1 Copyright (C) 2004, 2007 Internet Systems Consortium, Inc. ("ISC")
2 Copyright (C) 2000, 2001, 2003 Internet Software Consortium.
3 See COPYRIGHT in the source root or http://isc.org/copyright.html for terms.
5 BIND 8 to BIND 9 Migration Notes
7 BIND 9 is designed to be mostly upwards compatible with BIND 8, but
8 there is still a number of caveats you should be aware of when
9 upgrading an existing BIND 8 installation to use BIND 9.
12 1. Configuration File Compatibility
14 1.1. Unimplemented Options and Changed Defaults
16 BIND 9 supports most, but not all of the named.conf options of BIND 8.
17 For a complete list of implemented options, see doc/misc/options.
19 If your named.conf file uses an unimplemented option, named will log a
20 warning message. A message is also logged about each option whose
21 default has changed unless the option is set explicitly in named.conf.
23 The default of the "transfer-format" option has changed from
24 "one-answer" to "many-answers". If you have slave servers that do not
25 understand the many-answers zone transfer format (e.g., BIND 4.9.5 or
26 older) you need to explicitly specify "transfer-format one-answer;" in
27 either the options block or a server statement.
29 1.2. Handling of Configuration File Errors
31 In BIND 9, named refuses to start if it detects an error in
32 named.conf. Earlier versions would start despite errors, causing the
33 server to run with a partial configuration. Errors detected during
34 subsequent reloads do not cause the server to exit.
36 Errors in master files do not cause the server to exit, but they
37 do cause the zone not to load.
41 The set of logging categories in BIND 9 is different from that
42 in BIND 8. If you have customised your logging on a per-category
43 basis, you need to modify your logging statement to use the
46 Another difference is that the "logging" statement only takes effect
47 after the entire named.conf file has been read. This means that when
48 the server starts up, any messages about errors in the configuration
49 file are always logged to the default destination (syslog) when the
50 server first starts up, regardless of the contents of the "logging"
51 statement. In BIND 8, the new logging configuration took effect
52 immediately after the "logging" statement was read.
54 1.4. Notify messages and Refresh queries
56 The source address and port for these is now controlled by
57 "notify-source" and "transfer-source", respectively, rather that
58 query-source as in BIND 8.
60 1.5. Multiple Classes.
62 Multiple classes have to be put into explicit views for each class.
65 2. Zone File Compatibility
67 2.1. Strict RFC1035 Interpretation of TTLs in Zone Files
69 BIND 9 strictly complies with the RFC1035 and RFC2308 rules regarding
70 omitted TTLs in zone files. Omitted TTLs are replaced by the value
71 specified with the $TTL directive, or by the previous explicit TTL if
72 there is no $TTL directive.
74 If there is no $TTL directive and the first RR in the file does not
75 have an explicit TTL field, the zone file is illegal according to
76 RFC1035 since the TTL of the first RR is undefined. Unfortunately,
77 BIND 4 and many versions of BIND 8 accept such files without warning
78 and use the value of the SOA MINTTL field as a default for missing TTL
81 BIND 9.0 and 9.1 completely refused to load such files. BIND 9.2
82 emulates the nonstandard BIND 4/8 SOA MINTTL behaviour and loads the
83 files anyway (provided the SOA is the first record in the file), but
84 will issue the warning message "no TTL specified; using SOA MINTTL
87 To avoid problems, we recommend that you use a $TTL directive in each
90 2.2. Periods in SOA Serial Numbers Deprecated
92 Some versions of BIND allow SOA serial numbers with an embedded
93 period, like "3.002", and convert them into integers in a rather
94 unintuitive way. This feature is not supported by BIND 9; serial
95 numbers must be integers.
97 2.3. Handling of Unbalanced Quotes
99 TXT records with unbalanced quotes, like 'host TXT "foo', were not
100 treated as errors in some versions of BIND. If your zone files
101 contain such records, you will get potentially confusing error
102 messages like "unexpected end of file" because BIND 9 will interpret
103 everything up to the next quote character as a literal string.
105 2.4. Handling of Line Breaks
107 Some versions of BIND accept RRs containing line breaks that are not
108 properly quoted with parentheses, like the following SOA:
110 @ IN SOA ns.example. hostmaster.example.
111 ( 1 3600 1800 1814400 3600 )
113 This is not legal master file syntax and will be treated as an error
114 by BIND 9. The fix is to move the opening parenthesis to the first
117 2.5. Unimplemented BIND 8 Extensions
119 $GENERATE: The "$$" construct for getting a literal $ into a domain
120 name is deprecated. Use \$ instead.
122 2.6. TXT records are no longer automatically split.
124 Some versions of BIND accepted strings in TXT RDATA consisting of more
125 than 255 characters and silently split them to be able to encode the
126 strings in a protocol conformant way. You may now see errors like this
127 dns_rdata_fromtext: local.db:119: ran out of space
128 if you have TXT RRs with too longs strings. Make sure to split the
129 string in the zone data file at or before a single one reaches 255
132 3. Interoperability Impact of New Protocol Features
136 BIND 9 uses EDNS0 (RFC2671) to advertise its receive buffer size. It
137 also sets DO EDNS flag bit in queries to indicate that it wishes to
138 receive DNSSEC responses.
140 Most older servers that do not support EDNS0, including prior versions
141 of BIND, will send a FORMERR or NOTIMP response to these queries.
142 When this happens, BIND 9 will automatically retry the query without
145 Unfortunately, there exists at least one non-BIND name server
146 implementation that silently ignores these queries instead of sending
147 an error response. Resolving names in zones where all or most
148 authoritative servers use this server will be very slow or fail
149 completely. We have contacted the manufacturer of the name server in
150 case, and they are working on a solution.
152 When BIND 9 communicates with a server that does support EDNS0, such as
153 another BIND 9 server, responses of up to 4096 bytes may be
154 transmitted as a single UDP datagram which is subject to fragmentation
155 at the IP level. If a firewall incorrectly drops IP fragments, it can
156 cause resolution to slow down dramatically or fail.
160 Outgoing zone transfers now use the "many-answers" format by default.
161 This format is not understood by certain old versions of BIND 4.
162 You can work around this problem using the option "transfer-format
163 one-answer;", but since these old versions all have known security
164 problems, the correct fix is to upgrade the slave servers.
166 Zone transfers to Windows 2000 DNS servers sometimes fail due to a
167 bug in the Windows 2000 DNS server where DNS messages larger than
168 16K are not handled properly. Obtain the latest service pack for
169 Windows 2000 from Microsoft to address this issue. In the meantime,
170 the problem can be worked around by setting "transfer-format one-answer;".
171 http://support.microsoft.com/default.aspx?scid=kb;en-us;297936
173 4. Unrestricted Character Set
177 BIND 9 does not restrict the character set of domain names - it is
178 fully 8-bit clean in accordance with RFC2181 section 11.
180 It is strongly recommended that hostnames published in the DNS follow
181 the RFC952 rules, but BIND 9 will not enforce this restriction.
183 Historically, some applications have suffered from security flaws
184 where data originating from the network, such as names returned by
185 gethostbyaddr(), are used with insufficient checking and may cause a
186 breach of security when containing unexpected characters; see
187 <http://www.cert.org/advisories/CA-96.04.corrupt_info_from_servers.html>
188 for details. Some earlier versions of BIND attempt to protect these
189 flawed applications from attack by discarding data containing
190 characters deemed inappropriate in host names or mail addresses, under
191 the control of the "check-names" option in named.conf and/or "options
192 no-check-names" in resolv.conf. BIND 9 provides no such protection;
193 if applications with these flaws are still being used, they should
196 BIND 9.3 onwards implements check-names.
198 5. Server Administration Tools
200 5.1 Ndc Replaced by Rndc
202 The "ndc" program has been replaced by "rndc", which is capable of
203 remote operation. Unlike ndc, rndc requires a configuration file.
204 The easiest way to generate a configuration file is to run
205 "rndc-confgen -a"; see the man pages for rndc(8), rndc-confgen(8),
206 and rndc.conf(5) for details.
208 5.2. Nsupdate Differences
210 The BIND 8 implementation of nsupdate had an undocumented feature
211 where an update request would be broken down into multiple requests
212 based upon the discovered zones that contained the records. This
213 behaviour has not been implemented in BIND 9. Each update request
214 must pertain to a single zone, but it is still possible to do multiple
215 updates in a single invocation of nsupdate by terminating each update
216 with an empty line or a "send" command.
219 6. No Information Leakage between Zones
221 BIND 9 stores the authoritative data for each zone in a separate data
222 structure, as recommended in RFC1035 and as required by DNSSEC and
223 IXFR. When a BIND 9 server is authoritative for both a child zone and
224 its parent, it will have two distinct sets of NS records at the
225 delegation point: the authoritative NS records at the child's apex,
226 and a set of glue NS records in the parent.
228 BIND 8 was unable to properly distinguish between these two sets of NS
229 records and would "leak" the child's NS records into the parent,
230 effectively causing the parent zone to be silently modified: responses
231 and zone transfers from the parent contained the child's NS records
232 rather than the glue configured into the parent (if any). In the case
233 of children of type "stub", this behaviour was documented as a feature,
234 allowing the glue NS records to be omitted from the parent
237 Sites that were relying on this BIND 8 behaviour need to add any
238 omitted glue NS records, and any necessary glue A records, to the
241 Although stub zones can no longer be used as a mechanism for injecting
242 NS records into their parent zones, they are still useful as a way of
243 directing queries for a given domain to a particular set of name
247 7. Umask not Modified
249 The BIND 8 named unconditionally sets the umask to 022. BIND 9 does
250 not; the umask inherited from the parent process remains in effect.
251 This may cause files created by named, such as journal files, to be
252 created with different file permissions than they did in BIND 8. If
253 necessary, the umask should be set explicitly in the script used to
254 start the named process.
257 $Id: migration,v 1.45.18.2 2007/09/07 06:34:21 marka Exp $