| blob | b29de7df6dc03ed11acf278b335e6edb194009e7 |
1 <!DOCTYPE article PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
2 "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd" []>
3 <!--
4 - Copyright (C) 2004-2009 Internet Systems Consortium, Inc. ("ISC")
5 - Copyright (C) 2000-2003 Internet Software Consortium.
6 -
7 - Permission to use, copy, modify, and/or distribute this software for any
8 - purpose with or without fee is hereby granted, provided that the above
9 - copyright notice and this permission notice appear in all copies.
10 -
11 - THE SOFTWARE IS PROVIDED "AS IS" AND ISC DISCLAIMS ALL WARRANTIES WITH
12 - REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY
13 - AND FITNESS. IN NO EVENT SHALL ISC BE LIABLE FOR ANY SPECIAL, DIRECT,
14 - INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM
15 - LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE
16 - OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
17 - PERFORMANCE OF THIS SOFTWARE.
18 -->
20 <!-- Id: FAQ.xml,v 1.52 2009/11/03 14:02:20 marka Exp -->
22 <article class="faq">
23 <title>Frequently Asked Questions about BIND 9</title>
24 <articleinfo>
25 <copyright>
26 <year>2004</year>
27 <year>2005</year>
28 <year>2006</year>
29 <year>2007</year>
30 <year>2008</year>
31 <year>2009</year>
32 <holder>Internet Systems Consortium, Inc. ("ISC")</holder>
33 </copyright>
34 <copyright>
35 <year>2000</year>
36 <year>2001</year>
37 <year>2002</year>
38 <year>2003</year>
39 <holder>Internet Software Consortium.</holder>
40 </copyright>
41 </articleinfo>
42 <qandaset defaultlabel='qanda'>
44 <qandadiv><title>Compilation and Installation Questions</title>
46 <qandaentry>
47 <question>
48 <para>
49 I'm trying to compile BIND 9, and "make" is failing due to
50 files not being found. Why?
51 </para>
52 </question>
53 <answer>
54 <para>
55 Using a parallel or distributed "make" to build BIND 9 is
56 not supported, and doesn't work. If you are using one of
57 these, use normal make or gmake instead.
58 </para>
59 </answer>
60 </qandaentry>
62 <qandaentry>
63 <question>
64 <para>
65 Isn't "make install" supposed to generate a default named.conf?
66 </para>
67 </question>
68 <answer>
69 <para>
70 Short Answer: No.
71 </para>
72 <para>
73 Long Answer: There really isn't a default configuration which fits
74 any site perfectly. There are lots of decisions that need to
75 be made and there is no consensus on what the defaults should be.
76 For example FreeBSD uses /etc/namedb as the location where the
77 configuration files for named are stored. Others use /var/named.
78 </para>
79 <para>
80 What addresses to listen on? For a laptop on the move a lot
81 you may only want to listen on the loop back interfaces.
82 </para>
83 <para>
84 Who do you offer recursive service to? Is there are firewall
85 to consider? If so is it stateless or stateful. Are you
86 directly on the Internet? Are you on a private network? Are
87 you on a NAT'd network? The answers
88 to all these questions change how you configure even a
89 caching name server.
90 </para>
91 </answer>
92 </qandaentry>
94 </qandadiv> <!-- Compilation and Installation Questions -->
96 <qandadiv><title>Configuration and Setup Questions</title>
98 <qandaentry>
99 <!-- configuration, log -->
100 <question>
101 <para>
102 Why does named log the warning message <quote>no TTL specified -
103 using SOA MINTTL instead</quote>?
104 </para>
105 </question>
106 <answer>
107 <para>
108 Your zone file is illegal according to RFC1035. It must either
109 have a line like:
110 </para>
111 <informalexample>
112 <programlisting>
113 $TTL 86400</programlisting>
114 </informalexample>
115 <para>
116 at the beginning, or the first record in it must have a TTL field,
117 like the "84600" in this example:
118 </para>
119 <informalexample>
120 <programlisting>
121 example.com. 86400 IN SOA ns hostmaster ( 1 3600 1800 1814400 3600 )</programlisting>
122 </informalexample>
123 </answer>
124 </qandaentry>
126 <qandaentry>
127 <!-- configuration -->
128 <question>
129 <para>
130 Why do I get errors like <quote>dns_zone_load: zone foo/IN: loading
131 master file bar: ran out of space</quote>?
132 </para>
133 </question>
134 <answer>
135 <para>
136 This is often caused by TXT records with missing close
137 quotes. Check that all TXT records containing quoted strings
138 have both open and close quotes.
139 </para>
140 </answer>
141 </qandaentry>
143 <qandaentry>
144 <!-- security -->
145 <question>
146 <para>
147 How do I restrict people from looking up the server version?
148 </para>
149 </question>
150 <answer>
151 <para>
152 Put a "version" option containing something other than the
153 real version in the "options" section of named.conf. Note
154 doing this will not prevent attacks and may impede people
155 trying to diagnose problems with your server. Also it is
156 possible to "fingerprint" nameservers to determine their
157 version.
158 </para>
159 </answer>
160 </qandaentry>
162 <qandaentry>
163 <!-- security -->
164 <question>
165 <para>
166 How do I restrict only remote users from looking up the
167 server version?
168 </para>
169 </question>
170 <answer>
171 <para>
172 The following view statement will intercept lookups as the
173 internal view that holds the version information will be
174 matched last. The caveats of the previous answer still
175 apply, of course.
176 </para>
177 <informalexample>
178 <programlisting>
179 view "chaos" chaos {
180 match-clients { <those to be refused>; };
181 allow-query { none; };
182 zone "." {
183 type hint;
184 file "/dev/null"; // or any empty file
185 };
186 };</programlisting>
187 </informalexample>
188 </answer>
189 </qandaentry>
191 <qandaentry>
192 <!-- configuration -->
193 <question>
194 <para>
195 What do <quote>no source of entropy found</quote> or <quote>could not
196 open entropy source foo</quote> mean?
197 </para>
198 </question>
199 <answer>
200 <para>
201 The server requires a source of entropy to perform certain
202 operations, mostly DNSSEC related. These messages indicate
203 that you have no source of entropy. On systems with
204 /dev/random or an equivalent, it is used by default. A
205 source of entropy can also be defined using the random-device
206 option in named.conf.
207 </para>
208 </answer>
209 </qandaentry>
211 <qandaentry>
212 <!-- configuration -->
213 <question>
214 <para>
215 I'm trying to use TSIG to authenticate dynamic updates or
216 zone transfers. I'm sure I have the keys set up correctly,
217 but the server is rejecting the TSIG. Why?
218 </para>
219 </question>
220 <answer>
221 <para>
222 This may be a clock skew problem. Check that the the clocks
223 on the client and server are properly synchronised (e.g.,
224 using ntp).
225 </para>
226 </answer>
227 </qandaentry>
229 <qandaentry>
230 <question>
231 <para>
232 I see a log message like the following. Why?
233 </para>
234 <para>
235 couldn't open pid file '/var/run/named.pid': Permission denied
236 </para>
237 </question>
238 <answer>
239 <para>
240 You are most likely running named as a non-root user, and
241 that user does not have permission to write in /var/run.
242 The common ways of fixing this are to create a /var/run/named
243 directory owned by the named user and set pid-file to
244 "/var/run/named/named.pid", or set pid-file to "named.pid",
245 which will put the file in the directory specified by the
246 directory option (which, in this case, must be writable by
247 the named user).
248 </para>
249 </answer>
250 </qandaentry>
252 <qandaentry>
253 <question>
254 <para>
255 I can query the nameserver from the nameserver but not from other
256 machines. Why?
257 </para>
258 </question>
259 <answer>
260 <para>
261 This is usually the result of the firewall configuration stopping
262 the queries and / or the replies.
263 </para>
264 </answer>
265 </qandaentry>
267 <qandaentry>
268 <question>
269 <para>
270 How can I make a server a slave for both an internal and
271 an external view at the same time? When I tried, both views
272 on the slave were transferred from the same view on the master.
273 </para>
274 </question>
275 <answer>
276 <para>
277 You will need to give the master and slave multiple IP
278 addresses and use those to make sure you reach the correct
279 view on the other machine.
280 </para>
281 <informalexample>
282 <programlisting>
283 Master: 10.0.1.1 (internal), 10.0.1.2 (external, IP alias)
284 internal:
285 match-clients { !10.0.1.2; !10.0.1.4; 10.0.1/24; };
286 notify-source 10.0.1.1;
287 transfer-source 10.0.1.1;
288 query-source address 10.0.1.1;
289 external:
290 match-clients { any; };
291 recursion no; // don't offer recursion to the world
292 notify-source 10.0.1.2;
293 transfer-source 10.0.1.2;
294 query-source address 10.0.1.2;
296 Slave: 10.0.1.3 (internal), 10.0.1.4 (external, IP alias)
297 internal:
298 match-clients { !10.0.1.2; !10.0.1.4; 10.0.1/24; };
299 notify-source 10.0.1.3;
300 transfer-source 10.0.1.3;
301 query-source address 10.0.1.3;
302 external:
303 match-clients { any; };
304 recursion no; // don't offer recursion to the world
305 notify-source 10.0.1.4;
306 transfer-source 10.0.1.4;
307 query-source address 10.0.1.4;</programlisting>
308 </informalexample>
309 <para>
310 You put the external address on the alias so that all the other
311 dns clients on these boxes see the internal view by default.
312 </para>
313 </answer>
314 <answer>
315 <para>
316 BIND 9.3 and later: Use TSIG to select the appropriate view.
317 </para>
318 <informalexample>
319 <programlisting>
320 Master 10.0.1.1:
321 key "external" {
322 algorithm hmac-sha256;
323 secret "xxxxxxxxxxxxxxxxxxxxxxxx";
324 };
325 view "internal" {
326 match-clients { !key external; // reject message ment for the
327 // external view.
328 10.0.1/24; }; // accept from these addresses.
329 ...
330 };
331 view "external" {
332 match-clients { key external; any; };
333 server 10.0.1.2 { keys external; }; // tag messages from the
334 // external view to the
335 // other servers for the
336 // view.
337 recursion no;
338 ...
339 };
341 Slave 10.0.1.2:
342 key "external" {
343 algorithm hmac-sha256;
344 secret "xxxxxxxxxxxxxxxxxxxxxxxx";
345 };
346 view "internal" {
347 match-clients { !key external; 10.0.1/24; };
348 ...
349 };
350 view "external" {
351 match-clients { key external; any; };
352 server 10.0.1.1 { keys external; };
353 recursion no;
354 ...
355 };</programlisting>
356 </informalexample>
357 </answer>
358 </qandaentry>
360 <qandaentry>
361 <question>
362 <para>
363 I get error messages like <quote>multiple RRs of singleton type</quote>
364 and <quote>CNAME and other data</quote> when transferring a zone. What
365 does this mean?
366 </para>
367 </question>
368 <answer>
369 <para>
370 These indicate a malformed master zone. You can identify
371 the exact records involved by transferring the zone using
372 dig then running named-checkzone on it.
373 </para>
374 <informalexample>
375 <programlisting>
376 dig axfr example.com @master-server > tmp
377 named-checkzone example.com tmp</programlisting>
378 </informalexample>
379 <para>
380 A CNAME record cannot exist with the same name as another record
381 except for the DNSSEC records which prove its existence (NSEC).
382 </para>
383 <para>
384 RFC 1034, Section 3.6.2: <quote>If a CNAME RR is present at a node,
385 no other data should be present; this ensures that the data for a
386 canonical name and its aliases cannot be different. This rule also
387 insures that a cached CNAME can be used without checking with an
388 authoritative server for other RR types.</quote>
389 </para>
390 </answer>
391 </qandaentry>
393 <qandaentry>
394 <question>
395 <para>
396 I get error messages like <quote>named.conf:99: unexpected end
397 of input</quote> where 99 is the last line of named.conf.
398 </para>
399 </question>
400 <answer>
401 <para>
402 There are unbalanced quotes in named.conf.
403 </para>
404 </answer>
405 <answer>
406 <para>
407 Some text editors (notepad and wordpad) fail to put a line
408 title indication (e.g. CR/LF) on the last line of a
409 text file. This can be fixed by "adding" a blank line to
410 the end of the file. Named expects to see EOF immediately
411 after EOL and treats text files where this is not met as
412 truncated.
413 </para>
414 </answer>
415 </qandaentry>
417 <qandaentry>
418 <question>
419 <para>
420 How do I share a dynamic zone between multiple views?
421 </para>
422 </question>
423 <answer>
424 <para>
425 You choose one view to be master and the second a slave and
426 transfer the zone between views.
427 </para>
428 <informalexample>
429 <programlisting>
430 Master 10.0.1.1:
431 key "external" {
432 algorithm hmac-sha256;
433 secret "xxxxxxxxxxxxxxxxxxxxxxxx";
434 };
436 key "mykey" {
437 algorithm hmac-sha256;
438 secret "yyyyyyyyyyyyyyyyyyyyyyyy";
439 };
441 view "internal" {
442 match-clients { !key external; 10.0.1/24; };
443 server 10.0.1.1 {
444 /* Deliver notify messages to external view. */
445 keys { external; };
446 };
447 zone "example.com" {
448 type master;
449 file "internal/example.db";
450 allow-update { key mykey; };
451 also-notify { 10.0.1.1; };
452 };
453 };
455 view "external" {
456 match-clients { key external; any; };
457 zone "example.com" {
458 type slave;
459 file "external/example.db";
460 masters { 10.0.1.1; };
461 transfer-source 10.0.1.1;
462 // allow-update-forwarding { any; };
463 // allow-notify { ... };
464 };
465 };</programlisting>
466 </informalexample>
467 </answer>
468 </qandaentry>
470 <qandaentry>
471 <question>
472 <para>
473 I get a error message like <quote>zone wireless.ietf56.ietf.org/IN:
474 loading master file primaries/wireless.ietf56.ietf.org: no
475 owner</quote>.
476 </para>
477 </question>
478 <answer>
479 <para>
480 This error is produced when a line in the master file
481 contains leading white space (tab/space) but the is no
482 current record owner name to inherit the name from. Usually
483 this is the result of putting white space before a comment,
484 forgetting the "@" for the SOA record, or indenting the master
485 file.
486 </para>
487 </answer>
488 </qandaentry>
490 <qandaentry>
491 <question>
492 <para>
493 Why are my logs in GMT (UTC).
494 </para>
495 </question>
496 <answer>
497 <para>
498 You are running chrooted (-t) and have not supplied local timezone
499 information in the chroot area.
500 </para>
501 <simplelist>
502 <member>FreeBSD: /etc/localtime</member>
503 <member>Solaris: /etc/TIMEZONE and /usr/share/lib/zoneinfo</member>
504 <member>OSF: /etc/zoneinfo/localtime</member>
505 </simplelist>
506 <para>
507 See also tzset(3) and zic(8).
508 </para>
509 </answer>
510 </qandaentry>
512 <qandaentry>
513 <question>
514 <para>
515 I get <quote>rndc: connect failed: connection refused</quote> when
516 I try to run rndc.
517 </para>
518 </question>
519 <answer>
520 <para>
521 This is usually a configuration error.
522 </para>
523 <para>
524 First ensure that named is running and no errors are being
525 reported at startup (/var/log/messages or equivalent).
526 Running "named -g <usual arguments>" from a title
527 can help at this point.
528 </para>
529 <para>
530 Secondly ensure that named is configured to use rndc either
531 by "rndc-confgen -a", rndc-confgen or manually. The
532 Administrators Reference manual has details on how to do
533 this.
534 </para>
535 <para>
536 Old versions of rndc-confgen used localhost rather than
537 127.0.0.1 in /etc/rndc.conf for the default server. Update
538 /etc/rndc.conf if necessary so that the default server
539 listed in /etc/rndc.conf matches the addresses used in
540 named.conf. "localhost" has two address (127.0.0.1 and
541 ::1).
542 </para>
543 <para>
544 If you use "rndc-confgen -a" and named is running with -t or -u
545 ensure that /etc/rndc.conf has the correct ownership and that
546 a copy is in the chroot area. You can do this by re-running
547 "rndc-confgen -a" with appropriate -t and -u arguments.
548 </para>
549 </answer>
550 </qandaentry>
552 <qandaentry>
553 <question>
554 <para>
555 I get <quote>transfer of 'example.net/IN' from 192.168.4.12#53:
556 failed while receiving responses: permission denied</quote> error
557 messages.
558 </para>
559 </question>
560 <answer>
561 <para>
562 These indicate a filesystem permission error preventing
563 named creating / renaming the temporary file. These will
564 usually also have other associated error messages like
565 </para>
566 <informalexample>
567 <programlisting>
568 "dumping master file: sl/tmp-XXXX5il3sQ: open: permission denied"</programlisting>
569 </informalexample>
570 <para>
571 Named needs write permission on the directory containing
572 the file. Named writes the new cache file to a temporary
573 file then renames it to the name specified in named.conf
574 to ensure that the contents are always complete. This is
575 to prevent named loading a partial zone in the event of
576 power failure or similar interrupting the write of the
577 master file.
578 </para>
579 <para>
580 Note file names are relative to the directory specified in
581 options and any chroot directory ([<chroot
582 dir>/][<options dir>]).
583 </para>
584 <informalexample>
585 <para>
586 If named is invoked as "named -t /chroot/DNS" with
587 the following named.conf then "/chroot/DNS/var/named/sl"
588 needs to be writable by the user named is running as.
589 </para>
590 <programlisting>
591 options {
592 directory "/var/named";
593 };
595 zone "example.net" {
596 type slave;
597 file "sl/example.net";
598 masters { 192.168.4.12; };
599 };</programlisting>
600 </informalexample>
601 </answer>
602 </qandaentry>
604 <qandaentry>
605 <question>
606 <para>
607 I want to forward all DNS queries from my caching nameserver to
608 another server. But there are some domains which have to be
609 served locally, via rbldnsd.
610 </para>
611 <para>
612 How do I achieve this ?
613 </para>
614 </question>
615 <answer>
616 <programlisting>
617 options {
618 forward only;
619 forwarders { <ip.of.primary.nameserver>; };
620 };
622 zone "sbl-xbl.spamhaus.org" {
623 type forward; forward only;
624 forwarders { <ip.of.rbldns.server> port 530; };
625 };
627 zone "list.dsbl.org" {
628 type forward; forward only;
629 forwarders { <ip.of.rbldns.server> port 530; };
630 };
631 </programlisting>
632 </answer>
633 </qandaentry>
635 <qandaentry>
636 <question>
637 <para>
638 Can you help me understand how BIND 9 uses memory to store
639 DNS zones?
640 </para>
641 <para>
642 Some times it seems to take several times the amount of
643 memory it needs to store the zone.
644 </para>
645 </question>
646 <answer>
647 <para>
648 When reloading a zone named my have multiple copies of
649 the zone in memory at one time. The zone it is serving
650 and the one it is loading. If reloads are ultra fast it
651 can have more still.
652 </para>
653 <para>
654 e.g. Ones that are transferring out, the one that it is
655 serving and the one that is loading.
656 </para>
657 <para>
658 BIND 8 destroyed the zone before loading and also killed
659 off outgoing transfers of the zone.
660 </para>
661 <para>
662 The new strategy allows slaves to get copies of the new
663 zone regardless of how often the master is loaded compared
664 to the transfer time. The slave might skip some intermediate
665 versions but the transfers will complete and it will keep
666 reasonably in sync with the master.
667 </para>
668 <para>
669 The new strategy also allows the master to recover from
670 syntax and other errors in the master file as it still
671 has an in-core copy of the old contents.
672 </para>
673 </answer>
674 </qandaentry>
676 <qandaentry>
677 <question>
678 <para>
679 I want to use IPv6 locally but I don't have a external IPv6
680 connection. External lookups are slow.
681 </para>
682 </question>
683 <answer>
684 <para>
685 You can use server clauses to stop named making external lookups
686 over IPv6.
687 </para>
688 <programlisting>
689 server fd81:ec6c:bd62::/48 { bogus no; }; // site ULA prefix
690 server ::/0 { bogus yes; };
691 </programlisting>
692 </answer>
693 </qandaentry>
695 </qandadiv> <!-- Configuration and Setup Questions -->
697 <qandadiv><title>Operations Questions</title>
699 <qandaentry>
700 <question>
701 <para>
702 How to change the nameservers for a zone?
703 </para>
704 </question>
705 <answer>
706 <para>
707 Step 1: Ensure all nameservers, new and old, are serving the
708 same zone content.
709 </para>
710 <para>
711 Step 2: Work out the maximum TTL of the NS RRset in the parent and child
712 zones. This is the time it will take caches to be clear of a
713 particular version of the NS RRset.
714 If you are just removing nameservers you can skip to Step 6.
715 </para>
716 <para>
717 Step 3: Add new nameservers to the NS RRset for the zone and
718 wait until all the servers for the zone are answering with this
719 new NS RRset.
720 </para>
721 <para>
722 Step 4: Inform the parent zone of the new NS RRset then wait for all the
723 parent servers to be answering with the new NS RRset.
724 </para>
725 <para>
726 Step 5: Wait for cache to be clear of the old NS RRset.
727 See Step 2 for how long.
728 If you are just adding nameservers you are done.
729 </para>
730 <para>
731 Step 6: Remove any old nameservers from the zones NS RRset and
732 wait for all the servers for the zone to be serving the new NS RRset.
733 </para>
734 <para>
735 Step 7: Inform the parent zone of the new NS RRset then wait for all the
736 parent servers to be answering with the new NS RRset.
737 </para>
738 <para>
739 Step 8: Wait for cache to be clear of the old NS RRset.
740 See Step 2 for how long.
741 </para>
742 <para>
743 Step 9: Turn off the old nameservers or remove the zone entry from
744 the configuration of the old nameservers.
745 </para>
746 <para>
747 Step 10: Increment the serial number and wait for the change to
748 be visible in all nameservers for the zone. This ensures that
749 zone transfers are still working after the old servers are
750 decommissioned.
751 </para>
752 <para>
753 Note: the above procedure is designed to be transparent
754 to dns clients. Decommissioning the old servers too early
755 will result in some clients not being able to look up
756 answers in the zone.
757 </para>
758 <para>
759 Note: while it is possible to run the addition and removal
760 stages together it is not recommended.
761 </para>
762 </answer>
763 </qandaentry>
765 </qandadiv> <!-- Operations Questions -->
767 <qandadiv><title>General Questions</title>
769 <qandaentry>
770 <question>
771 <para>
772 I keep getting log messages like the following. Why?
773 </para>
774 <para>
775 Dec 4 23:47:59 client 10.0.0.1#1355: updating zone
776 'example.com/IN': update failed: 'RRset exists (value
777 dependent)' prerequisite not satisfied (NXRRSET)
778 </para>
779 </question>
780 <answer>
781 <para>
782 DNS updates allow the update request to test to see if
783 certain conditions are met prior to proceeding with the
784 update. The message above is saying that conditions were
785 not met and the update is not proceeding. See doc/rfc/rfc2136.txt
786 for more details on prerequisites.
787 </para>
788 </answer>
789 </qandaentry>
791 <qandaentry>
792 <question>
793 <para>
794 I keep getting log messages like the following. Why?
795 </para>
796 <para>
797 Jun 21 12:00:00.000 client 10.0.0.1#1234: update denied
798 </para>
799 </question>
800 <answer>
801 <para>
802 Someone is trying to update your DNS data using the RFC2136
803 Dynamic Update protocol. Windows 2000 machines have a habit
804 of sending dynamic update requests to DNS servers without
805 being specifically configured to do so. If the update
806 requests are coming from a Windows 2000 machine, see
807 <ulink
808 url="http://support.microsoft.com/support/kb/articles/q246/8/04.asp">
809 <http://support.microsoft.com/support/kb/articles/q246/8/04.asp></ulink>
810 for information about how to turn them off.
811 </para>
812 </answer>
813 </qandaentry>
815 <qandaentry>
816 <question>
817 <para>
818 When I do a "dig . ns", many of the A records for the root
819 servers are missing. Why?
820 </para>
821 </question>
822 <answer>
823 <para>
824 This is normal and harmless. It is a somewhat confusing
825 side effect of the way BIND 9 does RFC2181 trust ranking
826 and of the efforts BIND 9 makes to avoid promoting glue
827 into answers.
828 </para>
829 <para>
830 When BIND 9 first starts up and primes its cache, it receives
831 the root server addresses as additional data in an authoritative
832 response from a root server, and these records are eligible
833 for inclusion as additional data in responses. Subsequently
834 it receives a subset of the root server addresses as
835 additional data in a non-authoritative (referral) response
836 from a root server. This causes the addresses to now be
837 considered non-authoritative (glue) data, which is not
838 eligible for inclusion in responses.
839 </para>
840 <para>
841 The server does have a complete set of root server addresses
842 cached at all times, it just may not include all of them
843 as additional data, depending on whether they were last
844 received as answers or as glue. You can always look up the
845 addresses with explicit queries like "dig a.root-servers.net A".
846 </para>
847 </answer>
848 </qandaentry>
850 <qandaentry>
851 <question>
852 <para>
853 Why don't my zones reload when I do an "rndc reload" or SIGHUP?
854 </para>
855 </question>
856 <answer>
857 <para>
858 A zone can be updated either by editing zone files and
859 reloading the server or by dynamic update, but not both.
860 If you have enabled dynamic update for a zone using the
861 "allow-update" option, you are not supposed to edit the
862 zone file by hand, and the server will not attempt to reload
863 it.
864 </para>
865 </answer>
866 </qandaentry>
868 <qandaentry>
869 <question>
870 <para>
871 Why is named listening on UDP port other than 53?
872 </para>
873 </question>
874 <answer>
875 <para>
876 Named uses a system selected port to make queries of other
877 nameservers. This behaviour can be overridden by using
878 query-source to lock down the port and/or address. See
879 also notify-source and transfer-source.
880 </para>
881 </answer>
882 </qandaentry>
884 <qandaentry>
885 <question>
886 <para>
887 I get warning messages like <quote>zone example.com/IN: refresh:
888 failure trying master 1.2.3.4#53: timed out</quote>.
889 </para>
890 </question>
891 <answer>
892 <para>
893 Check that you can make UDP queries from the slave to the master
894 </para>
895 <informalexample>
896 <programlisting>
897 dig +norec example.com soa @1.2.3.4</programlisting>
898 </informalexample>
899 <para>
900 You could be generating queries faster than the slave can
901 cope with. Lower the serial query rate.
902 </para>
903 <informalexample>
904 <programlisting>
905 serial-query-rate 5; // default 20</programlisting>
906 </informalexample>
907 </answer>
908 </qandaentry>
910 <qandaentry>
911 <question>
912 <para>
913 I don't get RRSIG's returned when I use "dig +dnssec".
914 </para>
915 </question>
916 <answer>
917 <para>
918 You need to ensure DNSSEC is enabled (dnssec-enable yes;).
919 </para>
920 </answer>
921 </qandaentry>
923 <qandaentry>
924 <question>
925 <para>
926 Can a NS record refer to a CNAME.
927 </para>
928 </question>
929 <answer>
930 <para>
931 No. The rules for glue (copies of the *address* records
932 in the parent zones) and additional section processing do
933 not allow it to work.
934 </para>
935 <para>
936 You would have to add both the CNAME and address records
937 (A/AAAA) as glue to the parent zone and have CNAMEs be
938 followed when doing additional section processing to make
939 it work. No nameserver implementation supports either of
940 these requirements.
941 </para>
942 </answer>
943 </qandaentry>
945 <qandaentry>
946 <question>
947 <para>
948 What does <quote>RFC 1918 response from Internet for
949 0.0.0.10.IN-ADDR.ARPA</quote> mean?
950 </para>
951 </question>
952 <answer>
953 <para>
954 If the IN-ADDR.ARPA name covered refers to a internal address
955 space you are using then you have failed to follow RFC 1918
956 usage rules and are leaking queries to the Internet. You
957 should establish your own zones for these addresses to prevent
958 you querying the Internet's name servers for these addresses.
959 Please see <ulink url="http://as112.net/"><http://as112.net/></ulink>
960 for details of the problems you are causing and the counter
961 measures that have had to be deployed.
962 </para>
963 <para>
964 If you are not using these private addresses then a client
965 has queried for them. You can just ignore the messages,
966 get the offending client to stop sending you these messages
967 as they are most probably leaking them or setup your own zones
968 empty zones to serve answers to these queries.
969 </para>
970 <informalexample>
971 <programlisting>
972 zone "10.IN-ADDR.ARPA" {
973 type master;
974 file "empty";
975 };
977 zone "16.172.IN-ADDR.ARPA" {
978 type master;
979 file "empty";
980 };
982 ...
984 zone "31.172.IN-ADDR.ARPA" {
985 type master;
986 file "empty";
987 };
989 zone "168.192.IN-ADDR.ARPA" {
990 type master;
991 file "empty";
992 };
994 empty:
995 @ 10800 IN SOA <name-of-server>. <contact-email>. (
996 1 3600 1200 604800 10800 )
997 @ 10800 IN NS <name-of-server>.</programlisting>
998 </informalexample>
999 <para>
1000 <note>
1001 Future versions of named are likely to do this automatically.
1002 </note>
1003 </para>
1004 </answer>
1005 </qandaentry>
1007 <qandaentry>
1008 <question>
1009 <para>
1010 Will named be affected by the 2007 changes to daylight savings
1011 rules in the US.
1012 </para>
1013 </question>
1014 <answer>
1015 <para>
1016 No, so long as the machines internal clock (as reported
1017 by "date -u") remains at UTC. The only visible change
1018 if you fail to upgrade your OS, if you are in a affected
1019 area, will be that log messages will be a hour out during
1020 the period where the old rules do not match the new rules.
1021 </para>
1022 <para>
1023 For most OS's this change just means that you need to
1024 update the conversion rules from UTC to local time.
1025 Normally this involves updating a file in /etc (which
1026 sets the default timezone for the machine) and possibly
1027 a directory which has all the conversion rules for the
1028 world (e.g. /usr/share/zoneinfo). When updating the OS
1029 do not forget to update any chroot areas as well.
1030 See your OS's documentation for more details.
1031 </para>
1032 <para>
1033 The local timezone conversion rules can also be done on
1034 a individual basis by setting the TZ environment variable
1035 appropriately. See your OS's documentation for more
1036 details.
1037 </para>
1038 </answer>
1039 </qandaentry>
1041 <qandaentry>
1042 <question>
1043 <para>
1044 Is there a bugzilla (or other tool) database that mere
1045 mortals can have (read-only) access to for bind?
1046 </para>
1047 </question>
1048 <answer>
1049 <para>
1050 No. The BIND 9 bug database is kept closed for a number
1051 of reasons. These include, but are not limited to, that
1052 the database contains proprietory information from people
1053 reporting bugs. The database has in the past and may in
1054 future contain unfixed bugs which are capable of bringing
1055 down most of the Internet's DNS infrastructure.
1056 </para>
1057 <para>
1058 The release pages for each version contain up to date
1059 lists of bugs that have been fixed post release. That
1060 is as close as we can get to providing a bug database.
1061 </para>
1062 </answer>
1063 </qandaentry>
1065 <qandaentry>
1066 <question>
1067 <para>
1068 Why do queries for NSEC3 records fail to return the NSEC3 record?
1069 </para>
1070 </question>
1071 <answer>
1072 <para>
1073 NSEC3 records are strictly meta data and can only be
1074 returned in the authority section. This is done so that
1075 signing the zone using NSEC3 records does not bring names
1076 into existence that do not exist in the unsigned version
1077 of the zone.
1078 </para>
1079 </answer>
1080 </qandaentry>
1082 </qandadiv> <!-- General Questions -->
1084 <qandadiv><title>Operating-System Specific Questions</title>
1086 <qandadiv><title>HPUX</title>
1088 <qandaentry>
1089 <question>
1090 <para>I get the following error trying to configure BIND:
1091 <programlisting>checking if unistd.h or sys/types.h defines fd_set... no
1092 configure: error: need either working unistd.h or sys/select.h</programlisting>
1093 </para>
1094 </question>
1095 <answer>
1096 <para>
1097 You have attempted to configure BIND with the bundled C compiler.
1098 This compiler does not meet the minimum compiler requirements to
1099 for building BIND. You need to install a ANSI C compiler and / or
1100 teach configure how to find the ANSI C compiler. The later can
1101 be done by adjusting the PATH environment variable and / or
1102 specifying the compiler via CC.
1103 </para>
1104 <informalexample>
1105 <programlisting>./configure CC=<compiler> ...</programlisting>
1106 </informalexample>
1107 </answer>
1108 </qandaentry>
1110 </qandadiv> <!-- HPUX -->
1112 <qandadiv><title>Linux</title>
1114 <qandaentry>
1115 <question>
1116 <para>
1117 Why do I get the following errors:
1118 <programlisting>general: errno2result.c:109: unexpected error:
1119 general: unable to convert errno to isc_result: 14: Bad address
1120 client: UDP client handler shutting down due to fatal receive error: unexpected error</programlisting>
1121 </para>
1122 </question>
1123 <answer>
1124 <para>
1125 This is the result of a Linux kernel bug.
1126 </para>
1127 <para>
1128 See:
1129 <ulink url="http://marc.theaimsgroup.com/?l=linux-netdev&m=113081708031466&w=2"><http://marc.theaimsgroup.com/?l=linux-netdev&m=113081708031466&w=2></ulink>
1130 </para>
1131 </answer>
1132 </qandaentry>
1134 <qandaentry>
1135 <question>
1136 <para>
1137 Why does named lock up when it attempts to connect over IPSEC tunnels?
1138 </para>
1139 </question>
1140 <answer>
1141 <para>
1142 This is due to a kernel bug where the fact that a socket is marked
1143 non-blocking is ignored. It is reported that setting
1144 xfrm_larval_drop to 1 helps but this may have negative side effects.
1145 See:
1146 <ulink url="https://bugzilla.redhat.com/show_bug.cgi?id=427629"><https://bugzilla.redhat.com/show_bug.cgi?id=427629></ulink>
1147 and
1148 <ulink url="http://lkml.org/lkml/2007/12/4/260"><http://lkml.org/lkml/2007/12/4/260></ulink>.
1149 </para>
1150 <para>
1151 xfrm_larval_drop can be set to 1 by the following procedure:
1152 <programlisting>
1153 echo "1" > proc/sys/net/core/xfrm_larval_drop</programlisting>
1154 </para>
1155 </answer>
1156 </qandaentry>
1158 <qandaentry>
1159 <question>
1160 <para>
1161 Why do I see 5 (or more) copies of named on Linux?
1162 </para>
1163 </question>
1164 <answer>
1165 <para>
1166 Linux threads each show up as a process under ps. The
1167 approximate number of threads running is n+4, where n is
1168 the number of CPUs. Note that the amount of memory used
1169 is not cumulative; if each process is using 10M of memory,
1170 only a total of 10M is used.
1171 </para>
1172 <para>
1173 Newer versions of Linux's ps command hide the individual threads
1174 and require -L to display them.
1175 </para>
1176 </answer>
1177 </qandaentry>
1179 <qandaentry>
1180 <question>
1181 <para>
1182 Why does BIND 9 log <quote>permission denied</quote> errors accessing
1183 its configuration files or zones on my Linux system even
1184 though it is running as root?
1185 </para>
1186 </question>
1187 <answer>
1188 <para>
1189 On Linux, BIND 9 drops most of its root privileges on
1190 startup. This including the privilege to open files owned
1191 by other users. Therefore, if the server is running as
1192 root, the configuration files and zone files should also
1193 be owned by root.
1194 </para>
1195 </answer>
1196 </qandaentry>
1198 <qandaentry>
1199 <question>
1200 <para>
1201 I get the error message <quote>named: capset failed: Operation
1202 not permitted</quote> when starting named.
1203 </para>
1204 </question>
1205 <answer>
1206 <para>
1207 The capability module, part of "Linux Security Modules/LSM",
1208 has not been loaded into the kernel. See insmod(8), modprobe(8).
1209 </para>
1210 <para>
1211 The relevant modules can be loaded by running:
1212 <programlisting>
1213 modprobe commoncap
1214 modprobe capability</programlisting>
1215 </para>
1216 </answer>
1217 </qandaentry>
1219 <qandaentry>
1220 <question>
1221 <para>
1222 I'm running BIND on Red Hat Enterprise Linux or Fedora Core -
1223 </para>
1224 <para>
1225 Why can't named update slave zone database files?
1226 </para>
1227 <para>
1228 Why can't named create DDNS journal files or update
1229 the master zones from journals?
1230 </para>
1231 <para>
1232 Why can't named create custom log files?
1233 </para>
1234 </question>
1236 <answer>
1237 <para>
1238 Red Hat Security Enhanced Linux (SELinux) policy security
1239 protections :
1240 </para>
1242 <para>
1243 Red Hat have adopted the National Security Agency's
1244 SELinux security policy (see <ulink
1245 url="http://www.nsa.gov/selinux"><http://www.nsa.gov/selinux></ulink>)
1246 and recommendations for BIND security , which are more
1247 secure than running named in a chroot and make use of
1248 the bind-chroot environment unnecessary .
1249 </para>
1251 <para>
1252 By default, named is not allowed by the SELinux policy
1253 to write, create or delete any files EXCEPT in these
1254 directories:
1255 <informalexample>
1256 <programlisting>
1257 $ROOTDIR/var/named/slaves
1258 $ROOTDIR/var/named/data
1259 $ROOTDIR/var/tmp
1260 </programlisting>
1261 </informalexample>
1262 where $ROOTDIR may be set in /etc/sysconfig/named if
1263 bind-chroot is installed.
1264 </para>
1266 <para>
1267 The SELinux policy particularly does NOT allow named to modify
1268 the $ROOTDIR/var/named directory, the default location for master
1269 zone database files.
1270 </para>
1272 <para>
1273 SELinux policy overrules file access permissions - so
1274 even if all the files under /var/named have ownership
1275 named:named and mode rw-rw-r--, named will still not be
1276 able to write or create files except in the directories
1277 above, with SELinux in Enforcing mode.
1278 </para>
1280 <para>
1281 So, to allow named to update slave or DDNS zone files,
1282 it is best to locate them in $ROOTDIR/var/named/slaves,
1283 with named.conf zone statements such as:
1284 <informalexample>
1285 <programlisting>
1286 zone "slave.zone." IN {
1287 type slave;
1288 file "slaves/slave.zone.db";
1289 ...
1290 };
1291 zone "ddns.zone." IN {
1292 type master;
1293 allow-updates {...};
1294 file "slaves/ddns.zone.db";
1295 };
1296 </programlisting>
1297 </informalexample>
1298 </para>
1300 <para>
1301 To allow named to create its cache dump and statistics
1302 files, for example, you could use named.conf options
1303 statements such as:
1304 <informalexample>
1305 <programlisting>
1306 options {
1307 ...
1308 dump-file "/var/named/data/cache_dump.db";
1309 statistics-file "/var/named/data/named_stats.txt";
1310 ...
1311 };
1312 </programlisting>
1313 </informalexample>
1314 </para>
1316 <para>
1317 You can also tell SELinux to allow named to update any
1318 zone database files, by setting the SELinux tunable boolean
1319 parameter 'named_write_master_zones=1', using the
1320 system-config-securitylevel GUI, using the 'setsebool'
1321 command, or in /etc/selinux/targeted/booleans.
1322 </para>
1324 <para>
1325 You can disable SELinux protection for named entirely by
1326 setting the 'named_disable_trans=1' SELinux tunable boolean
1327 parameter.
1328 </para>
1330 <para>
1331 The SELinux named policy defines these SELinux contexts for named:
1332 <informalexample>
1333 <programlisting>
1334 named_zone_t : for zone database files - $ROOTDIR/var/named/*
1335 named_conf_t : for named configuration files - $ROOTDIR/etc/{named,rndc}.*
1336 named_cache_t: for files modifiable by named - $ROOTDIR/var/{tmp,named/{slaves,data}}
1337 </programlisting>
1338 </informalexample>
1339 </para>
1341 <para>
1342 If you want to retain use of the SELinux policy for named,
1343 and put named files in different locations, you can do
1344 so by changing the context of the custom file locations
1345 .
1346 </para>
1348 <para>
1349 To create a custom configuration file location, e.g.
1350 '/root/named.conf', to use with the 'named -c' option,
1351 do:
1352 <informalexample>
1353 <programlisting>
1354 # chcon system_u:object_r:named_conf_t /root/named.conf
1355 </programlisting>
1356 </informalexample>
1357 </para>
1359 <para>
1360 To create a custom modifiable named data location, e.g.
1361 '/var/log/named' for a log file, do:
1362 <informalexample>
1363 <programlisting>
1364 # chcon system_u:object_r:named_cache_t /var/log/named
1365 </programlisting>
1366 </informalexample>
1367 </para>
1369 <para>
1370 To create a custom zone file location, e.g. /root/zones/, do:
1371 <informalexample>
1372 <programlisting>
1373 # chcon system_u:object_r:named_zone_t /root/zones/{.,*}
1374 </programlisting>
1375 </informalexample>
1376 </para>
1378 <para>
1379 See these man-pages for more information : selinux(8),
1380 named_selinux(8), chcon(1), setsebool(8)
1381 </para>
1382 </answer>
1383 </qandaentry>
1385 <qandaentry>
1386 <question>
1387 <para>
1388 Listening on individual IPv6 interfaces does not work.
1389 </para>
1390 </question>
1391 <answer>
1392 <para>
1393 This is usually due to "/proc/net/if_inet6" not being available
1394 in the chroot file system. Mount another instance of "proc"
1395 in the chroot file system.
1396 </para>
1397 <para>
1398 This can be be made permanent by adding a second instance to
1399 /etc/fstab.
1400 <informalexample>
1401 <programlisting>
1402 proc /proc proc defaults 0 0
1403 proc /var/named/proc proc defaults 0 0</programlisting>
1404 </informalexample>
1405 </para>
1406 </answer>
1407 </qandaentry>
1409 </qandadiv> <!-- Linux -->
1411 <qandadiv><title>Windows</title>
1413 <qandaentry>
1414 <question>
1415 <para>
1416 Zone transfers from my BIND 9 master to my Windows 2000
1417 slave fail. Why?
1418 </para>
1419 </question>
1420 <answer>
1421 <para>
1422 This may be caused by a bug in the Windows 2000 DNS server
1423 where DNS messages larger than 16K are not handled properly.
1424 This can be worked around by setting the option "transfer-format
1425 one-answer;". Also check whether your zone contains domain
1426 names with embedded spaces or other special characters,
1427 like "John\032Doe\213s\032Computer", since such names have
1428 been known to cause Windows 2000 slaves to incorrectly
1429 reject the zone.
1430 </para>
1431 </answer>
1432 </qandaentry>
1434 <qandaentry>
1435 <question>
1436 <para>
1437 I get <quote>Error 1067</quote> when starting named under Windows.
1438 </para>
1439 </question>
1440 <answer>
1441 <para>
1442 This is the service manager saying that named exited. You
1443 need to examine the Application log in the EventViewer to
1444 find out why.
1445 </para>
1446 <para>
1447 Common causes are that you failed to create "named.conf"
1448 (usually "C:\windows\dns\etc\named.conf") or failed to
1449 specify the directory in named.conf.
1450 </para>
1451 <informalexample>
1452 <programlisting>
1453 options {
1454 Directory "C:\windows\dns\etc";
1455 };</programlisting>
1456 </informalexample>
1457 </answer>
1458 </qandaentry>
1460 </qandadiv> <!-- Windows -->
1462 <qandadiv><title>FreeBSD</title>
1464 <qandaentry>
1465 <question>
1466 <para>
1467 I have FreeBSD 4.x and "rndc-confgen -a" just sits there.
1468 </para>
1469 </question>
1470 <answer>
1471 <para>
1472 /dev/random is not configured. Use rndcontrol(8) to tell
1473 the kernel to use certain interrupts as a source of random
1474 events. You can make this permanent by setting rand_irqs
1475 in /etc/rc.conf.
1476 </para>
1477 <informalexample>
1478 <programlisting>
1479 rand_irqs="3 14 15"</programlisting>
1480 </informalexample>
1481 <para>
1482 See also
1483 <ulink url="http://people.freebsd.org/~dougb/randomness.html">
1484 <http://people.freebsd.org/~dougb/randomness.html></ulink>.
1485 </para>
1486 </answer>
1487 </qandaentry>
1489 </qandadiv> <!-- FreeBSD -->
1491 <qandadiv><title>Solaris</title>
1493 <qandaentry>
1494 <question>
1495 <para>
1496 How do I integrate BIND 9 and Solaris SMF
1497 </para>
1498 </question>
1499 <answer>
1500 <para>
1501 Sun has a blog entry describing how to do this.
1502 </para>
1503 <para>
1504 <ulink
1505 url="http://blogs.sun.com/roller/page/anay/Weblog?catname=%2FSolaris">
1506 <http://blogs.sun.com/roller/page/anay/Weblog?catname=%2FSolaris>
1507 </ulink>
1508 </para>
1509 </answer>
1510 </qandaentry>
1512 </qandadiv>
1514 <qandadiv><title>Apple Mac OS X</title>
1516 <qandaentry>
1517 <question>
1518 <para>
1519 How do I run BIND 9 on Apple Mac OS X?
1520 </para>
1521 </question>
1522 <answer>
1523 <para>
1524 If you run Tiger(Mac OS 10.4) or later then this is all you need to do:
1525 </para>
1526 <informalexample>
1527 <programlisting>
1528 % sudo rndc-confgen > /etc/rndc.conf</programlisting>
1529 </informalexample>
1530 <para>
1531 Copy the key statement from /etc/rndc.conf into /etc/rndc.key, e.g.:
1532 </para>
1533 <informalexample>
1534 <programlisting>
1535 key "rndc-key" {
1536 algorithm hmac-md5;
1537 secret "uvceheVuqf17ZwIcTydddw==";
1538 };</programlisting>
1539 </informalexample>
1540 <para>
1541 Then start the relevant service:
1542 </para>
1543 <informalexample>
1544 <programlisting>
1545 % sudo service org.isc.named start</programlisting>
1546 </informalexample>
1547 <para>
1548 This is persistent upon a reboot, so you will have to do it only once.
1549 </para>
1550 </answer>
1552 <answer>
1553 <para>
1554 Alternatively you can just generate /etc/rndc.key by running:
1555 </para>
1556 <informalexample>
1557 <programlisting>
1558 % sudo rndc-confgen -a</programlisting>
1559 </informalexample>
1560 <para>
1561 Then start the relevant service:
1562 </para>
1563 <informalexample>
1564 <programlisting>
1565 % sudo service org.isc.named start</programlisting>
1566 </informalexample>
1567 <para>
1568 Named will look for /etc/rndc.key when it starts if it
1569 doesn't have a controls section or the existing controls are
1570 missing keys sub-clauses. This is persistent upon a
1571 reboot, so you will have to do it only once.
1572 </para>
1573 </answer>
1574 </qandaentry>
1576 </qandadiv>
1578 </qandadiv> <!-- Operating-System Specific Questions -->
1580 </qandaset>
1581 </article>