update request, using either TSIG or SIG(0). In most
5870cases, ``update-policy`` rules only apply to key-based identities. There
5871is no way to specify update permissions based on the client source address.
5872
5873``update-policy`` rules are only meaningful for zones of type
5874``primary``, and are not allowed in any other zone type. It is a
5875configuration error to specify both ``allow-update`` and
5876``update-policy`` at the same time.
5877
5878A pre-defined ``update-policy`` rule can be switched on with the command
5879``update-policy local;``. ``named`` automatically
5880generates a TSIG session key when starting and stores it in a file;
5881this key can then be used by local clients to update the zone while
5882``named`` is running. By default, the session key is stored in the file
5883``/var/run/named/session.key``, the key name is "local-ddns", and the
5884key algorithm is HMAC-SHA256. These values are configurable with the
5885``session-keyfile``, ``session-keyname``, and ``session-keyalg`` options,
5886respectively. A client running on the local system, if run with
5887appropriate permissions, may read the session key from the key file and
5888use it to sign update requests. The zone's update policy is set to
5889allow that key to change any record within the zone. Assuming the key
5890name is "local-ddns", this policy is equivalent to:
5891
5892::
5893
5894   update-policy { grant local-ddns zonesub any; };
5895
5896with the additional restriction that only clients connecting from the
5897local system are permitted to send updates.
5898
5899Note that only one session key is generated by ``named``; all zones
5900configured to use ``update-policy local`` accept the same key.
5901
5902The command ``nsupdate -l`` implements this feature, sending requests to
5903localhost and signing them using the key retrieved from the session key
5904file.
5905
5906Other rule definitions look like this:
5907
5908::
5909
5910   ( grant | deny ) identity ruletype  name   types
5911
5912Each rule grants or denies privileges. Rules are checked in the order in
5913which they are specified in the ``update-policy`` statement. Once a
5914message has successfully matched a rule, the operation is immediately
5915granted or denied, and no further rules are examined. There are 13 types
5916of rules; the rule type is specified by the ``ruletype`` field, and the
5917interpretation of other fields varies depending on the rule type.
5918
5919In general, a rule is matched when the key that signed an update request
5920matches the ``identity`` field, the name of the record to be updated
5921matches the ``name`` field (in the manner specified by the ``ruletype``
5922field), and the type of the record to be updated matches the ``types``
5923field. Details for each rule type are described below.
5924
5925The ``identity`` field must be set to a fully qualified domain name. In
5926most cases, this represents the name of the TSIG or SIG(0) key that
5927must be used to sign the update request. If the specified name is a
5928wildcard, it is subject to DNS wildcard expansion, and the rule may
5929apply to multiple identities. When a TKEY exchange has been used to
5930create a shared secret, the identity of the key used to authenticate the
5931TKEY exchange is used as the identity of the shared secret. Some
5932rule types use identities matching the client's Kerberos principal (e.g,
5933``"host/machine@REALM"``) or Windows realm (``machine$@REALM``).
5934
5935The ``name`` field also specifies a fully qualified domain name. This often
5936represents the name of the record to be updated. Interpretation of this
5937field is dependent on rule type.
5938
5939If no ``types`` are explicitly specified, then a rule matches all types
5940except RRSIG, NS, SOA, NSEC, and NSEC3. Types may be specified by name,
5941including ``ANY``; ANY matches all types except NSEC and NSEC3, which can
5942never be updated. Note that when an attempt is made to delete all
5943records associated with a name, the rules are checked for each existing
5944record type.
5945
5946The ruletype field has 16 values: ``name``, ``subdomain``, ``zonesub``, ``wildcard``,
5947``self``, ``selfsub``, ``selfwild``, ``ms-self``, ``ms-selfsub``, ``ms-subdomain``,
5948``krb5-self``, ``krb5-selfsub``, ``krb5-subdomain``,
5949``tcp-self``, ``6to4-self``, and ``external``.
5950
5951``name``
5952    With exact-match semantics, this rule matches when the name being updated is identical to the contents of the ``name`` field.
5953
5954``subdomain``
5955    This rule matches when the name being updated is a subdomain of, or identical to, the contents of the ``name`` field.
5956
5957``zonesub``
5958    This rule is similar to subdomain, except that it matches when the name being updated is a subdomain of the zone in which the ``update-policy`` statement appears. This obviates the need to type the zone name twice, and enables the use of a standard ``update-policy`` statement in multiple zones without modification.
5959    When this rule is used, the ``name`` field is omitted.
5960
5961``wildcard``
5962    The ``name`` field is subject to DNS wildcard expansion, and this rule matches when the name being updated is a valid expansion of the wildcard.
5963
5964``self``
5965    This rule matches when the name of the record being updated matches the contents of the ``identity`` field. The ``name`` field is ignored. To avoid confusion, it is recommended that this field be set to the same value as the ``identity`` field or to "."
5966    The ``self`` rule type is most useful when allowing one key per name to update, where the key has the same name as the record to be updated. In this case, the ``identity`` field can be specified as ``*`` (asterisk).
5967
5968``selfsub``
5969    This rule is similar to ``self``, except that subdomains of ``self`` can also be updated.
5970
5971``selfwild``
5972    This rule is similar to ``self``, except that only subdomains of ``self`` can be updated.
5973
5974``ms-self``
5975    When a client sends an UPDATE using a Windows machine principal (for example, ``machine$@REALM``), this rule allows records with the absolute name of ``machine.REALM`` to be updated.
5976
5977    The realm to be matched is specified in the ``identity`` field.
5978
5979    The ``name`` field has no effect on this rule; it should be set to "." as a placeholder.
5980
5981    For example, ``grant EXAMPLE.COM ms-self . A AAAA`` allows any machine with a valid principal in the realm ``EXAMPLE.COM`` to update its own address records.
5982
5983``ms-selfsub``
5984    This is similar to ``ms-self``, except it also allows updates to any subdomain of the name specified in the Windows machine principal, not just to the name itself.
5985
5986``ms-subdomain``
5987    When a client sends an UPDATE using a Windows machine principal (for example, ``machine$@REALM``), this rule allows any machine in the specified realm to update any record in the zone or in a specified subdomain of the zone.
5988
5989    The realm to be matched is specified in the ``identity`` field.
5990
5991    The ``name`` field specifies the subdomain that may be updated. If set to "." or any other name at or above the zone apex, any name in the zone can be updated.
5992
5993    For example, if ``update-policy`` for the zone "example.com" includes ``grant EXAMPLE.COM ms-subdomain hosts.example.com. AA AAAA``, any machine with a valid principal in the realm ``EXAMPLE.COM`` is able to update address records at or below ``hosts.example.com``.
5994
5995``krb5-self``
5996    When a client sends an UPDATE using a Kerberos machine principal (for example, ``host/machine@REALM``), this rule allows records with the absolute name of ``machine`` to be updated, provided it has been authenticated by REALM. This is similar but not identical to ``ms-self``, due to the ``machine`` part of the Kerberos principal being an absolute name instead of an unqualified name.
5997
5998    The realm to be matched is specified in the ``identity`` field.
5999
6000    The ``name`` field has no effect on this rule; it should be set to "." as a placeholder.
6001
6002    For example, ``grant EXAMPLE.COM krb5-self . A AAAA`` allows any machine with a valid principal in the realm ``EXAMPLE.COM`` to update its own address records.
6003
6004``krb5-selfsub``
6005    This is similar to ``krb5-self``, except it also allows updates to any subdomain of the name specified in the ``machine`` part of the Kerberos principal, not just to the name itself.
6006
6007``krb5-subdomain``
6008    This rule is identical to ``ms-subdomain``, except that it works with Kerberos machine principals (i.e., ``host/machine@REALM``) rather than Windows machine principals.
6009
6010``tcp-self``
6011    This rule allows updates that have been sent via TCP and for which the standard mapping from the client's IP address into the ``in-addr.arpa`` and ``ip6.arpa`` namespaces matches the name to be updated. The ``identity`` field must match that name. The ``name`` field should be set to ".". Note that, since identity is based on the client's IP address, it is not necessary for update request messages to be signed.
6012
6013    .. note::
6014        It is theoretically possible to spoof these TCP sessions.
6015
6016``6to4-self``
6017    This allows the name matching a 6to4 IPv6 prefix, as specified in :rfc:`3056`, to be updated by any TCP connection from either the 6to4 network or from the corresponding IPv4 address. This is intended to allow NS or DNAME RRsets to be added to the ``ip6.arpa`` reverse tree.
6018
6019    The ``identity`` field must match the 6to4 prefix in ``ip6.arpa``. The ``name`` field should be set to ".". Note that, since identity is based on the client's IP address, it is not necessary for update request messages to be signed.
6020
6021    In addition, if specified for an ``ip6.arpa`` name outside of the ``2.0.0.2.ip6.arpa`` namespace, the corresponding /48 reverse name can be updated. For example, TCP/IPv6 connections from 2001:DB8:ED0C::/48 can update records at ``C.0.D.E.8.B.D.0.1.0.0.2.ip6.arpa``.
6022
6023    .. note::
6024        It is theoretically possible to spoof these TCP sessions.
6025
6026``external``
6027    This rule allows ``named`` to defer the decision of whether to allow a given update to an external daemon.
6028
6029    The method of communicating with the daemon is specified in the ``identity`` field, the format of which is "``local:``\ path", where "path" is the location of a Unix-domain socket. (Currently, "local" is the only supported mechanism.)
6030
6031    Requests to the external daemon are sent over the Unix-domain socket as datagrams with the following format:
6032
6033    ::
6034
6035        Protocol version number (4 bytes, network byte order, currently 1)
6036        Request length (4 bytes, network byte order)
6037        Signer (null-terminated string)
6038        Name (null-terminated string)
6039        TCP source address (null-terminated string)
6040        Rdata type (null-terminated string)
6041        Key (null-terminated string)
6042        TKEY token length (4 bytes, network byte order)
6043        TKEY token (remainder of packet)
6044
6045    The daemon replies with a four-byte value in network byte order, containing either 0 or 1; 0 indicates that the specified update is not permitted, and 1 indicates that it is.
6046
6047       .. warning:: The external daemon must not delay communication. This policy is evaluated synchronously; any wait period negatively affects :iscman:`named` performance.
6048
6049.. _multiple_views:
6050
6051Multiple Views
6052^^^^^^^^^^^^^^
6053
6054When multiple views are in use, a zone may be referenced by more than
6055one of them. Often, the views contain different zones with the same
6056name, allowing different clients to receive different answers for the
6057same queries. At times, however, it is desirable for multiple views to
6058contain identical zones. The ``in-view`` zone option provides an
6059efficient way to do this; it allows a view to reference a zone that was
6060defined in a previously configured view. For example:
6061
6062::
6063
6064   view internal {
6065       match-clients { 10/8; };
6066
6067       zone example.com {
6068       type primary;
6069       file "example-external.db";
6070       };
6071   };
6072
6073   view external {
6074       match-clients { any; };
6075
6076       zone example.com {
6077       in-view internal;
6078       };
6079   };
6080
6081An ``in-view`` option cannot refer to a view that is configured later in
6082the configuration file.
6083
6084A ``zone`` statement which uses the ``in-view`` option may not use any
6085other options, with the exception of ``forward`` and ``forwarders``.
6086(These options control the behavior of the containing view, rather than
6087change the zone object itself.)
6088
6089Zone-level ACLs (e.g., allow-query, allow-transfer), and other
6090configuration details of the zone, are all set in the view the referenced
6091zone is defined in. Be careful to ensure that ACLs are wide
6092enough for all views referencing the zone.
6093
6094An ``in-view`` zone cannot be used as a response policy zone.
6095
6096An ``in-view`` zone is not intended to reference a ``forward`` zone.
6097
6098.. _zone_file:
6099
6100Zone File
6101---------
6102
6103.. _types_of_resource_records_and_when_to_use_them:
6104
6105Types of Resource Records and When to Use Them
6106~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
6107
6108This section, largely borrowed from :rfc:`1034`, describes the concept of a
6109Resource Record (RR) and explains when each type is used. Since the
6110publication of :rfc:`1034`, several new RRs have been identified and
6111implemented in the DNS. These are also included.
6112
6113Resource Records
6114^^^^^^^^^^^^^^^^
6115
6116A domain name identifies a node. Each node has a set of resource
6117information, which may be empty. The set of resource information
6118associated with a particular name is composed of separate RRs. The order
6119of RRs in a set is not significant and need not be preserved by name
6120servers, resolvers, or other parts of the DNS. However, sorting of
6121multiple RRs is permitted for optimization purposes: for example, to
6122specify that a particular nearby server be tried first. See
6123:ref:`the_sortlist_statement` and :ref:`rrset_ordering`.
6124
6125The components of a Resource Record are:
6126
6127owner name
6128    The domain name where the RR is found.
6129
6130type
6131    An encoded 16-bit value that specifies the type of the resource record.
6132
6133TTL
6134    The time-to-live of the RR. This field is a 32-bit integer in units of seconds, and is primarily used by resolvers when they cache RRs. The TTL describes how long a RR can be cached before it should be discarded.
6135
6136class
6137    An encoded 16-bit value that identifies a protocol family or an instance of a protocol.
6138
6139RDATA
6140    The resource data. The format of the data is type- and sometimes class-specific.
6141
6142For a complete list of *types* of valid RRs, including those that have been obsoleted, please refer to https://en.wikipedia.org/wiki/List_of_DNS_record_types.
6143
6144The following *classes* of resource records are currently valid in the
6145DNS:
6146
6147IN
6148    The Internet.
6149
6150CH
6151    Chaosnet, a LAN protocol created at MIT in the mid-1970s. It was rarely used for its historical purpose, but was reused for BIND's built-in server information zones, e.g., ``version.bind``.
6152
6153HS
6154    Hesiod, an information service developed by MIT's Project Athena. It was used to share information about various systems databases, such as users, groups, printers, etc.
6155
6156The owner name is often implicit, rather than forming an integral part
6157of the RR. For example, many name servers internally form tree or hash
6158structures for the name space, and chain RRs off nodes. The remaining RR
6159parts are the fixed header (type, class, TTL), which is consistent for
6160all RRs, and a variable part (RDATA) that fits the needs of the resource
6161being described.
6162
6163The TTL field is a time limit on how long an RR can be
6164kept in a cache. This limit does not apply to authoritative data in
6165zones; that also times out, but follows the refreshing policies for the
6166zone. The TTL is assigned by the administrator for the zone where the
6167data originates. While short TTLs can be used to minimize caching, and a
6168zero TTL prohibits caching, the realities of Internet performance
6169suggest that these times should be on the order of days for the typical
6170host. If a change is anticipated, the TTL can be reduced prior to
6171the change to minimize inconsistency, and then
6172increased back to its former value following the change.
6173
6174The data in the RDATA section of RRs is carried as a combination of
6175binary strings and domain names. The domain names are frequently used as
6176"pointers" to other data in the DNS.
6177
6178.. _rr_text:
6179
6180Textual Expression of RRs
6181^^^^^^^^^^^^^^^^^^^^^^^^^
6182
6183RRs are represented in binary form in the packets of the DNS protocol,
6184and are usually represented in highly encoded form when stored in a name
6185server or resolver. In the examples provided in :rfc:`1034`, a style
6186similar to that used in primary files was employed in order to show the
6187contents of RRs. In this format, most RRs are shown on a single line,
6188although continuation lines are possible using parentheses.
6189
6190The start of the line gives the owner of the RR. If a line begins with a
6191blank, then the owner is assumed to be the same as that of the previous
6192RR. Blank lines are often included for readability.
6193
6194Following the owner are listed the TTL, type, and class of the RR. Class
6195and type use the mnemonics defined above, and TTL is an integer before
6196the type field. To avoid ambiguity in parsing, type and class
6197mnemonics are disjoint, TTLs are integers, and the type mnemonic is
6198always last. The IN class and TTL values are often omitted from examples
6199in the interest of clarity.
6200
6201The resource data or RDATA section of the RR is given using knowledge
6202of the typical representation for the data.
6203
6204For example, the RRs carried in a message might be shown as:
6205
6206 +---------------------+---------------+--------------------------------+
6207 | ``ISI.EDU.``        | ``MX``        | ``10 VENERA.ISI.EDU.``         |
6208 +---------------------+---------------+--------------------------------+
6209 |                     | ``MX``        | ``10 VAXA.ISI.EDU``            |
6210 +---------------------+---------------+--------------------------------+
6211 | ``VENERA.ISI.EDU``  | ``A``         | ``128.9.0.32``                 |
6212 +---------------------+---------------+--------------------------------+
6213 |                     | ``A``         | ``10.1.0.52``                  |
6214 +---------------------+---------------+--------------------------------+
6215 | ``VAXA.ISI.EDU``    | ``A``         | ``10.2.0.27``                  |
6216 +---------------------+---------------+--------------------------------+
6217 |                     | ``A``         | ``128.9.0.33``                 |
6218 +---------------------+---------------+--------------------------------+
6219
6220The MX RRs have an RDATA section which consists of a 16-bit number
6221followed by a domain name. The address RRs use a standard IP address
6222format to contain a 32-bit Internet address.
6223
6224The above example shows six RRs, with two RRs at each of three domain
6225names.
6226
6227Here is another possible example:
6228
6229 +----------------------+---------------+-------------------------------+
6230 | ``XX.LCS.MIT.EDU.``  | ``IN A``      | ``10.0.0.44``                 |
6231 +----------------------+---------------+-------------------------------+
6232 |                      | ``CH A``      | ``MIT.EDU. 2420``             |
6233 +----------------------+---------------+-------------------------------+
6234
6235This shows two addresses for ``XX.LCS.MIT.EDU``, each of a
6236different class.
6237
6238.. _mx_records:
6239
6240Discussion of MX Records
6241~~~~~~~~~~~~~~~~~~~~~~~~
6242
6243As described above, domain servers store information as a series of
6244resource records, each of which contains a particular piece of
6245information about a given domain name (which is usually, but not always,
6246a host). The simplest way to think of an RR is as a typed pair of data, a
6247domain name matched with a relevant datum and stored with some
6248additional type information, to help systems determine when the RR is
6249relevant.
6250
6251MX records are used to control delivery of email. The data specified in
6252the record is a priority and a domain name. The priority controls the
6253order in which email delivery is attempted, with the lowest number
6254first. If two priorities are the same, a server is chosen randomly. If
6255no servers at a given priority are responding, the mail transport agent
6256falls back to the next largest priority. Priority numbers do not
6257have any absolute meaning; they are relevant only respective to other
6258MX records for that domain name. The domain name given is the machine to
6259which the mail is delivered. It *must* have an associated address
6260record (A or AAAA); CNAME is not sufficient.
6261
6262For a given domain, if there is both a CNAME record and an MX record,
6263the MX record is in error and is ignored. Instead, the mail is
6264delivered to the server specified in the MX record pointed to by the
6265CNAME. For example:
6266
6267 +------------------------+--------+--------+--------------+------------------------+
6268 | ``example.com.``       | ``IN`` | ``MX`` | ``10``       | ``mail.example.com.``  |
6269 +------------------------+--------+--------+--------------+------------------------+
6270 |                        | ``IN`` | ``MX`` | ``10``       | ``mail2.example.com.`` |
6271 +------------------------+--------+--------+--------------+------------------------+
6272 |                        | ``IN`` | ``MX`` | ``20``       | ``mail.backup.org.``   |
6273 +------------------------+--------+--------+--------------+------------------------+
6274 | ``mail.example.com.``  | ``IN`` | ``A``  | ``10.0.0.1`` |                        |
6275 +------------------------+--------+--------+--------------+------------------------+
6276 | ``mail2.example.com.`` | ``IN`` | ``A``  | ``10.0.0.2`` |                        |
6277 +------------------------+--------+--------+--------------+------------------------+
6278
6279Mail delivery is attempted to ``mail.example.com`` and
6280``mail2.example.com`` (in any order); if neither of those succeeds,
6281delivery to ``mail.backup.org`` is attempted.
6282
6283.. _Setting_TTLs:
6284
6285Setting TTLs
6286~~~~~~~~~~~~
6287
6288The time-to-live (TTL) of the RR field is a 32-bit integer represented in
6289units of seconds, and is primarily used by resolvers when they cache
6290RRs. The TTL describes how long an RR can be cached before it should be
6291discarded. The following three types of TTLs are currently used in a zone
6292file.
6293
6294SOA
6295    The last field in the SOA is the negative caching TTL. This controls how long other servers cache no-such-domain (NXDOMAIN) responses from this server.
6296
6297    The maximum time for negative caching is 3 hours (3h).
6298
6299$TTL
6300    The $TTL directive at the top of the zone file (before the SOA) gives a default TTL for every RR without a specific TTL set.
6301
6302RR TTLs
6303    Each RR can have a TTL as the second field in the RR, which controls how long other servers can cache it.
6304
6305All of these TTLs default to units of seconds, though units can be
6306explicitly specified: for example, ``1h30m``.
6307
6308.. _ipv4_reverse:
6309
6310Inverse Mapping in IPv4
6311~~~~~~~~~~~~~~~~~~~~~~~
6312
6313Reverse name resolution (that is, translation from IP address to name)
6314is achieved by means of the ``in-addr.arpa`` domain and PTR records.
6315Entries in the in-addr.arpa domain are made in least-to-most significant
6316order, read left to right. This is the opposite order to the way IP
6317addresses are usually written. Thus, a machine with an IP address of
631810.1.2.3 would have a corresponding in-addr.arpa name of
63193.2.1.10.in-addr.arpa. This name should have a PTR resource record whose
6320data field is the name of the machine or, optionally, multiple PTR
6321records if the machine has more than one name. For example, in the
6322``example.com`` domain:
6323
6324 +--------------+-------------------------------------------------------+
6325 | ``$ORIGIN``  | ``2.1.10.in-addr.arpa``                               |
6326 +--------------+-------------------------------------------------------+
6327 | ``3``        | ``IN PTR foo.example.com.``                           |
6328 +--------------+-------------------------------------------------------+
6329
6330.. note::
6331
6332   The ``$ORIGIN`` line in this example is only to provide context;
6333   it does not necessarily appear in the actual
6334   usage. It is only used here to indicate that the example is
6335   relative to the listed origin.
6336
6337.. _zone_directives:
6338
6339Other Zone File Directives
6340~~~~~~~~~~~~~~~~~~~~~~~~~~
6341
6342The DNS "master file" format was initially defined in :rfc:`1035` and has
6343subsequently been extended. While the format itself is class-independent,
6344all records in a zone file must be of the same class.
6345
6346Master file directives include ``$ORIGIN``, ``$INCLUDE``, and ``$TTL.``
6347
6348.. _atsign:
6349
6350The ``@`` (at-sign)
6351^^^^^^^^^^^^^^^^^^^
6352
6353When used in the label (or name) field, the asperand or at-sign (@)
6354symbol represents the current origin. At the start of the zone file, it
6355is the <``zone_name``>, followed by a trailing dot (.).
6356
6357.. _origin_directive:
6358
6359The ``$ORIGIN`` Directive
6360^^^^^^^^^^^^^^^^^^^^^^^^^
6361
6362Syntax: ``$ORIGIN`` domain-name [comment]
6363
6364``$ORIGIN`` sets the domain name that is appended to any
6365unqualified records. When a zone is first read, there is an implicit
6366``$ORIGIN`` <``zone_name``>``.``; note the trailing dot. The
6367current ``$ORIGIN`` is appended to the domain specified in the
6368``$ORIGIN`` argument if it is not absolute.
6369
6370::
6371
6372   $ORIGIN example.com.
6373   WWW     CNAME   MAIN-SERVER
6374
6375is equivalent to
6376
6377::
6378
6379   WWW.EXAMPLE.COM. CNAME MAIN-SERVER.EXAMPLE.COM.
6380
6381.. _include_directive:
6382
6383The ``$INCLUDE`` Directive
6384^^^^^^^^^^^^^^^^^^^^^^^^^^
6385
6386Syntax: ``$INCLUDE`` filename [origin] [comment]
6387
6388This reads and processes the file ``filename`` as if it were included in the
6389file at this point. The ``filename`` can be an absolute path, or a relative
6390path. In the latter case it is read from ``named``'s working directory. If
6391``origin`` is specified, the file is processed with ``$ORIGIN`` set to that
6392value; otherwise, the current ``$ORIGIN`` is used.
6393
6394The origin and the current domain name revert to the values they had
6395prior to the ``$INCLUDE`` once the file has been read.
6396
6397.. note::
6398
6399   :rfc:`1035` specifies that the current origin should be restored after
6400   an ``$INCLUDE``, but it is silent on whether the current domain name
6401   should also be restored. BIND 9 restores both of them. This could be
6402   construed as a deviation from :rfc:`1035`, a feature, or both.
6403
6404.. _ttl_directive:
6405
6406The ``$TTL`` Directive
6407^^^^^^^^^^^^^^^^^^^^^^
6408
6409Syntax: ``$TTL`` default-ttl [comment]
6410
6411This sets the default Time-To-Live (TTL) for subsequent records with undefined
6412TTLs. Valid TTLs are of the range 0-2147483647 seconds.
6413
6414``$TTL`` is defined in :rfc:`2308`.
6415
6416.. _generate_directive:
6417
6418BIND Primary File Extension: the ``$GENERATE`` Directive
6419~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
6420
6421Syntax: ``$GENERATE`` range owner [ttl] [class] type rdata [comment]
6422
6423``$GENERATE`` is used to create a series of resource records that only
6424differ from each other by an iterator.
6425
6426``range``
6427    This can be one of two forms: start-stop or start-stop/step.
6428    If the first form is used, then step is set to 1. "start",
6429    "stop", and "step" must be positive integers between 0 and
6430    (2^31)-1. "start" must not be larger than "stop".
6431
6432``owner``
6433    This describes the owner name of the resource records to be created.
6434
6435    The ``owner`` string may include one or more ``$`` (dollar sign)
6436    symbols, which will be replaced with the iterator value when
6437    generating records; see below for details.
6438
6439``ttl``
6440    This specifies the time-to-live of the generated records. If
6441    not specified, this is inherited using the normal TTL inheritance
6442    rules.
6443
6444    ``class`` and ``ttl`` can be entered in either order.
6445
6446``class``
6447    This specifies the class of the generated records. This must
6448    match the zone class if it is specified.
6449
6450    ``class`` and ``ttl`` can be entered in either order.
6451
6452``type``
6453    This can be any valid type.
6454
6455``rdata``
6456    This is a string containing the RDATA of the resource record
6457    to be created. As with ``owner``, the ``rdata`` string may
6458    include one or more ``$`` symbols, which are replaced with the
6459    iterator value. ``rdata`` may be quoted if there are spaces in
6460    the string; the quotation marks do not appear in the generated
6461    record.
6462
6463    Any single ``$`` (dollar sign) symbols within the ``owner`` or
6464    ``rdata`` strings are replaced by the iterator value. To get a ``$``
6465    in the output, escape the ``$`` using a backslash ``\\``, e.g.,
6466    ``\$``. (For compatibility with earlier versions, ``$$`` is also
6467    recognized as indicating a literal ``$`` in the output.)
6468
6469    The ``$`` may optionally be followed by modifiers which change
6470    the offset from the iterator, field width, and base.  Modifiers
6471    are introduced by a ``{`` (left brace) immediately following
6472    the ``$``, as in  ``${offset[,width[,base]]}``. For example,
6473    ``${-20,3,d}`` subtracts 20 from the current value and prints
6474    the result as a decimal in a zero-padded field of width 3.
6475    Available output forms are decimal (``d``), octal (``o``),
6476    hexadecimal (``x`` or ``X`` for uppercase), and nibble (``n``
6477    or ``N`` for uppercase). The modfiier cannot contain whitespace
6478    or newlines.
6479
6480    The default modifier is ``${0,0,d}``. If the ``owner`` is not
6481    absolute, the current ``$ORIGIN`` is appended to the name.
6482
6483    In nibble mode, the value is treated as if it were a reversed
6484    hexadecimal string, with each hexadecimal digit as a separate
6485    label. The width field includes the label separator.
6486
6487Examples:
6488
6489``$GENERATE`` can be used to easily generate the sets of records required
6490to support sub-/24 reverse delegations described in :rfc:`2317`:
6491
6492::
6493
6494   $ORIGIN 0.0.192.IN-ADDR.ARPA.
6495   $GENERATE 1-2 @ NS SERVER$.EXAMPLE.
6496   $GENERATE 1-127 $ CNAME $.0
6497
6498is equivalent to
6499
6500::
6501
6502   0.0.0.192.IN-ADDR.ARPA. NS SERVER1.EXAMPLE.
6503   0.0.0.192.IN-ADDR.ARPA. NS SERVER2.EXAMPLE.
6504   1.0.0.192.IN-ADDR.ARPA. CNAME 1.0.0.0.192.IN-ADDR.ARPA.
6505   2.0.0.192.IN-ADDR.ARPA. CNAME 2.0.0.0.192.IN-ADDR.ARPA.
6506   ...
6507   127.0.0.192.IN-ADDR.ARPA. CNAME 127.0.0.0.192.IN-ADDR.ARPA.
6508
6509This example creates a set of A and MX records. Note the MX's ``rdata``
6510is a quoted string; the quotes are stripped when ``$GENERATE`` is processed:
6511
6512::
6513
6514   $ORIGIN EXAMPLE.
6515   $GENERATE 1-127 HOST-$ A 1.2.3.$
6516   $GENERATE 1-127 HOST-$ MX "0 ."
6517
6518is equivalent to
6519
6520::
6521
6522   HOST-1.EXAMPLE.   A  1.2.3.1
6523   HOST-1.EXAMPLE.   MX 0 .
6524   HOST-2.EXAMPLE.   A  1.2.3.2
6525   HOST-2.EXAMPLE.   MX 0 .
6526   HOST-3.EXAMPLE.   A  1.2.3.3
6527   HOST-3.EXAMPLE.   MX 0 .
6528   ...
6529   HOST-127.EXAMPLE. A  1.2.3.127
6530   HOST-127.EXAMPLE. MX 0 .
6531
6532
6533This example generates A and AAAA records using modifiers; the AAAA
6534``owner`` names are generated using nibble mode:
6535
6536::
6537
6538   $ORIGIN EXAMPLE.
6539   $GENERATE 0-2 HOST-${0,4,d} A 1.2.3.${1,0,d}
6540   $GENERATE 1024-1026 ${0,3,n} AAAA 2001:db8::${0,4,x}
6541
6542is equivalent to:
6543
6544::
6545   HOST-0000.EXAMPLE.   A      1.2.3.1
6546   HOST-0001.EXAMPLE.   A      1.2.3.2
6547   HOST-0002.EXAMPLE.   A      1.2.3.3
6548   0.0.4.EXAMPLE.       AAAA   2001:db8::400
6549   1.0.4.EXAMPLE.       AAAA   2001:db8::401
6550   2.0.4.EXAMPLE.       AAAA   2001:db8::402
6551
6552The ``$GENERATE`` directive is a BIND extension and not part of the
6553standard zone file format.
6554
6555.. _zonefile_format:
6556
6557Additional File Formats
6558~~~~~~~~~~~~~~~~~~~~~~~
6559
6560In addition to the standard text format, BIND 9 supports the ability
6561to read or dump to zone files in other formats.
6562
6563The ``raw`` format is a binary representation of zone data in a manner
6564similar to that used in zone transfers. Since it does not require
6565parsing text, load time is significantly reduced.
6566
6567An even faster alternative is the ``map`` format, which is an image of a
6568BIND 9 in-memory zone database; it can be loaded directly into memory via
6569the ``mmap()`` function and the zone can begin serving queries almost
6570immediately.  Because records are not indivdually processed when loading a
6571``map`` file, zones using this format cannot be used in ``response-policy``
6572statements.
6573
6574For a primary server, a zone file in ``raw`` or ``map`` format is expected
6575to be generated from a text zone file by the ``named-compilezone`` command.
6576For a secondary server or a dynamic zone, the zone file is automatically
6577generated when ``named`` dumps the zone contents after zone transfer or
6578when applying prior updates, if one of these formats is specified by the
6579``masterfile-format`` option.
6580
6581If a zone file in a binary format needs manual modification, it first must
6582be converted to ``text`` format by the ``named-compilezone`` command,
6583then converted back after editing.  For example:
6584
6585::
6586    named-compilezone -f map -F text -o zonefile.text <origin> zonefile.map
6587    [edit zonefile.text]
6588    named-compilezone -f text -F map -o zonefile.map <origin> zonefile.text
6589
6590Note that the ``map`` format is highly architecture-specific. A ``map``
6591file *cannot* be used on a system with different pointer size, endianness,
6592or data alignment than the system on which it was generated, and should in
6593general be used only inside a single system.
6594
6595The ``map`` format is also dependent on the internal memory representation
6596of a zone database, which may change from one release of BIND 9 to another.
6597``map`` files are never compatible across major releases, and may not be
6598compatible across minor releases; any upgrade to BIND 9 may cause ``map``
6599files to be rejected when loading. If a ``map`` file is being used for a
6600primary zone, it will need to be regenerated from text before restarting
6601the server.  If it used for a secondary zone, this is unnecessary; the
6602rejection of the file will trigger a retransfer of the zone from the
6603primary. (To avoid a spike in traffic upon restart, it may be desirable in
6604some cases to convert ``map`` files to ``text`` format using
6605``named-compilezone`` before an upgrade, then back to ``map`` format with
6606the new version of ``named-compilezone`` afterward.)
6607
6608The use of ``map`` format may also be limited by operating system
6609mmap(2) limits like ``sysctl vm.max_map_count``.  For Linux, this
6610defaults to 65536, which limits the number of mapped zones that can
6611be used without increasing ``vm.max_map_count``.
6612
6613``raw`` format uses network byte order and avoids architecture-
6614dependent data alignment so that it is as portable as possible, but it is
6615still primarily expected to be used inside the same single system. To
6616export a zone file in either ``raw`` or ``map`` format, or make a portable
6617backup of such a file, conversion to ``text`` format is recommended.
6618
6619.. _statistics:
6620
6621BIND 9 Statistics
6622-----------------
6623
6624BIND 9 maintains lots of statistics information and provides several
6625interfaces for users to access those statistics. The available
6626statistics include all statistics counters that are meaningful in BIND 9,
6627and other information that is considered useful.
6628
6629The statistics information is categorized into the following sections:
6630
6631Incoming Requests
6632   The number of incoming DNS requests for each OPCODE.
6633
6634Incoming Queries
6635   The number of incoming queries for each RR type.
6636
6637Outgoing Queries
6638   The number of outgoing queries for each RR type sent from the internal
6639   resolver, maintained per view.
6640
6641Name Server Statistics
6642   Statistics counters for incoming request processing.
6643
6644Zone Maintenance Statistics
6645   Statistics counters regarding zone maintenance operations, such as zone
6646   transfers.
6647
6648Resolver Statistics
6649   Statistics counters for name resolutions performed in the internal resolver,
6650   maintained per view.
6651
6652Cache DB RRsets
6653   Statistics counters related to cache contents, maintained per view.
6654
6655   The "NXDOMAIN" counter is the number of names that have been cached as
6656   nonexistent.  Counters named for RR types indicate the number of active
6657   RRsets for each type in the cache database.
6658
6659   If an RR type name is preceded by an exclamation point (!), it represents the
6660   number of records in the cache which indicate that the type does not exist
6661   for a particular name; this is also known as "NXRRSET". If an RR type name
6662   is preceded by a hash mark (#), it represents the number of RRsets for this
6663   type that are present in the cache but whose TTLs have expired; these RRsets
6664   may only be used if stale answers are enabled.  If an RR type name is
6665   preceded by a tilde (~), it represents the number of RRsets for this type
6666   that are present in the cache database but are marked for garbage collection;
6667   these RRsets cannot be used.
6668
6669Socket I/O Statistics
6670   Statistics counters for network-related events.
6671
6672A subset of Name Server Statistics is collected and shown per zone for
6673which the server has the authority, when ``zone-statistics`` is set to
6674``full`` (or ``yes``), for backward compatibility. See the description of
6675``zone-statistics`` in :ref:`options` for further details.
6676
6677These statistics counters are shown with their zone and view names. The
6678view name is omitted when the server is not configured with explicit
6679views.
6680
6681There are currently two user interfaces to get access to the statistics.
6682One is in plain-text format, dumped to the file specified by the
6683``statistics-file`` configuration option; the other is remotely
6684accessible via a statistics channel when the ``statistics-channels``
6685statement is specified in the configuration file (see :ref:`statschannels`.)
6686
6687.. _statsfile:
6688
6689The Statistics File
6690~~~~~~~~~~~~~~~~~~~
6691
6692The text format statistics dump begins with a line, like:
6693
6694``+++ Statistics Dump +++ (973798949)``
6695
6696The number in parentheses is a standard Unix-style timestamp, measured
6697in seconds since January 1, 1970. Following that line is a set of
6698statistics information, which is categorized as described above. Each
6699section begins with a line, like:
6700
6701``++ Name Server Statistics ++``
6702
6703Each section consists of lines, each containing the statistics counter
6704value followed by its textual description; see below for available
6705counters. For brevity, counters that have a value of 0 are not shown in
6706the statistics file.
6707
6708The statistics dump ends with the line where the number is identical to
6709the number in the beginning line; for example:
6710
6711``--- Statistics Dump --- (973798949)``
6712
6713.. _statistics_counters:
6714
6715Statistics Counters
6716~~~~~~~~~~~~~~~~~~~
6717
6718The following lists summarize the statistics counters that BIND 9 provides.
6719For each counter, the abbreviated
6720symbol name is given; these symbols are shown in the statistics
6721information accessed via an HTTP statistics channel.
6722The description of the counter is also shown in the
6723statistics file but, in this document, may be slightly
6724modified for better readability.
6725
6726.. _stats_counters:
6727
6728Name Server Statistics Counters
6729^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6730
6731``Requestv4``
6732    This indicates the number of IPv4 requests received. Note: this also counts non-query requests.
6733
6734``Requestv6``
6735    This indicates the number of IPv6 requests received. Note: this also counts non-query requests.
6736
6737``ReqEdns0``
6738    This indicates the number of requests received with EDNS(0).
6739
6740``ReqBadEDN SVer``
6741    This indicates the number of requests received with an unsupported EDNS version.
6742
6743``ReqTSIG``
6744    This indicates the number of requests received with TSIG.
6745
6746``ReqSIG0``
6747    This indicates the number of requests received with SIG(0).
6748
6749``ReqBadSIG``
6750    This indicates the number of requests received with an invalid (TSIG or SIG(0)) signature.
6751
6752``ReqTCP``
6753    This indicates the number of TCP requests received.
6754
6755``AuthQryRej``
6756    This indicates the number of rejected authoritative (non-recursive) queries.
6757
6758``RecQryRej``
6759    This indicates the number of rejected recursive queries.
6760
6761``XfrRej``
6762    This indicates the number of rejected zone transfer requests.
6763
6764``UpdateRej``
6765    This indicates the number of rejected dynamic update requests.
6766
6767``Response``
6768    This indicates the number of responses sent.
6769
6770``RespTruncated``
6771    This indicates the number of truncated responses sent.
6772
6773``RespEDNS0``
6774    This indicates the number of responses sent with EDNS(0).
6775
6776``RespTSIG``
6777    This indicates the number of responses sent with TSIG.
6778
6779``RespSIG0``
6780    This indicates the number of responses sent with SIG(0).
6781
6782``QrySuccess``
6783    This indicates the number of queries that resulted in a successful answer, meaning queries which return a NOERROR response with at least one answer RR. This corresponds to the ``success`` counter of previous versions of BIND 9.
6784
6785``QryAuthAns``
6786    This indicates the number of queries that resulted in an authoritative answer.
6787
6788``QryNoauthAns``
6789    This indicates the number of queries that resulted in a non-authoritative answer.
6790
6791``QryReferral``
6792    This indicates the number of queries that resulted in a referral answer. This corresponds to the ``referral`` counter of previous versions of BIND 9.
6793
6794``QryNxrrset``
6795    This indicates the number of queries that resulted in NOERROR responses with no data. This corresponds to the ``nxrrset`` counter of previous versions of BIND 9.
6796
6797``QrySERVFAIL``
6798    This indicates the number of queries that resulted in SERVFAIL.
6799
6800``QryFORMERR``
6801    This indicates the number of queries that resulted in FORMERR.
6802
6803``QryNXDOMAIN``
6804    This indicates the number of queries that resulted in NXDOMAIN. This corresponds to the ``nxdomain`` counter of previous versions of BIND 9.
6805
6806``QryRecursion``
6807    This indicates the number of queries that caused the server to perform recursion in order to find the final answer. This corresponds to the ``recursion`` counter of previous versions of BIND 9.
6808
6809``QryDuplicate``
6810    This indicates the number of queries which the server attempted to recurse but for which it discovered an existing query with the same IP address, port, query ID, name, type, and class already being processed. This corresponds to the ``duplicate`` counter of previous versions of BIND 9.
6811
6812``QryDropped``
6813    This indicates the number of recursive queries for which the server discovered an excessive number of existing recursive queries for the same name, type, and class, and which were subsequently dropped. This is the number of dropped queries due to the reason explained with the ``clients-per-query`` and ``max-clients-per-query`` options (see :ref:`clients-per-query <clients-per-query>`). This corresponds to the ``dropped`` counter of previous versions of BIND 9.
6814
6815``QryFailure``
6816    This indicates the number of query failures. This corresponds to the ``failure`` counter of previous versions of BIND 9. Note: this counter is provided mainly for backward compatibility with previous versions; normally, more fine-grained counters such as ``AuthQryRej`` and ``RecQryRej`` that would also fall into this counter are provided, so this counter is not of much interest in practice.
6817
6818``QryNXRedir``
6819    This indicates the number of queries that resulted in NXDOMAIN that were redirected.
6820
6821``QryNXRedirRLookup``
6822    This indicates the number of queries that resulted in NXDOMAIN that were redirected and resulted in a successful remote lookup.
6823
6824``XfrReqDone``
6825    This indicates the number of requested and completed zone transfers.
6826
6827``UpdateReqFwd``
6828    This indicates the number of forwarded update requests.
6829
6830``UpdateRespFwd``
6831    This indicates the number of forwarded update responses.
6832
6833``UpdateFwdFail``
6834    This indicates the number of forwarded dynamic updates that failed.
6835
6836``UpdateDone``
6837    This indicates the number of completed dynamic updates.
6838
6839``UpdateFail``
6840    This indicates the number of failed dynamic updates.
6841
6842``UpdateBadPrereq``
6843    This indicates the number of dynamic updates rejected due to a prerequisite failure.
6844
6845``UpdateQuota``
6846    This indicates the number of times a dynamic update or update
6847    forwarding request was rejected because the number of pending
6848    requests exceeded ``update-quota``.
6849
6850``RateDropped``
6851    This indicates the number of responses dropped due to rate limits.
6852
6853``RateSlipped``
6854    This indicates the number of responses truncated by rate limits.
6855
6856``RPZRewrites``
6857    This indicates the number of response policy zone rewrites.
6858
6859.. _zone_stats:
6860
6861Zone Maintenance Statistics Counters
6862^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6863
6864``NotifyOutv4``
6865    This indicates the number of IPv4 notifies sent.
6866
6867``NotifyOutv6``
6868    This indicates the number of IPv6 notifies sent.
6869
6870``NotifyInv4``
6871    This indicates the number of IPv4 notifies received.
6872
6873``NotifyInv6``
6874    This indicates the number of IPv6 notifies received.
6875
6876``NotifyRej``
6877    This indicates the number of incoming notifies rejected.
6878
6879``SOAOutv4``
6880    This indicates the number of IPv4 SOA queries sent.
6881
6882``SOAOutv6``
6883    This indicates the number of IPv6 SOA queries sent.
6884
6885``AXFRReqv4``
6886    This indicates the number of requested IPv4 AXFRs.
6887
6888``AXFRReqv6``
6889    This indicates the number of requested IPv6 AXFRs.
6890
6891``IXFRReqv4``
6892    This indicates the number of requested IPv4 IXFRs.
6893
6894``IXFRReqv6``
6895    This indicates the number of requested IPv6 IXFRs.
6896
6897``XfrSuccess``
6898    This indicates the number of successful zone transfer requests.
6899
6900``XfrFail``
6901    This indicates the number of failed zone transfer requests.
6902
6903.. _resolver_stats:
6904
6905Resolver Statistics Counters
6906^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6907
6908``Queryv4``
6909    This indicates the number of IPv4 queries sent.
6910
6911``Queryv6``
6912    This indicates the number of IPv6 queries sent.
6913
6914``Responsev4``
6915    This indicates the number of IPv4 responses received.
6916
6917``Responsev6``
6918    This indicates the number of IPv6 responses received.
6919
6920``NXDOMAIN``
6921    This indicates the number of NXDOMAINs received.
6922
6923``SERVFAIL``
6924    This indicates the number of SERVFAILs received.
6925
6926``FORMERR``
6927    This indicates the number of FORMERRs received.
6928
6929``OtherError``
6930    This indicates the number of other errors received.
6931
6932``EDNS0Fail``
6933    This indicates the number of EDNS(0) query failures.
6934
6935``Mismatch``
6936    This indicates the number of mismatched responses received, meaning the DNS ID, response's source address, and/or the response's source port does not match what was expected. (The port must be 53 or as defined by the ``port`` option.) This may be an indication of a cache poisoning attempt.
6937
6938``Truncated``
6939    This indicates the number of truncated responses received.
6940
6941``Lame``
6942    This indicates the number of lame delegations received.
6943
6944``Retry``
6945    This indicates the number of query retries performed.
6946
6947``QueryAbort``
6948    This indicates the number of queries aborted due to quota control.
6949
6950``QuerySockFail``
6951    This indicates the number of failures in opening query sockets. One common reason for such failures is due to a limitation on file descriptors.
6952
6953``QueryTimeout``
6954    This indicates the number of query timeouts.
6955
6956``GlueFetchv4``
6957    This indicates the number of IPv4 NS address fetches invoked.
6958
6959``GlueFetchv6``
6960    This indicates the number of IPv6 NS address fetches invoked.
6961
6962``GlueFetchv4Fail``
6963    This indicates the number of failed IPv4 NS address fetches.
6964
6965``GlueFetchv6Fail``
6966    This indicates the number of failed IPv6 NS address fetches.
6967
6968``ValAttempt``
6969    This indicates the number of attempted DNSSEC validations.
6970
6971``ValOk``
6972    This indicates the number of successful DNSSEC validations.
6973
6974``ValNegOk``
6975    This indicates the number of successful DNSSEC validations on negative information.
6976
6977``ValFail``
6978    This indicates the number of failed DNSSEC validations.
6979
6980``QryRTTnn``
6981    This provides a frequency table on query round-trip times (RTTs). Each ``nn`` specifies the corresponding frequency. In the sequence of ``nn_1``, ``nn_2``, ..., ``nn_m``, the value of ``nn_i`` is the number of queries whose RTTs are between ``nn_(i-1)`` (inclusive) and ``nn_i`` (exclusive) milliseconds. For the sake of convenience, we define ``nn_0`` to be 0. The last entry should be represented as ``nn_m+``, which means the number of queries whose RTTs are equal to or greater than ``nn_m`` milliseconds.
6982
6983.. _socket_stats:
6984
6985Socket I/O Statistics Counters
6986^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6987
6988Socket I/O statistics counters are defined per socket type, which are
6989``UDP4`` (UDP/IPv4), ``UDP6`` (UDP/IPv6), ``TCP4`` (TCP/IPv4), ``TCP6``
6990(TCP/IPv6), ``Unix`` (Unix Domain), and ``FDwatch`` (sockets opened
6991outside the socket module). In the following list, ``<TYPE>`` represents
6992a socket type. Not all counters are available for all socket types;
6993exceptions are noted in the descriptions.
6994
6995``<TYPE>Open``
6996    This indicates the number of sockets opened successfully. This counter does not apply to the ``FDwatch`` type.
6997
6998``<TYPE>OpenFail``
6999    This indicates the number of failures to open sockets. This counter does not apply to the ``FDwatch`` type.
7000
7001``<TYPE>Close``
7002    This indicates the number of closed sockets.
7003
7004``<TYPE>BindFail``
7005    This indicates the number of failures to bind sockets.
7006
7007``<TYPE>ConnFail``
7008    This indicates the number of failures to connect sockets.
7009
7010``<TYPE>Conn``
7011    This indicates the number of connections established successfully.
7012
7013``<TYPE>AcceptFail``
7014    This indicates the number of failures to accept incoming connection requests. This counter does not apply to the ``UDP`` and ``FDwatch`` types.
7015
7016``<TYPE>Accept``
7017    This indicates the number of incoming connections successfully accepted. This counter does not apply to the ``UDP`` and ``FDwatch`` types.
7018
7019``<TYPE>SendErr``
7020    This indicates the number of errors in socket send operations.
7021
7022``<TYPE>RecvErr``
7023    This indicates the number of errors in socket receive operations, including errors of send operations on a connected UDP socket, notified by an ICMP error message.
7024