chiark-named-conf - check and generate nameserver configuration
chiark-named-conf [options] -n|-y|-f
chiark-named-conf [options] zone ...
chiark-named-conf is a tool for managing nameserver configurations and
checking for suspected DNS problems. Its main functions are to check
that delegations are appropriate and working, that secondary zones are
slaved from the right places, and to generate a configuration for BIND,
from its own input file.
By default, for each zone, in addition to any warnings, the output
lists the zone’s configuration type. If the zone is checked, the
serial number at each of the nameservers is shown, with any unpublished
primary having * after the serial number.
If one of the options -n, -y, or -f is supplied then chiark-named-conf
will read its main configuration file for the list of relevant zones.
It will then check the configuration and delegation for each zone
and/or generate and install a new configuration file for the
Generate and install new nameserver config, as well as checking
configuration, for all listed zones.
Check configuration, for all listed zones, but do not generate
new nameserver config.
Generate and install new nameserver config, without doing any
configuration cross-checking. (Syntax errors in our input
configuration will still abort this operation.)
Do nothing: do no checks, and don’t write a new config. This
can be used to get a list of the zones being processed.
--mail-first | --mail-middle | --mail-final
Send mails to zone SOA MNAMEs reporting zones with problems.
You must call chiark-named-conf at least twice, once with
--mail-first, and later with --mail-final, and preferably with
one or more calls to --mail-middle in between. All three
options carry out a check and store the results; --mail-final
also sends a mail to the zone SOA MNAME or local administrator,
if too many of the calls had errors or warnings (calls before
the most recent --mail-first being ignored).
just like --mail-final except that it always sends mail to the
local server admin and never to remote zone contacts, adding
(testing!) to the start of the To: field.
Alternatively, one or more zone names may be supplied as arguments, in
which case their delegations will be checked, and compared with the
data for that zone in the main configuration (if any). In this case no
new configuration file for the nameserver will be made.
Checks even zones known to be broken. Ie, ignores the ? zone
style modifier in the configuration.
Use config-file instead of /etc/bind/chiark-conf-gen.zones.
Also changes the default directory.
-D Enables debugging. Useful for debugging chiark-named-conf, but
probably not useful for debugging your DNS configuration.
Repeat to increase the debugging level. (Maximum is -DD.)
Do not warn about glueless referrals (strictly, makes the zone
style modifier ~ the default). Not recommended - see the
section GLUELESSNESS, below.
Only checks for mistakes which are the responsibility of the
local administrator (to fix or get fixed). This means that for
published and stealth zones we only check that we’re slaving
from the right place and that any names and addresses for
ourself are right. For primary zones all checks are still done.
It is a mistake to specify -l with foreign zones (zones supplied
explictly on the command line but not relevant to the local
server); doing so produces a warning.
Overrides a modifiers directive in the configuration file. The
modifiers specified in the directive are completely replaced by
those specified in this command line option. (Note that
modifiers specified in per-zone directives still override these
per-group settings.) If more than one modifiers directive
specifies the same group, they are all affected. modifiers
directives which don’t specify a group cannot be affected. It
is an error if the group does not appear in the config file.
See ZONE STYLE MODIFIERS, below.
The special group foreign is used for zones which don’t appear in the
Suppress the usual report of the list of nameservers for each
zone and the serial number from each. When specified twice, do
not print any information except warnings.
When a problem is detected, warn for all sources of the same
imperfect data, rather than only the first we come across
Print additional information about what is being checked, as we
The file /etc/bind/chiark-conf-gen.zones (or other file specified with
the -C option) contains a sequence of directives, one per line. Blank
lines are permitted. Leading and trailing whitespace on each line is
ignored. Comments are lines starting with #. Ending a line with a
joins it to the next line, so that long directives can be split across
several physical lines.
These directives specify general configuration details. They should
appear before directives specifying zones, as each will affect only
later zone directives. Foreign zones (zones explicitly specified on
the command line but not mentioned in the configuration) use the
configuration settings prevailing at the end of the config file.
Specifies the email address of the local administrator. This is
used in the From: line of mails sent out, and will also receive
copies of the reports. There is no default.
Makes directory be the default directory (which affects the
interpretation of relative filenames). The default is the
directory containing the main configuration file, ie /etc/bind
if no -C option is specified.
forbid-addr [ip-address ...]
Specifies the list of addresses that are forbidden as any
nameserver for any zone. The default is no such addresses.
forbid-addr [ip-address ...]
Specifies the list of addresses that are forbidden as a
nameserver for a zone for which we are the primary - ie, the
list of our old or to-be-obsoleted slaves. The default is no
serverless-glueless domain ...
Specifies a list of domains under which we do not expect to find
any nameservers without glue; for these zones it is OK to find
glueless referrals. Each domain listed names a complete subtree
of the DNS, starting at the named point. The default is
in-addr.arpa ip6.arpa ip6.int.
To avoid indefinitely long or even circularly glueless referrals
(which delay or prevent lookups) it is necessary for all sites
to effectively implement similar conventions; currently the
author believes that only the reverse lookup namespaces are
conventionally devoid of (glueless) nameservers, and therefore
fine to provide glueless referrals for. See GLUELESSNESS below.
allow--indirect-glue nameserver-superdomain ...
Specifies a list of domains under which we expect to find
glueless nameservers, with up to one layer of indirection. For
nameservers under these domains it is OK to to find glueless
referrals, but only when listed as a nameserver for a zone which
is not itself a subdomain of an allow-indirect-glue nameserver-
This supports to common configuration style where DNS
operator(s) set up all of their nameservers with names within a
small subsection of the DNS (the portions under nameserver-
superdomains), and provide glueless referrals naming these
nameservers for all other zones. This provides at most one
level of missing glue.
Note that if the DNS administrators collectively able to
influence the service for some zone (including the admins for
its superzones, the zones containing its nameservers, and their
superzones and so forth) are not in sufficiently close
communication do not all agree on the proper set of nameserver-
superdomain then they might still set up circular glue and
chiark-named-conf would not necessarily be able to detect this
even if it was run on every relevant nameserver.
Uses directory for storing information about recent failures for
mailing to zone admins. See --mail-first et al. Old files in
here should be cleaned up periodically out of cron. There is no
When --mail-final is used, a mail will be sent to all zones
which had warnings or errors more than percentage% of the times
--mail-* was used (since the last --mail-first). The default is
modifiers !*$@~?] [group]
Applies the specified zone style modifiers (see below) to
subsequently declared zones (until the next modifiers
directive), as if the modifiers specified were written out for
each zone. You must specify at least one character for the
modifiers; if you want to reset everything to the default, just
say !. If style modifiers specified in the zone directive
conflict with the modifiers directive, those specified in the
zone directive take effect. group may contain alphanumerics and
underscores, and is used for the -m command-line option.
self-addr ip-address ...
Specifies the list of addresses that this server may be known by
in A records. There is no default.
output format filename [format filename ...]
Arranges that each filename will be overwritten when -y or -f
are used; its new contents will be configuration directives for
the zones which follow for the nameserver in question.
Currently the only format supported is bind8 which indicates
new-style BIND 8. If no zones follow, then each file will still
be overwritten, by an effectively empty file. Default: if there
is no output directive in the configuration then the default is
to use bind8 chiark-conf-gen.bind8; otherwise it is an error for
there to be any zones in the configuration before the first
self-ns fqdn ...
Specifies the list of names that this server may be known by in
NS records. There is no default. Any trailing * is replaced by
the name of the zone being checked, so for example self-ns
isp.ns.* before the zone example.com would mean to expect us to
be listed as isp.ns.example.com in the NS RRset.
self-soa fqdn ...
Specifies the list of names that this server may be known by in
the ORIGIN field of SOA records. There is no default. Any
trailing * is replaced by the name of the zone, as for self-ns.
self fqdn ...
Equivalent to both self-ns and self-soa with the same set of
slave-dir directory [[prefix] suffix]
Specifies the directory in which slave (published and stealth)
zonefiles should be placed. The default directory is
/var/cache/bind/chiark-slave. The default suffix and prefix are
empty; they also will be reset to these defaults by a slave-dir
directive which does not specify them.
These directives specify one or more zones.
primary[!*$@~?] zone filename
Specifies that this server is supposed to be the primary
nameserver for zone and that the zone data is to be found in
primary-dir[!*$@~?] directory[/prefix] [suffix[/subfile]]
Search directory for files whose names start with prefix and end
with suffix. Each such file is taken to represent a zone file
for which this server is supposed to be the primary; the part of
the filename between prefix and suffix is the name of the zone.
If /subfile is specified, then instead of looking for files, we
search for directories containing subfile; directories which do
not contain the subfile are simply skipped.
If directory[/prefix] exists as specified and is a directory
then it is interpreted as directory with an empty prefix;
otherwise the final path component is assumed to be the prefix.
If no suffix/subfile is specified then the default is _db.
published[!*$@~?] zone origin-addr
Specifies that this server is supposed to be a published slave
nameserver for the zone in question.
stealth[!*$@~?] zone server-addr ...
Specifies that this server is supposed to be an unpublished
secondary (aka stealth secondary) for the zone in question.
ZONE STYLE MODIFIERS
Each of the zone directives may optionally be followed by one or more
of the following characters (each at most once):
! Reverses the meaning of all style modifiers after the !. Only
one ! must appear in the modifier list. In this list, other
modifiers which default to ‘enabled’ are described by describing
the effect of their inverse - see the description for !@ below.
* Indicates that the zone is unofficial, ie that it is not
delegated as part of the global Internet DNS and that no attempt
should be made to find the superzone and check delegations.
Note that unofficial, local zones should be created with
caution. They should be in parts of the namespace which are
reserved for private use, or belong to the actual zone
$ Indicates that any mails should be sent about the zone to the
nameserver admin rather than to the zone SOA MNAME. This is the
default unless we are supposedly a published server for the
!@ Indicates that no mails should be sent about the zone to anyone.
~ Indicates that the zone’s delegation is known to be glueless,
and that lack of glue should not be flagged. Not recommended -
see the section GLUELESSNESS, below.
? Indicates that the zone is known to be broken and no checks
should be carried out on it, unless the -A option is specified.
Reads file as if it were included here.
end Ends processing of this file; any data beyond this point is
chiark-named-conf makes the following checks:
Delegations: Each delegation from a server for the superzone should
contain the same set of nameservers. None of the delegations should
lack glue. The glue addresses should be the same in each delegation,
and agree with the local default nameserver.
Delegated servers: Each server mentioned in the delegation should have
the same SOA record (and obviously, should be authoritative).
All published nameservers - including delegated servers and servers
named in the zone’s nameserver set: All nameservers for the zone should
supply the same list of nameservers for the zone, and none of this
authority information should be glueless. All the glue should always
give the same addresses.
Origin server’s data: The set of nameservers in the origin server’s
version of the zone should be a superset of those in the delegations.
Our zone configuration: For primary zones, the SOA origin should be one
of the names specified with self-soa (or self). For published zones,
the address should be that of the SOA origin. For stealth zones, the
address should be that of the SOA origin or one of the published
Glue is the name given for the addresses of nameservers which are often
supplied in a referral. In fact, it turns out that it is important for
the reliability and performance of the DNS that referrals, in general,
always come with glue.
Firstly, glueless referrals usually cause extra delays looking up
names. BIND 8, when it receives a completely glueless referral and
does not have the nameservers’ addresses in its cache, will start
queries for the nameserver addresses; but it will throw the original
client’s question away, so that when these queries arrive, it won’t
restart the query from where it left off. This means that the client
won’t get its answer until it retries, typically at least 1 second
later - longer if you have more than one nameserver listed. Worse, if
the nameserver to which the glueless referral points is itself under
another glueless referral, another retry will be required.
Even for better resolvers than BIND 8, long chains of glueless
referrals can cause performance and reliability problems, turning a
simple two or three query exchange into something needing more than a
Even worse, one might accidentally create a set of circularly glueless
referrals such as
example.com NS ns0.example.net.uk
example.com NS ns1.example.net.uk
example.net.uk NS ns0.example.com
example.net.uk NS ns1.example.com
Here it is impossible to look up anything in either example.com or
There are, as far as the author is aware, no generally agreed
conventions or standards for avoiding unreasonably long glueless
chains, or even circular glueless situations. The only way to
guarantee that things will work properly is therefore to always supply
However, the situation is further complicated by the fact that many
implementations (including BIND 8.2.3, and many registry systems), will
refuse to accept glue RRs for delegations in a parent zonefile unless
they are under the parent’s zone apex. In these cases it can be
necessary to create names for the child’s nameservers which are
underneath the child’s apex, so that the glue records are both in the
parent’s bailiwick and obviously necessary.
In the past, the ‘shared registry system’ managing .com, .net and .org
did not allow a single IPv4 address to be used for more than one
nameserver name. However, at the time of writing (October 2002) this
problem seems to have been fixed, and the workaround I previously
recommended (creating a single name for your nameserver somewhere in
.com, .net or .org, and using that for all the delegations from .com,
.net and .org) should now be avoided.
Finally, a note about ‘reverse’ zones, such as those in in-addr.arpa:
It does not seem at all common practice to create nameservers in in-
addr.arpa zones (ie, no NS RRs seem to point into in-addr.arpa, even
those for in-addr.arpa zones). Current practice seems to be to always
use nameservers for in-addr.arpa which are in the normal, forward,
address space. If everyone sticks to the rule of always publishing
nameservers names in the ‘main’ part of the namespace, and publishing
glue for them, there is no chance of anything longer than a 1-step
glueless chain might occur for a in-addr.arpa zone. It is probably
best to maintain this as the status quo, despite the performance
problem this implies for BIND 8 caches. This is what the
serverless-glueless directive is for.
Dan Bernstein has some information and examples about this at
http://cr.yp.to/djbdns/notes.html#gluelessness but be warned that it is
I recommend that every nameserver should have its own name in every
forward zone that it serves. For example:
zone.example.com NS servus.ns.example.com
servus.ns.example.com A 127.0.0.2
18.104.22.168.in-addr.arpa PTR servus.example.net
servus.example.net A 127.0.0.2
Domain names in in-addr.arpa should not be used in the right hand side
of NS records.
chiark-named-conf is supposed to be resistant to malicious data in the
DNS. It is not resistant to malicious data in its own options,
configuration file or environment. It is not supposed to read its
stdin, but is not guaranteed to be safe if stdin is dangerous.
Killing chiark-named-conf suddenly should be safe, even with -y or -f
(though of course it may not complete its task if killed), provided
that only one invocation is made at once.
Slow remote nameservers will cause chiark-named-conf to take
0 All went well and there were no warnings.
There were warnings or errors.
Default input configuration file. (Override with -C.)
Default directory. (Override with -C or default-dir.)
Default output file.
Default location for slave zones.
Setting variables used by dig(1) and adnshost(1) will affect the
operation of chiark-named-conf. Avoid messing with these if possible.
PATH is used to find subprograms such as dig and adnshost.
The determination of the parent zone for each zone to be checked, and
its nameservers, is done simply using the system default nameserver.
The processing of output from dig is not very reliable or robust, but
this is mainly the fault of dig. This can lead to somewhat unhelpful
error reporting for lookup failures.
chiark-named-conf and this manpage were written by Ian Jackson
<email@example.com>. They are Copyright 2002 Ian Jackson.
chiark-named-conf and this manpage are free software; you can
redistribute it and/or modify it under the terms of the GNU General
Public License as published by the Free Software Foundation; either
version 2, or (at your option) any later version.
This is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
for more details.
You should have received a copy of the GNU General Public License along
with this program; if not, consult the Free Software Foundation’s
website at www.fsf.org, or the GNU Project website at www.gnu.org.