Name Server Operations Guide for BIND Release 4.9.3 Releases from 4.9 Paul Vixie[1] Internet Software Consortium La Honda, CA Releases through 4.8.3 Kevin J. Dunlap[2] Michael J. Karels Computer Systems Research Group Computer Science Division Department of Electrical Engineering and Computer Sciences University of California Berkeley, CA 94720 1. Introduction The Berkeley Internet Name Domain (BIND) implements an Internet name server for BSD-derived operating sys- tems. The BIND consists of a server (or ``daemon'') called named and a resolver library. A name server is a network service that enables clients to name resources or objects and share this information with other objects in the network. This in effect is a distributed data base system for objects in a computer network. The BIND server runs in the background, servicing queries on a ____________________ [1] This author was employed by Digital Equipment Corporation's Network Systems Laboratory during the development and release of BIND 4.9. Release 4.9.2 was sponsored by Vixie Enterprises. Releases from 4.9.3 were sponsored by the Internet Software Consortium. [2] This author was an employee of Digital Equipment Corporation's ULTRIX Engineering Advanced Development Group and was on loan to CSRG when this work was done. ULTRIX is a trademark of Digital Equipment Corporation. SMM:10-2 Name Server Operations Guide for BIND well known network port. The standard port for UDP and TCP is specified in /etc/services. The resolver is a set of routines residing in a system library that provides the interface that programs can use to access the domain name services. BIND is fully integrated into BSD (4.3 and later releases) network programs for use in storing and retrieving host names and address. The system adminis- trator can configure the system to use BIND as a replace- ment to the older host table lookup of information in the network hosts file /etc/hosts. The default configuration for BSD uses BIND. 2. The Name Service The basic function of the name server is to provide information about network objects by answering queries. The specifications for this name server are defined in RFC1034, RFC1035 and RFC974. These documents can be found in /usr/src/etc/named/doc in 4.3BSD or ftped from ftp.rs.internic.net. It is also recommended that you read the related manual pages, named(8), resolver(3), and resolver(5). The advantage of using a name server over the host table lookup for host name resolution is to avoid the need for a single centralized clearinghouse for all names. The authority for this information can be delegated to the different organizations on the network responsible for it. The host table lookup routines require that the mas- ter file for the entire network be maintained at a cen- tral location by a few people. This works fine for small networks where there are only a few machines and the dif- ferent organizations responsible for them cooperate. But this does not work well for large networks where machines cross organizational boundaries. With the name server, the network can be broken into a hierarchy of domains. The name space is organized as a tree according to organizational or administrative boun- daries. Each node, called a domain, is given a label, and the name of the domain is the concatenation of all the labels of the domains from the root to the current domain, listed from right to left separated by dots. A label need only be unique within its domain. The whole space is partitioned into several areas called zones, each starting at a domain and extending down to the leaf domains or to domains where other zones start. Zones usu- ally represent administrative boundaries. An example of Name Server Operations Guide for BIND SMM:10-3 a host address for a host at the University of Califor- nia, Berkeley would look as follows: monet.Berkeley.EDU The top level domain for educational organizations is EDU; Berkeley is a subdomain of EDU and monet is the name of the host. 3. Types of Zones A ``zone'' is a point of delegation in the DNS tree. It contains all names from a certain point ``downward'' except those which are delegated to other zones. A ``delegation point'' has one or more NS records in the ``parent zone'', which should be matched by equivalent NS records at the root of the ``delegated zone'' (i.e., the ``@'' name in the zone file). Understanding the difference between a ``zone'' and a ``domain'' is crucial to the proper operation of a name server. As an example, consider the DEC.COM domain, which includes names such as POBOX1.PA.DEC.COM and QUABBIN.CRL.DEC.COM even though the DEC.COM zone includes only delegations for the PA.DEC.COM and CRL.DEC.COM zones. A zone can map exactly to a single domain, but could also include only part of a domain (the rest of which could be delegated to other name servers). Techni- cally speaking, every name in the DNS tree is a ``domain'', even if it is ``terminal'', that is, has no ``subdomains''. Technically speaking, every subdomain is a domain and every domain except the root is also a sub- domain. The terminology is not intuitive and you would do well to read RFC's 1033, 1034, and 1035 to gain a com- plete understanding of this difficult and subtle topic. Though BIND is a Domain Name Server, it deals pri- marily in terms of zones. The primary and secondary declarations in the named.boot file specify zones, not domains. When you ask someone if they are willing to be a secondary server for your ``domain'', you are actually asking for secondary service for some collection of zones. Each zone will have one ``primary'' server, which loads the zone contents from some local file which is edited by humans or perhaps generated mechanically from some other local file which is edited by humans. Then there will be some number of ``secondary'' servers, which load the zone contents using the IP/DNS protocol (that is, the secondary servers will contact the primary and fetch the zone using IP/TCP). This set of servers (the SMM:10-4 Name Server Operations Guide for BIND primary and all of the secondaries) should be listed in the NS records in the parent zone, which will constitute a ``delegation''. This set of servers must also be listed in the zone file itself, usually under the ``@'' name which is a magic cookie that means the ``top level'' or ``root'' of current zone. You can list servers in the zone's top-level ``@'' NS records that are not in the parent's NS delegation, but you cannot list servers in the parent's delegation that are not present in the zone's ``@''. Any servers listed in the NS records must be configured as authoritative (either primary or secon- dary) for the zone. If a server listed in a NS record is not authoritative, it will respond with a ``lame delega- tion'' when queried. 4. Types of Servers Servers do not really have ``types''. A server can be a primary for some zones and a secondary for others, or it can be only a primary, or only a secondary, or it can serve no zones and just answer queries via its ``cache''. Previous versions of this document referred to servers as ``master'' and ``slave'' but we now feel that those distinctions - and the assignment of a ``type'' to a name server - are not useful. 4.1. Caching Only Server All servers are caching servers. This means that the server caches the information that it receives for use until the data expires. A Caching Only Server is a server that is not authoritative for any zone. This server services queries and asks other servers, who have the authority, for the information needed. All servers keep data in their cache until the data expires, based on a TTL (``Time To Live'') field which is maintained for all resource records. 4.2. Remote Server A Remote Server is an option given to people who would like to use a name server from their workstation or on a machine that has a limited amount of memory and CPU cycles. With this option you can run all of the networking programs that use the name server without the name server running on the local machine. All of the queries are serviced by a name server that is running on another machine on the network. A host which has an /etc/resolv.conf file listing only remote hosts, and which does not run a name server of its own, is sometimes called a Remote Server (because the actual server is remote?) but more often it is called Name Server Operations Guide for BIND SMM:10-5 simply a DNS Client. This kind of host is technically not a ``server'', since it has no cache and does not answer queries. 4.3. Slave Server A Slave Server is a server that always forwards queries it cannot satisfy from its cache, to a fixed list of forwarding servers instead of interacting with the name servers for the root and other domains. The queries to the forwarding servers are recursive queries. There may be one or more forwarding servers, and they are tried in turn until the list is exhausted. A Slave and forwarder configuration is typically used when you do not wish all the servers at a given site to interact with the rest of the Internet servers. A typical scenario would involve a number of workstations and a departmental timesharing machine with Internet access. The workstations might be administratively prohibited from having Internet access. To give the workstations the appearance of access to the Internet domain system, the workstations could be Slave servers to the timesharing machine which would forward the queries and interact with other name servers to resolve the query before return- ing the answer. An added benefit of using the for- warding feature is that the central machine develops a much more complete cache of information that all the workstations can take advantage of. The use of Slave mode and forwarding is discussed further under the description of the named bootfile commands. There is no prohibition against declaring a server to be a slave even though it has primary and/or secondary zones as well; the effect will still be that anything in the local server's cache or zones will be answered, and anything else will be forwarded using the forwarders list. 5. Files The name server uses several files to load its data base. This section covers the files and their formats needed for named. 5.1. Boot File This is the file that is first read when named starts up. This tells the server what type of server it is, which zones it has authority over and where to get its initial data. The default location for this file is /etc/named.boot. However this can be changed SMM:10-6 Name Server Operations Guide for BIND by setting the BOOTFILE variable when you compile named or by specifying the location on the command line when named is started up. 5.1.1. Domain A default domain may be specified for the name server using a line such as domain Berkeley.Edu Older name servers use this information when they receive a query for a name without a ``.'' that is not known. Newer designs assume that the resolver library will append its own idea of a ``default domain'' to any unqualified names. Though the name server can still be compiled with support for the domain directive in the boot file, the default is to leave it out and we strenuously recommend against its use. If you use this feature, clients outside your local domain which send you requests about unqualified names will have the implicit qualification of your domain rather than theirs. The proper place for this function is on the client, in their /etc/resolv.conf (or equivalent) file. Use of the domain directive in your boot file is strongly discouraged. 5.1.2. Directory The directory directive specifies the direc- tory in which the name server should run, allowing the other file names in the boot file to use rela- tive path names. There can be only one directory directive and it should be given before any other directives that specify file names. directory /var/named If you have more than a couple of named files to be maintained, you may wish to place the named files in a directory such as /var/named and adjust the directory command properly. The main purposes of this command are to make sure named is in the proper directory when trying to include files by relative path names with $INCLUDE and to allow named to run in a location that is reasonable to dump core if it feels the urge. 5.1.3. Primary Service The line in the boot file that designates the server as a primary master server for a zone looks Name Server Operations Guide for BIND SMM:10-7 as follows: primary Berkeley.Edu ucbhosts The first field specifies that the server is a pri- mary one for the zone stated in the second field. The third field is the name of the file from which the data is read. The above assumes that the zone you are speci- fying is a class IN zone. If you wish to designate a different class you can append /class to the first field, where class is either the integer value or the standard mnemonic for the class. For example the line for a primary server for a hesiod class zone looks as follows: primary/HS Berkeley.Edu hesiod.data Note that this support for specifying other than class IN zones is a compile-time option which your vendor may not have enabled when they built your operating system. 5.1.4. Secondary Service The line for a secondary server is similar to the primary except that it lists addresses of other servers (usually primary servers) from which the zone data will be obtained. secondary Berkeley.Edu 128.32.0.10 128.32.0.4 ucbhosts.bak The first field specifies that the server is a secondary server for the zone stated in the second field. The two network addresses specify the name servers which have data for the zone. Note that at least one of these will be a primary, and, unless you are using some protocol other than IP/DNS for your zone transfer mechanism, the others will all be other secondary servers. Having your secondary server pull data from other secondary servers is usually unwise, since you can add delay to the pro- pagation of zone updates if your network's connec- tivity varies in pathological but common ways. The intended use for multiple addresses on a secondary declaration is when the primary server has multiple network interfaces and therefore multiple host addresses. The secondary server gets its data across the network from one of the listed servers. The server addresses are tried in the order listed. If a filename is present after the list of primary servers, data for the zone will be dumped into that SMM:10-8 Name Server Operations Guide for BIND file as a backup. When the server is first started, the data is loaded from the backup file if possible, and a primary server is then consulted to check that the zone is still up-to-date. Note that listing your server as a secondary server does not necessarily make it one - the parent zone must delegate authority to your server as well as the primary and the other secondaries, or you will be transferring a zone over for no reason; no other server will have a reason to query you for that zone unless the parent zone lists you as a server for the zone. As with primary you may specify a secondary server for a class other than IN by appending /class to the secondary keyword, e.g., secondary/HS. 5.1.5. Stub Service The line for a stub server is similar to a secondary. (This feature is experimental as of 4.9.3.) stub Berkeley.Edu 128.32.0.10 128.32.0.4 ucbhosts.bak The first field specifies that the server is a stub server for the zone stated in the second field. Stub zones are intended to ensure that a pri- mary for a zone always has the correct NS records for children of that zone. If the primary is not a secondary for a child zone it should be configured with stub zones for all its children. Stub zones provide a mechanism to allow NS records for a zone to be specified in only one place. primary CSIRO.AU csiro.dat stub dms.CSIRO.AU 130.155.16.1 dms.stub stub dap.CSIRO.AU 130.155.98.1 dap.stub 5.1.6. Cache Initialization All servers, including ``caching only'' servers, should have a line as follows in the boot file to prime the name servers cache: cache . root.cache Do not put anything into your cache files other than root server information. Name Server Operations Guide for BIND SMM:10-9 All cache files listed will be read in at named boot time and any values still valid will be reinstated in the cache. The root name server information in the cache files will be used until a root query is actually answered by one of the name servers in the cache file, after which that answer will be used instead of the cache file until the answer times out. As with primary and secondary, you may specify a secondary server for a class other than IN by appending /class to the cache keyword, e.g., class/HS. 5.1.7. Forwarders Any server can make use of forwarders. A for- warder is another server capable of processing recursive queries that is willing to try resolving queries on behalf of other systems. The forwarders command specifies forwarders by internet address as follows: forwarders 128.32.0.10 128.32.0.4 There are two main reasons for wanting to do so. First, some systems may not have full network access and may be prevented from sending any IP packets into the rest of the Internet and therefore must rely on a forwarder which does have access to the full net. The second reason is that the for- warder sees a union of all queries as they pass through its server and therefore it builds up a very rich cache of data compared to the cache in a typical workstation name server. In effect, the forwarder becomes a meta-cache that all hosts can benefit from, thereby reducing the total number of queries from that site to the rest of the net. The effect of ``forwarders'' is to prepend some fixed addresses to the list of name servers to be tried for every query. Normally that list is made up only of higher-authority servers discovered via NS record lookups for the relevant domain. If the forwarders do not answer, then unless the slave directive was given, the appropriate servers for the domains will be queried directly. 5.1.8. Slave Servers Slave mode is used if the use of forwarders is the only possible way to resolve queries due to SMM:10-10 Name Server Operations Guide for BIND lack of full net access or if you wish to prevent the name server from using other than the listed forwarders. Slave mode is activated by placing the simple command options forward-only in the bootfile. If this option is used, then you must specify forwarders. When in slave mode, the server will forward each query to each of the the forwarders until an answer is found or the list of forwarders is exhausted. The server will not try to contact any remote name server other than those named in the forwarders list. So while forwarders prepends addresses to the ``server list'' for each query, options forward- only causes the ``server list'' to contain only those addresses listed in the forwarders declara- tions. Careless use of the options forward-only directive can cause really horrible forwarding loops, since you could end up forwarding queries only to some set of hosts which are also slaves, and one or several of them could be forwarding queries back to you. Use of the options forward-only directive should be considered very carefully. Note that this same behaviour can be achieved using the deprecated directive, slave. 5.1.9. Nonrecursive Servers BIND's separation of authoritative (zone) and nonauthoritiative (cache) data has always been somewhat weak, and pollution of the former via the latter has been known to occur. One way to prevent this, as well as to save memory on servers carrying a lot of authoritative data (e.g., root servers) is to make such servers ``nonrecursive.'' This can be achieved via the directive options no-recursion in the bootfile. A server with this option enabled will not attempt to fetch data to help answer queries - if you ask it for data it does not have, it will send you a referral to a more authoritative server or, if it is itself authoritative for the zone of the query, it will send you an negative answer. Name Server Operations Guide for BIND SMM:10-11 A nonrecursive server can be named in an NS RR but it cannot be listed in the resolv.conf file. 5.1.10. Query Logging If the file system containing your syslog file has quite a bit of space, you can consider using the options query-log directive in your bootfile. This will cause your name server to log every query it receives, which when combined with a Perl or AWK script to postpro- cess the logs, can be a useful management tool. 5.1.11. Inverse Query Pseudosupport BIND by default does not support inverse queries, and this has been known to cause problems for certain microcomputer operating systems and for older versions of BIND's nslookup tool. You may decide that rather than answering with ``operation not implemented,'' named should detect the most common inverse queries and answer them with bogus information. It is better to upgrade your clients to stop depending on inverse queries, but if that is not possible, you should use the options fake-iquery directive in your bootfile. NOTE: the responses are in fact bogus, in that they contain ISO8859 square brackets ([ and ]), so your clients will not be able to do anything useful with these responses. It has been observed that no client ever did any- thing useful with real inverse query responses, either. 5.1.12. Setting Name Server Limits Some name server operations can be quite resource intensive, and in order to tune your sys- tem properly it is sometimes necessary to change BIND's internal quotas. This is accomplished via limit directives in the bootfile. Limits, and their default values, are as follows: SMM:10-12 Name Server Operations Guide for BIND limit transfers-in 10 This is the number of simultaneous named-xfer processes BIND is willing to start. Higher numbers yield faster convergence to primary servers if your secondary server has hundreds or thousands of zones to maintain, but setting this number too high can cause thrashing due to starvation of resources such as network bandwidth or swap space. NOTE: this limit can also be expressed via the deprecated directive max-fetch NN. limit transfers-per-ns 2 This is the number of simultaneous named-xfer processes BIND is willing to initiate to any given name server. In most cases, you should not need to change it. If your secondary server is pulling hundreds or thousands of zones from a single pri- mary server, increasing transfers-per-ns may speed convergence. It should be kept as small as possi- ble, to avoid causing thrashing and resource star- vation on the primary server. limit datasize Most systems have a quota that limits the size of the so-called ``data segment,'' which is where BIND keeps all of its authority and cache data. BIND will behave suboptimally (perhaps even exiting) if it runs up against this quota. If your system sup- ports a system call to change this quota for a given process, you can ask BIND to use that system call via the limit datasize NN directive. The value given here may be scaled by postfixing k for 1024X, m for (1024^2)X, and g for (1024^3)X. In 1995, the root servers all use limit datasize 64m. 5.1.13. Zone Transfer Restrictions It may be the case that your organization does not wish to give complete lists of your hosts to anyone on the Internet who can reach your name servers. While it is still possible for people to ``iterate'' through your address range, looking for PTR records, and build a list of your hosts the ``slow'' way, it is still considered reasonable to restrict your export of zones via the zone transfer protocol. To limit the list of neighbors who can transfer zones from your server, use the xfrnets directive. Name Server Operations Guide for BIND SMM:10-13 This directive has the same syntax as forward- ers except that you can list network numbers in addition to host addresses. For example, you could add the directive xfrnets 16.0.0.0 if you wanted to permit only hosts on Class A net- work number 16 to transfer zones from your server. This is not nearly granular enough, and a future version of BIND will permit such access-control to be specified on a per-host basis rather than the current per-net basis. Note that while addresses without explicit masks are assumed by this direc- tive to be networks, you can specify a mask which is as granular as you wish, perhaps including all bits of the address such that only a single host is given transfer permission. For example, consider xfrnets 16.1.0.2&255.255.255.255 which would permit only host 16.1.0.2 to transfer zones from you. Note that no spaces are allowed surrounding the ``&'' character that introduces a netmask. The xfrnets directive may also be given as tcplist for compatibility with interim releases of BIND 4.9. 5.1.14. Sorting Addresses If there are multiple addresses available for a name server which BIND wants to contact, BIND will try the ones it believes are ``closest'' first. ``Closeness'' is defined in terms of similarity-of-address; that is, if one address is on the same subnet as some interface of the local host, then that address will be tried first. Fail- ing that, an address which is on the same network will be tried first. Failing that, they will be tried in a more-or-less random order unless the sortlist directive was given in the named.boot file. sortlist has a syntax similar to forwarders, xfrnets, and bogusns - you give it a list of dotted-quad networks and it uses these to ``prefer'' some remote name server addresses over others. If no explicit mask is provided with each element of a sortlist, one will be inferred based on the high order address bits. SMM:10-14 Name Server Operations Guide for BIND If you are on a Class C net which has a Class B net between you and the rest of the Internet, you could try to improve the name server's luck in get- ting answers by listing the Class B network's number in a sortlist directive. This should have the effect of trying ``closer'' servers before the more ``distant'' ones. Note that this behaviour is new as of BIND 4.9. The other and older effect of the sortlist directive is to cause BIND to sort the A records in any response it generates, so as to put those which appear on the sortlist earlier than those which do not. This is not as helpful as you might think, since many clients will reorder the A records either at random or using LIFO; also, consider the fact that the server won't be able to guess the client's network topology, and so will not be able to accurately order for ``closeness'' to all possi- ble clients. Doing the ordering in the resolver is clearly superior. In actual practice, this directive is used only rarely since it hardwires information which changes rapidly; a network which is ``close'' today may be ``distant'' next month. Since BIND builds up a cache of the remote name servers' response times, it will quickly converge on ``reasonable'' behaviour, which isn't the same as ``optimal'' but it's close enough. Future directions for BIND include choosing addresses based on local interface metrics (on hosts that have more than one) and perhaps on routing table information. We do not intend to solve the generalized ``multihomed host'' problem, but we should be able to do a little better than we're doing now. Likewise, we hope to see a higher level resolver library that sorts responses using topology information that only exists on the client's host. 5.1.15. Bogus Name Servers It happens occasionally that some remote name server goes ``bad''. You can tell your name server to refuse to listen to or ask questions of certain other name servers by listing them in a bogusns directive in your named.boot file. Its syntax is the same as forwarders, xfrnets, and sortlist - you just give it a list of dotted-quad Internet addresses. Note that zones delegated to such servers will not be reachable from clients of your servers; thus you should use this directive Name Server Operations Guide for BIND SMM:10-15 sparingly or not at all. 5.1.16. Segmented Boot Files If you are secondary for a lot of zones, you may find it convenient to split your named.boot file into a static portion which hardly ever changes (directives such as directory, sortlist, xfrnets and cache could go here), and dynamic por- tions that change frequently (all of your primary directives might go in one file, and all of your secondary directives might go in another file - and either or both of these might be fetched automati- cally from some neighbor so that they can change your list of secondary zones without requiring your active intervention). You can accomplish this via the include directive, which takes just a single file name as its argument. No quotes are needed around the file name. The file name will be evaluated after the name server has changed its working directory to that specified in the direc- tory directive, so you can use relative pathnames if your system supports them. 5.2. Resolver Configuration The configuration file's name is /etc/resolv.conf. This file designates the name servers on the network that should be sent queries. The resolver will try to contact a name server on the localhost if it cannot find its configuration file. You should install the configuration file on every host anyway, since this is the only recommended way to specify a system-level default domain, and you can still list the local host's address if it runs a name server. It is considered reasonable to create this file even if you run a local server, since its con- tents will be cached by each client of the resolver library when the client makes its first call to a resolver routine. The resolv.conf file contains directives, one per line, of the following forms: ; comment # another comment domain local-domain search search-list nameserver server-address sortlist sort-list options option-list SMM:10-16 Name Server Operations Guide for BIND Exactly one of the domain or search directives should be given, exactly once. If the search directive is given, the first item in the given search-list will override any previously-specified local-domain. The nameserver directive may be given up to three times; additional nameserver directives will be ignored. Comments may be given by starting a line with a ``;'' or ``#''; note that comments were not permitted in versions of the resolver earlier than the one included with BIND 4.9 - so if your vendor's resolver supports comments, you know they are really on the ball. The local-domain will be appended to any query- name that does not contain a ``.''. local-domain can be overridden on a per-process basis by setting the LOCALDOMAIN environment variable. Note that local- domain processing can be disabled by setting an option in the resolver. The search-list is a list of domains which are tried, in order, as qualifying domains for query-names which do not contain a ``.''. Note that search-list processing can be disabled by setting an option in the resolver. Also note that the environment variable ``LOCALDOMAIN'' can override this search-list on a per-process basis. The server-address's are aggregated and then used as the default destination of queries generated through the resolver. In other words, this is the way you tell the resolver which name servers it should use. It is possible for a given client application to override this list, and this is often done inside the name server (which is itself a resolver client) and in test programs such as nslookup. Note that if you wish to list the local host in your resolver configuration file, you should probably use its primary Internet address rather than a local-host alias such as 127.0.0.1 or 0.0.0.0. This is due to a bug in the handling of connected SOCK_DGRAM sockets in some ver- sions of the BSD networking code. If you must use an address-alias, you should prefer 0.0.0.0 (or simply ``0'') over 127.0.0.1, though be warned that depending on the vintage of your BSD-derived networking code, both of them are capable of failing in their own ways. If your host's IP implementation does not create a short-circuit route between the default interface and the loopback interface, then you might also want to add a static route (eg. in /etc/rc.local) to do so: route add myhost.domain.name localhost 1 Name Server Operations Guide for BIND SMM:10-17 The sort-list is a list of IP address, netmask pairs. Addresses returned by gethostbyname are sorted to the order specified by this list. Any addresses that do not match the address netmask pair will be returned after those that do. The netmask is optional and the natural netmask will be used if not specified. The option-list is a list of options which each override some internal resolver variable. Supported options at this time are: debug sets the RES_DEBUG bit in _res.options. ndots:n sets the lower threshold (measured in ``number of dots'') on names given to res_query() such that names with more than this number of dots will be tried as absolute names before any local-domain or search-list processing is done. The default for this internal variable is ``1''. 5.3. Cache Initialization File 5.3.1. root.cache The name server needs to know the servers that are the authoritative name servers for the root domain of the network. To do this we have to prime the name server's cache with the addresses of these higher authorities. The location of this file is specified in the boot file. This file uses the Standard Resource Record Format (aka. Masterfile Format) covered further on in this paper. 5.4. Domain Data Files There are two standard files for specifying the data for a domain. These are hosts and host.rev. These files use the Standard Resource Record Format covered later in this paper. Note that the file names are arbitrary; many network administrators prefer to name their zone files after the domains they contain, especially in the average case which is where a given server is primary and/or secondary for many different zones. 5.4.1. hosts This file contains all the data about the machines in this zone. The location of this file SMM:10-18 Name Server Operations Guide for BIND is specified in the boot file. 5.4.2. hosts.rev This file specifies the IN-ADDR.ARPA domain. This is a special domain for allowing address to name mapping. As internet host addresses do not fall within domain boundaries, this special domain was formed to allow inverse mapping. The IN- ADDR.ARPA domain has four labels preceding it. These labels correspond to the 4 octets of an Internet address. All four octets must be specified even if an octet contains zero. The Internet address 128.32.0.4 is located in the domain 4.0.32.128.IN-ADDR.ARPA. This reversal of the address is awkward to read but allows for the natural grouping of hosts in a network. 5.4.3. named.local This file specifies the PTR record for the local loopback interface, better known as localhost, whose network address is 127.0.0.1. The location of this file is specified in the boot file. It is vitally important to the proper opera- tion of every name server that the 127.0.0.1 address have a PTR record pointing back to the name ``localhost.''. The name of this PTR record is always ``1.0.0.127.IN-ADDR.ARPA''. This is neces- sary if you want your users to be able to use hostname-authentication (hosts.equiv or ~/.rhosts) on the name ``localhost''. As implied by this PTR record, there should be a ``localhost.my.dom.ain'' A record (with address 127.0.0.1) in every domain that contains hosts. ``localhost.'' will lose its trailing dot when 1.0.0.127.in-addr.arpa is queried for; then, the DEFNAMES and/or DNSRCH resolver options will cause ``localhost'' to be evaluated as a host name in the local domain, and that means the top domains (or ideally, every domain) in your resolver's search path had better have something by that name. 5.5. Standard Resource Record Format The records in the name server data files are called resource records. The Standard Resource Record Format (RR) is specified in RFC1035. The following is a general description of these records: {name} {ttl} addr-class Record Type Record Specific data Resource records have a standard format shown above. Name Server Operations Guide for BIND SMM:10-19 The first field is always the name of the domain record and it must always start in column 1. For all RR's other than the first in a file, the name may be left blank; in that case it takes on the name of the previous RR. The second field is an optional time to live field. This specifies how long this data will be stored in the data base. By leaving this field blank the default time to live is specified in the Start Of Authority resource record (see below). The third field is the address class; currently, only one class is supported: IN for internet addresses and other internet information. Limited support is included for the HS class, which is for MIT/Athena ``Hesiod'' information. The fourth field states the type of the resource record. The fields after that are dependent on the type of the RR. Case is preserved in names and data fields when loaded into the name server. All comparisons and lookups in the name server data base are case insensitive. The following characters have special meanings: ``.'' A free standing dot in the name field refers to the root domain. ``@'' A free standing @ in the name field denotes the current origin. ``\X'' Where X is any character other than a digit (0- 9), quotes that character so that its special meaning does not apply. For example, ``\.'' can be used to place a dot character in a label. ``\DDD'' Where each D is a digit, is the octet correspond- ing to the decimal number described by DDD. The resulting octet is assumed to be text and is not checked for special meaning. ``( )'' Parentheses are used to group data that crosses a line. In effect, line terminations are not recog- nized within parentheses. (At present, this notation only works for SOA RR's and is not optional.) ``;'' Semicolon starts a comment; the remainder of the line is ignored. Note that a completely blank line is also considered a comment, and ignored. SMM:10-20 Name Server Operations Guide for BIND ``*'' An asterisk signifies wildcarding. Note that this is just another data character whose special meaning comes about only during internal name server search operations. Wildcarding is only meaningful for some RR types (notably MX), and then only in the name field - not in the data fields. Anywhere a name appears - either in the name field or in some data field defined to contain names - the current origin will be appended if the name does not end in a ``.''. This is useful for appending the current domain name to the data, such as machine names, but may cause problems where you do not want this to happen. A good rule of thumb is that, if the name is not in the domain for which you are creating the data file, end the name with a ``.''. 5.5.1. $INCLUDE An include line begins with $INCLUDE, starting in column 1, and is followed by a file name, and, optionally, by a new temporary $ORIGIN to be used while reading this file. This feature is particu- larly useful for separating different types of data into multiple files. An example would be: $INCLUDE /usr/local/adm/named/data/mail-exchanges The line would be interpreted as a request to load the file /usr/local/adm/named/data/mail-exchanges. The $INCLUDE command does not cause data to be loaded into a different zone or tree. This is sim- ply a way to allow data for a given primary zone to be organized in separate files. Not even the ``tem- porary $ORIGIN'' feature described above is suffi- cient to cause your data to branch out into some other zone - zone boundaries can only be introduced in the boot file. A $INCLUDE file must have a name on its first RR. That is, the first character of the first non-comment line must not be a space. The current default name in the parent file does not carry into the $INCLUDE file. 5.5.2. $ORIGIN The origin is a way of changing the origin in a data file. The line starts in column 1, and is followed by a domain origin. This seems like it could be useful for putting more then one zone into Name Server Operations Guide for BIND SMM:10-21 a data file, but that's not how it works. The name server fundamentally requires a given zone to map entirely to some specific file. You should there- fore be very careful to use $ORIGIN only once at the top of a file, or, within a file, to change to a ``lower'' domain in the zone - never to some other zone altogether. 5.5.3. SOA - Start Of Authority name {ttl} addr-class SOA Origin Person in charge @ IN SOA ucbvax.Berkeley.Edu. kjd.ucbvax.Berkeley.Edu. ( 1995122103 ; Serial 10800 ; Refresh 1800 ; Retry 3600000 ; Expire 259200 ) ; Minimum The Start of Authority, SOA, record designates the start of a zone. The name is the name of the zone and is often given as ``@'' since this is always the current $ORIGIN and the SOA RR is usually the first record of the primary zone file. Origin is the name of the host on which this data file resides (in other words, the primary master server for this zone.) Person in charge is the e-mail address for the person responsible for the name server, with ``@'' changed to a ``.''. The serial number is the version number of this data file and must be a positive integer. This number must be incremented whenever a change is made to the data. Older servers permitted the use of a phantom ``.'' in this and other numbers in a zone file; the mean- ing of n.m was ``n000m'' rather than the more intuitive ``n*1000+m'' (such that 1.234 translated to 1000234 rather than to 1234). This feature has been deprecated due to its obscurity, unpredicta- bility, and lack of necessity. Note that using a ``YYYYMMDDNN'' notation you can still make 100 changes per day until the year 4294. You should choose a notation that works for you. If you're a clever perl programmer you could even use RCS ver- sion numbers to help generate your zone serial numbers. The refresh indicates how often, in seconds, the secondary name servers are to check with the primary name server to see if an update is needed. The retry indicates how long, in seconds, a secondary server should wait before retrying a failed zone transfer. Expire is the upper limit, in seconds, that a secondary name server is to use the data before it expires for lack of getting a SMM:10-22 Name Server Operations Guide for BIND refresh. Minimum is the default number of seconds to be used for the Time To Live field on resource records which do not specify one in the zone file. It is also an enforced minimum on Time To Live if it is specified on some resource record (RR) in the zone. There must be exactly one SOA record per zone. 5.5.4. NS - Name Server {name} {ttl} addr-class NS Name servers name IN NS ucbarpa.Berkeley.Edu. The Name Server record, NS, lists a name server responsible for a given domain, creating a delega- tion point and a subzone. The first name field specifies the zone that is serviced by the name server specified by the second name. Every zone needs at least two name servers. Name Server Operations Guide for BIND SMM:10-23 5.5.5. A - Address {name} {ttl} addr-class A address ucbarpa IN A 128.32.0.4 IN A 10.0.0.78 The Address record, A, lists the address for a given machine. The name field is the machine name and the address is the network address. There should be one A record for each address of the machine. 5.5.6. HINFO - Host Information {name} {ttl} addr-class HINFO Hardware OS IN HINFO VAX-11/780 UNIX Host Information resource record, HINFO, is for host specific data. This lists the hardware and operating system that are running at the listed host. If you want to include a space in the machine name you must quote the name (using ``"'' characters.) There could be one HINFO record for each host, though for security reasons most domains don't have any HINFO records at all. No applica- tion depends on them. 5.5.7. WKS - Well Known Services {name} {ttl} addr-class WKS address protocol list of services IN WKS 128.32.0.10 UDP who route timed domain IN WKS 128.32.0.10 TCP ( echo telnet discard sunrpc sftp uucp-path systat daytime netstat qotd nntp link chargen ftp auth time whois mtp pop rje finger smtp supdup hostnames domain nameserver ) The Well Known Services record, WKS, describes the well known services supported by a particular protocol at a specified address. The list of services and port numbers come from the list of services specified in /etc/services. There should be only one WKS record per protocol per address. Note that RFC1123 says of WKS records: 2.2 Using Domain Name Service SMM:10-24 Name Server Operations Guide for BIND ... An application SHOULD NOT rely on the ability to locate a WKS record containing an accurate listing of all services at a particular host address, since the WKS RR type is not often used by Internet sites. To confirm that a service is present, simply attempt to use it. ... 5.2.12 WKS Use in MX Processing: RFC-974, p. 5 RFC-974 [SMTP:3] recommended that the domain system be queried for WKS ("Well-Known Service") records, to verify that each proposed mail target does support SMTP. Later experience has shown that WKS is not widely supported, so the WKS step in MX processing SHOULD NOT be used. ... 6.1.3.6 Status of RR Types ... The TXT and WKS RR types have not been widely used by Internet sites; as a result, an application cannot rely on the the existence of a TXT or WKS RR in most domains. 5.5.8. CNAME - Canonical Name alias {ttl} addr-class CNAME Canonical name ucbmonet IN CNAME monet The Canonical Name resource record, CNAME, speci- fies an alias or nickname for the official, or canonical, host name. This record must be the only one associated with the alias name. All other resource records must be associated with the canon- ical name, not with the nickname. Any resource records that include a domain name as their value (e.g., NS or MX) must list the canonical name, not the nickname. Similarly, a CNAME will be followed when searching for A RRs, but not for MX RRs or NS RRs or most other types of RRs. CNAMEs are allowed to point to other CNAMEs, but this is considered sloppy. Nicknames are useful when a well known host changes its name. In that case, it is usually a good idea to have a CNAME record so that people still using the old name will get to the right place. 5.5.9. PTR - Domain Name Pointer name {ttl} addr-class PTR real name 7.0 IN PTR monet.Berkeley.Edu. Name Server Operations Guide for BIND SMM:10-25 A Domain Name Pointer record, PTR, allows special names to point to some other location in the domain. The above example of a PTR record is used in setting up reverse pointers for the special IN- ADDR.ARPA domain. This line is from the example hosts.rev file. PTR records are needed by the gethostbyaddr function. Note the trailing ``.'' which prevents BIND from appending the current $ORIGIN to that domain name. 5.5.10. MX - Mail Exchange name {ttl} addr-class MX preference value mail exchange Munnari.OZ.AU. IN MX 0 Seismo.CSS.GOV. *.IL. IN MX 0 RELAY.CS.NET. Mail eXchange records, MX, are used to specify a list of hosts which are configured to receive mail sent to this domain name. Every name which receives mail should have an MX since if one is not found at the time mail is being delivered, an MX will be ``imputed'' with a cost of 0 and a destina- tion of the host itself. If you want a host to receive its own mail, you should create an MX for your host's name, pointing at your host's name. It is better to have this be explicit than to let it be imputed by remote mailers. In the first exam- ple, above, Seismo.CSS.GOV. is a mail gateway that knows how to deliver mail to Munnari.OZ.AU.. These two machines may have a private connection or use a different transport medium. The preference value is the order that a mailer should follow when there is more than one way to deliver mail to a single machine. Note that lower numbers indicate higher precedence, and that mailers are supposed to ran- domize same-valued MX hosts so as to distribute the load evenly if the costs are equal. See RFC974 for more detailed information. Wildcard names containing the character ``*'' may be used for mail routing with MX records. There are likely to be servers on the network that simply state that any mail to a domain is to be routed through a relay. Second example, above, all mail to hosts in the domain IL is routed through RELAY.CS.NET. This is done by creating a wildcard resource record, which states that *.IL has an MX of RELAY.CS.NET. Wildcard MX records are not very useful in practice, though, since once a mail mes- sage gets to the gateway for a given domain it still has to be routed within that domain and it is not currently possible to have an apparently- different set of MX records inside and outside of a SMM:10-26 Name Server Operations Guide for BIND domain. If you won't be needing any Mail Exchanges inside your domain, go ahead and use a wildcard. If you want to use both wildcard ``top-level'' and specific ``interior'' MX records, note that each specific record will have to ``end with'' a com- plete recitation of the same data that is carried in the top-level record. This is because the specific MX records will take precedence over the top-level wildcard records, and must be able to perform the top-level's if a given interior domain is to be able to receive mail from outside the gateway. Wildcard MX records are very subtle and you should be careful with them. 5.5.11. TXT - Text name {ttl} addr-class TXT string Munnari.OZ.AU. IN TXT "foo" A TXT record contains free-form textual data. The syntax of the text depends on the domain where it is found; many systems use TXT records to encode local data in a stylized format. MIT Hesiod is one such system. 5.5.12. RP - Responsible Person owner {ttl} addr-class RP mbox-domain-name TXT-domain-name franklin IN RP ben.franklin.berkeley.edu. sysadmins.berkeley.edu. The Responsible Person record, RP, identifies the name or group name of the responsible person for a host. Often it is desirable to be able to identify the responsible entity for a particular host. When that host is down or malfunctioning, you would want to contact those parties who might be able to repair the host. The first field, mbox-domain-name, is a domain name that specifies the mailbox for the responsible person. Its format in a zone file uses the DNS convention for mailbox encoding, identical to that used for the Person-in-charge mailbox field in the SOA record. In the example above, the mbox- domain-name shows the encoding for ``''. The root domain name (just ``.'') may be specified to indicate that no mailbox is available. The second field, TXT-domain-name, is a domain name for which TXT records exist. A subsequent query can be performed to retrieve the associated Name Server Operations Guide for BIND SMM:10-27 TXT resource records at TXT-domain-name. This pro- vides a level of indirection so that the entity can be referred to from multiple places in the DNS. The root domain name (just ``.'') may be specified for TXT-domain-name to indicate that no associated TXT RR exists. In the example above, ``sysadmins.berkeley.edu.'' is the name of a TXT record that might contain some text with names and phone numbers. The format of the RP record is class- insensitive. Multiple RP records at a single name may be present in the database, though they should have identical TTLs. The RP record is still experimental; not all name servers implement or recognize it. 5.5.13. AFSDB - DCE or AFS Server name {ttl} addr-class AFSDB subtype server host name toaster.com. IN AFSDB 1 jack.toaster.com. toaster.com. IN AFSDB 1 jill.toaster.com. toaster.com. IN AFSDB 2 tracker.toaster.com. AFSDB records are used to specify the hosts that provide a style of distributed service advertised under this domain name. A subtype value (analogous to the ``preference'' value in the MX record) indi- cates which style of distributed service is pro- vided with the given name. Subtype 1 indicates that the named host is an AFS (R) database server for the AFS cell of the given domain name. Subtype 2 indicates that the named host provides intra-cell name service for the DCE (R) cell named by the given domain name. In the example above, jack.toaster.com and jill.toaster.com are declared to be AFS database servers for the toaster.com AFS cell, so that AFS clients wishing service from toaster.com are directed to those two hosts for further information. The third record declares that tracker.toaster.com houses a directory server for the root of the DCE cell toaster.com, so that DCE clients that wish to refer to DCE services should consult with the host tracker.toaster.com for further information. The DCE sub-type of record is usually accompanied by a TXT record for other information specifying other details to be used in accessing the DCE cell. RFC1183 contains more detailed information on the use of this record type. SMM:10-28 Name Server Operations Guide for BIND The AFSDB record is still experimental; not all name servers implement or recognize it. 5.5.14. PX - Pointer to X.400/RFC822 mapping informa- tion name {ttl} addr-class PX prefer 822-dom X.400-dom *.ADMD-garr.X42D.it. IN PX 50 it. ADMD-garr.C-it. *.infn.it. IN PX 50 infn.it. O.PRMD-infn.ADMD-garr.C-it. *.it. IN PX 50 it. O-gate.PRMD-garr.ADMD-garr.C-it. The PX records (Pointer to X.400/RFC822 map- ping information) are used to specify address map- ping rules between X.400 O/R addresses and RFC822 style (domain-style) mail addresses. For a detailed description of the mapping process please refer to RFC1327. Mapping rules are of 3 different types: 1) mapping from X.400 to RFC822 (defined as "table 1 rules" in RFC1327) 2) mapping from RFC822 to X.400 (defined as "table 2 rules" in RFC1327) 3) encoding RFC822 into X.400 (defined as "gate table" in RFC1327) All three types of mapping rules are specified using PX Resource Records in DNS, although the name value is different: for case 1, the name value is an X.400 domain in DNS syntax, whereas for cases 2 and 3 the name value is an RFC822 domain. Refer to RFC-1664 for details on specifying an X.400 domain in DNS syntax and for the use of the X42D keyword in it. Tools are available to convert from RFC1327 tables format into DNS files syntax. Preference is analogous to the MX RR Preference parameter: it is currently advised to use a fixed value of 50 for it. 822-dom gives the RFC822 part of the mapping rules, and X.400-dom gives the X.400 part of the mapping rule (in DNS syntax). It is currently advised always to use wildcarded name values, as the RFC1327 tables specifications permit wildcard specifications only. This is to keep compatibility with existing services using static RFC1327 tables instead of DNS PX information. Specifications of mapping rules from X.400 to RFC822 syntax requires the creation of an Name Server Operations Guide for BIND SMM:10-29 appropriate X.400 domain tree into DNS, including thus specific SOA and NS records for the domain itself. Specification of mapping rules from RFC822 into X.400 can be embedded directly into the normal direct name tree. Again, refer to RFC1664 for details about organization of this structure. Tools and library routines, based on the stan- dard resolver ones, are available to retrieve from DNS the appropriate mapping rules in RFC1327 or DNS syntax. Once again, refer to RFC1664 to use the PX resource record, and be careful in coordinating the mapping information you can specify in DNS with the same information specified into the RFC1327 static tables. The PX record is still experimental; not all servers implement or recognize it. 5.6. Discussion about the TTL The Time To Live assigned to the records and to the zone via the Minimum field in the SOA record is very important. High values will lead to lower BIND network traffic and faster response time. Lower values will tend to generate lots of requests but will allow faster propagation of changes. Only changes and deletions from the zone are affected by the TTLs. Additions propagate according to the Refresh value in the SOA. Experience has shown that sites use default TTLs for their zones varying from around 0.5 day to around 7 days. You may wish to consider boosting the default TTL shown in former versions of this guide from one day (86400 seconds) to three days (259200 seconds). This will drastically reduce the number of requests made to your name servers. If you need fast propagation of changes and dele- tions, it might be wise to reduce the Minimum field a few days before the change, then do the modification itself and augment the TTL to its former value. If you know that your zone is pretty stable (you mainly add new records without deleting or changing old ones) then you may even wish to consider a TTL higher than three days. SMM:10-30 Name Server Operations Guide for BIND Note that in any case, it makes no sense to have records with a TTL below the SOA Refresh delay, as Delay is the time required for secondaries to get a copy of the newly modified zone. 5.7. About ``secure zones'' Secure zones implement named security on a zone by zone basis. It is designed to use a permission list of networks or hosts which may obtain particular information from the zone. In order to use zone security, named must be com- piled with SECURE_ZONES defined and you must have at least one secure_zone TXT RR. Unless a secure_zone record exists for a given zone, no restrictions will be applied to the data in that zone. The format of the secure_zone TXT RR is: secure_zone addr-class TXT string The addr-class may be either HS or IN. The syn- tax for the TXT string is either ``network address:netmask'' or ``host IP address:H''. ``network address:netmask'' allows queries from an entire network. If the netmask is omitted, named will use the default netmask for the network address specified. ``host IP address:H'' allows queries from a host. The ``H'' after the ``:'' is required to differen- tiate the host address from a network address. Multi- ple secure_zone TXT RRs are allowed in the same zone file. For example, you can set up a zone to only answer Hesiod requests from the masked class B network 130.215.0.0 and from host 128.23.10.56 by adding the following two TXT RR's: secure_zone HS TXT ``130.215.0.0:255.255.0.0'' secure_zone HS TXT ``128.23.10.56:H'' This feature can be used to restrict access to a Hesiod password map or to separate internal and exter- nal internet address resolution on a firewall machine without needing to run a separate named for internal and external address resolution. Note that you will need to include your loopback interface (127.0.0.1) in your secure_zone record, or Name Server Operations Guide for BIND SMM:10-31 your local clients won't be able to resolve names. 5.8. About Hesiod, and HS-class Resource Records Hesiod, developed by MIT Project Athena, is an information service built upon BIND. Its intent is similar to that of Sun's NIS: to furnish information about users, groups, network-accessible file systems, printcaps, and mail service throughout an installa- tion. Aside from its use of BIND rather than separate server code another important difference between Hesiod and NIS is that Hesiod is not intended to deal with passwords and authentication, but only with data that are not security sensitive. Hesiod servers can be implemented by adding resource records to BIND servers; or they can be implemented as separate servers separately administered. To learn about and obtain Hesiod make an anonymous FTP connection to host ATHENA-DIST.MIT.EDU and retrieve the compressed tar file /pub/ATHENA/hesiod.tar.Z. You will not need the named and resolver library portions of the distribution because their functionality has already been integrated into BIND as of 4.9. To learn how Hesiod functions as part of the Athena computing environment obtain the paper /pub/ATHENA/usenix/athena-changes.PS from the above FTP server host. There is also a tar file of sample Hesiod resource files. Whether one should use Hesiod class is open to question, since the same services can probably be pro- vided with class IN, type TXT and type CNAME records. In either case, the code and documents for Hesiod will suggest how to set up and use the service. Note that while BIND includes support for HS- class queries, the zone transfer logic for non-IN- class zones is still experimental. 5.9. Sample Files The following section contains sample files for the name server. This covers example boot files for the different types of servers and example domain data SMM:10-32 Name Server Operations Guide for BIND base files. 5.9.1. Boot Files 5.9.1.1. Primary Server ; ; Boot file for Primary Name Server ; ; type domain source file or host ; directory /usr/local/adm/named primary Berkeley.Edu ucbhosts primary 32.128.in-addr.arpa ucbhosts.rev primary 0.0.127.in-addr.arpa named.local cache . root.cache 5.9.1.2. Secondary Server ; ; Boot file for Secondary Name Server ; ; type domain source file or host ; directory /usr/local/adm/named secondary Berkeley.Edu 128.32.0.4 128.32.0.10 ucbhosts.bak secondary 32.128.in-addr.arpa 128.32.0.4 128.32.0.10 ucbhosts.rev.bak primary 0.0.127.in-addr.arpa named.local cache . root.cache 5.9.1.3. Caching Only Server Name Server Operations Guide for BIND SMM:10-33 ; ; Boot file for Caching Only Name Server ; ; type domain source file or host ; directory /usr/local/adm/named cache . root.cache primary 0.0.127.in-addr.arpa named.local 5.9.2. Remote Server / DNS Client 5.9.2.1. /etc/resolv.conf domain Berkeley.Edu nameserver 128.32.0.4 nameserver 128.32.0.10 sortlist 130.155.160.0/255.255.240.0 130.155.0.0 SMM:10-34 Name Server Operations Guide for BIND 5.9.3. root.cache ; ; This file holds the information on root name servers needed to ; initialize cache of Internet domain name servers ; (e.g. reference this file in the "cache . " ; configuration file of BIND domain name servers). ; ; This file is made available by InterNIC registration services ; under anonymous FTP as ; file /domain/named.root ; on server FTP.RS.INTERNIC.NET ; -OR- under Gopher at RS.INTERNIC.NET ; under menu InterNIC Registration Services (NSI) ; submenu InterNIC Registration Archives ; file named.root ; ; last update: Oct 5, 1994 ; related version of root zone: 1994100500 ; . 604800 IN NS NS.INTERNIC.NET. NS.INTERNIC.NET. 604800 IN A 198.41.0.4 . 604800 IN NS NS1.ISI.EDU. NS1.ISI.EDU. 604800 IN A 128.9.0.107 . 604800 IN NS C.PSI.NET. C.PSI.NET. 604800 IN A 192.33.4.12 . 604800 IN NS TERP.UMD.EDU. TERP.UMD.EDU. 604800 IN A 128.8.10.90 . 604800 IN NS NS.NASA.GOV. NS.NASA.GOV. 604800 IN A 128.102.16.10 604800 IN A 192.52.195.10 . 604800 IN NS NS.ISC.ORG. NS.ISC.ORG. 604800 IN A 192.5.5.241 . 604800 IN NS NS.NIC.DDN.MIL. NS.NIC.DDN.MIL. 604800 IN A 192.112.36.4 . 604800 IN NS AOS.ARL.ARMY.MIL. AOS.ARL.ARMY.MIL. 604800 IN A 128.63.4.82 604800 IN A 192.5.25.82 . 604800 IN NS NIC.NORDU.NET. NIC.NORDU.NET. 604800 IN A 192.36.148.17 ; End of File Name Server Operations Guide for BIND SMM:10-35 5.9.4. named.local @ IN SOA ucbvax.Berkeley.Edu. kjd.ucbvax.Berkeley.Edu. ( 1994072100 ; Serial 10800 ; Refresh 1800 ; Retry 3600000 ; Expire 259200 ) ; Minimum IN NS ucbvax.Berkeley.Edu. ; pedantic 1 IN PTR localhost. 5.9.5. host.rev ; ; @(#)ucb-hosts.rev 1.1 (Berkeley) 86/02/05 ; @ IN SOA ucbvax.Berkeley.Edu. kjd.monet.Berkeley.Edu. ( 1986020501 ; Serial 10800 ; Refresh 1800 ; Retry 3600000 ; Expire 259200 ) ; Minimum IN NS ucbarpa.Berkeley.Edu. IN NS ucbvax.Berkeley.Edu. 0.0 IN PTR Berkeley-net.Berkeley.EDU. IN A 255.255.255.0 0.130 IN PTR csdiv-net.Berkeley.EDU. 4.0 IN PTR ucbarpa.Berkeley.Edu. 6.0 IN PTR ernie.Berkeley.Edu. 7.0 IN PTR monet.Berkeley.Edu. 10.0 IN PTR ucbvax.Berkeley.Edu. 6.130 IN PTR monet.Berkeley.Edu. 5.9.6. Hosts SMM:10-36 Name Server Operations Guide for BIND ; ; @(#)ucb-hosts 1.2 (berkeley) 88/02/05 ; @ IN SOA ucbvax.Berkeley.Edu. kjd.monet.Berkeley.Edu. ( 1988020501 ; Serial 10800 ; Refresh 1800 ; Retry 3600000 ; Expire 259200 ) ; Minimum IN NS ucbarpa.Berkeley.Edu. IN NS ucbvax.Berkeley.Edu. localhost IN A 127.1 ; note that 127.1 is the same as 127.0.0.1; see inet(3n) ucbarpa IN A 128.32.4 IN A 10.0.0.78 IN HINFO VAX-11/780 UNIX arpa IN CNAME ucbarpa ernie IN A 128.32.6 IN HINFO VAX-11/780 UNIX ucbernie IN CNAME ernie monet IN A 128.32.7 IN A 128.32.130.6 IN HINFO VAX-11/750 UNIX ucbmonet IN CNAME monet ucbvax IN A 10.2.0.78 ; 128.32.10 means 128.32.0.10; see inet(3n) IN A 128.32.10 ; HINFO and WKS are widely unused, ; but we'll show them as examples. IN HINFO VAX-11/750 UNIX IN WKS 128.32.0.10 TCP ( echo telnet discard sunrpc sftp uucp-path systat daytime netstat qotd nntp link chargen ftp auth time whhois mtp pop rje finger smtp supdup hostnames domain nameserver ) vax IN CNAME ucbvax toybox IN A 128.32.131.119 IN HINFO Pro350 RT11 toybox IN MX 0 monet.Berkeley.Edu. csrg IN MX 0 Ralph.CS IN MX 0 Zhou.CS IN MX 0 Painter.CS IN MX 0 Riggle.CS IN MX 0 Terry.CS IN MX 0 Kevin.CS Name Server Operations Guide for BIND SMM:10-37 6. Setting up Your Own Domain When setting up a domain that is going to be on a public network the site administrator should contact the organization in charge of the network and request the appropriate domain registration form. An organization that belongs to multiple networks (such as the Internet and BITNET) should register with only one network. 6.1. Internet Sites on the Internet who need information on setting up a domain should contact the registrar for their network, which is one of the following: MILnet HOSTMASTER@NIC.DDN.MIL other HOSTMASTER@INTERNIC.NET You may also want to be placed on the BIND mailing list, which is a mail group for people on the Internet who run BIND. The group discusses future design deci- sions, operational problems, and other related topic. The address to request being placed on this mailing list is: bind-request@uunet.uu.net 6.2. Subdomains of Existing Domains If you want a subdomain of some existing domain, you should find the contact point for the parent domain rather than asking one of the above top-level registrars. There should be a convention that registrar@domain or hostmaster@domain for any given domain will always be an alias for that domain's registrar (somewhat analogous to postmaster), but there is no such convention. Try it as a last resort, but first you should examine the SOA record for the domain and send mail to the ``responsible person'' shown therein. You can also try whois. 7. Domain Management This section contains information for starting, con- trolling and debugging named. 7.1. /etc/rc.local The hostname should be set to the full domain style name in /etc/rc.local using hostname(1). The following entry should be added to /etc/rc.local to SMM:10-38 Name Server Operations Guide for BIND start up named at system boot time: if [ -f /usr/sbin/named ]; then /usr/sbin/named [options] & echo -n ' named' >/dev/console fi This usually directly follows the lines that start syslogd. Do Not attempt to run named from inetd. This will continuously restart the name server and defeat the purpose of the cache. 7.2. /var/run/named.pid When named is successfully started up it writes its process id into the file /var/run/named.pid. This is useful to programs that want to send signals to named. The name of this file may be changed by defin- ing PIDFILE to the new name when compiling named. 7.3. /etc/hosts The gethostbyname() library call can detect if named is running. If it is determined that named is not running it will look in /etc/hosts to resolve an address. This option was added to allow ifconfig(8C) to configure the machines local interfaces and to enable a system manager to access the network while the system is in single user mode. It is advisable to put the local machines interface addresses and a cou- ple of machine names and address in /etc/hosts so the system manager can rcp files from another machine when the system is in single user mode. The format of /etc/hosts has not changed. See hosts(5) for more information. Since the process of reading /etc/hosts is slow, it is not advisable to use this option when the system is in multi user mode. 7.4. Signals There are several signals that can be sent to the named process to have it do tasks without restarting the process. 7.4.1. Reload SIGHUP - Causes named to read named.boot and reload the database. This is useful when you have made a change to a ``primary'' data file and you want named's internal database to reflect the change. If you build BIND with the FORCED_RELOAD option, then SIGHUP also has the effect of schedul- ing all ``secondary'' zones for serial-number Name Server Operations Guide for BIND SMM:10-39 checks, which could lead to zone transfers ahead of the usual schedule. Normally serial-number com- pares are done only at the intervals specified in the zone's SOA record. 7.4.2. Debugging When named is running incorrectly, look first in /var/log/messages and check for any messages logged by syslog. Next send it a signal to see what is happening. Unless you run it with the ``- d'' option, named has very little to say on its standard output or standard error. Everything named has to say, it says to syslog. SIGINT - Dumps the current data base and cache to /var/tmp/named_dump.db This should give you an indication to whether the data base was loaded correctly. The name of the dump file may be changed by defining DUMPFILE to the new name when compiling named. Note: the following two signals only work when named is built with DEBUG defined. SIGUSR1 - Turns on debugging. Each following SIGUSR1 increments the debug level. The output goes to /var/tmp/named.run The name of this debug file may be changed by defining DEBUGFILE to the new name before compiling named. SIGUSR2 - Turns off debugging completely. For more detailed debugging, define DEBUG when com- piling the resolver routines into /lib/libc.a. SIGWINCH - Toggles tracing of all incoming queries if named has been compiled with QRYLOG defined. The trace is sent to syslog, and is huge, but it is very useful for tracking down problems. To run with tracing of all queries specify the -q flag on the command line. If you routinely log queries you will probably want to analyze the results using the dnsstats stats script in the con- trib directory. SIGIOT - Dumps statistics data into /var/tmp/named.stats if the server is built with STATS defined. Statistics are appended to the file. SMM:10-40 Name Server Operations Guide for BIND 8. Building a System with a Name Server BIND is composed of two parts. One is the user interface called the resolver which consists of a group of routines that reside in the C library /lib/libc.a. Second is the actual server called named. This is a dae- mon that runs in the background and services queries on a given network port. The standard port for UDP and TCP is specified in /etc/services. 8.1. Resolver Routines in libc When building your 4.3BSD system you may either build the C library to use the name server resolver routines or use the host table lookup routines to do host name and address resolution. The default resolver for 4.3BSD uses the name server. Newer BSD systems include both name server and host table func- tionality with preference given to the name server if there is one or if there is a /etc/resolv.conf file. Building the C library to use the name server changes the way gethostbyname(3N), gethostbyaddr(3N), and sethostent(3N) do their functions. The name server renders gethostent(3N) obsolete, since it has no concept of a next line in the database. These library calls are built with the resolver routines needed to query the name server. The resolver contains functions that build query packets and exchange them with name servers. Before building the 4.3BSD C library, set the variable HOSTLOOKUP equal to named in /usr/src/lib/libc/Makefile. You then make and install the C library and compiler and then compile the rest of the 4.3BSD system. For more information see sec- tion 6.6 of ``Installing and Operating 4.3BSD on the VAX=''. If your operating system isn't VAX= 4.3BSD, it is probably the case that your vendor has included resolver support in the supplied C Library. You should consult your vendor's documentation to find out what has to be done to enable resolver support. Note that your vendor's resolver may be out of date with respect to the one shipped with BIND, and that you might want to build BIND's resolver library and install it, and its include files, into your system's ____________________ =VAX is a Trademark of Digital Equipment Corporation Name Server Operations Guide for BIND SMM:10-41 compile/link path so that your own network applica- tions will be able to use the newer features. SMM:10-42 Name Server Operations Guide for BIND ACKNOWLEDGEMENTS - 4.9.3 The mailing list was once again of great help; this release would not be nearly as ready for prime time if not for their efforts. Special commendations are owed to Robert Elz, Don "Truck" Lewis, Bob Halley, Mark Andrews, Berthold Paffrath, Ruediger Volk, and Peter Koch. Digital Equipment Corporation, Hewlett Packard, Silicon Graphics, and SunSoft all made hardware available for integration testing; this made the release far more solid than it would otherwise have been. More hardware loans are welcome - if you are a system vendor and you would like BIND to run ``out of the box'' on your platform and are willing to lend some rusty old hardware for the purpose, please con- tact me () to make the arrangements. Special thanks to the Internet Software Consortium for funding this work. Contact if your organization would like to participate in funding future releases of BIND and other freely redistributable software packages that are in wide use on the Internet. ACKNOWLEDGEMENTS - through 4.9 The alpha-test group was extremely helpful in furnish- ing improvements, finding and repairing bugs, and being patient. I would like to express special thanks to Brian Reid of Digital Equipment corporation for funding this work. Robert Elz, Alan Barrett, Paul Albitz, Bryan Beecher, Andrew Partan, Andy Cherenson, Tom Limoncelli, Berthold Paffrath, Fuat Baran, Anant Kumar, Art Harkin, Win Treese, Don Lewis, Christophe Wolfhugel, and a cast of dozens all helped out above and beyond the call of duty. Special thanks to Phil Almquist, who got the project started and contributed a lot of the code and fixed several of the worst bugs. ACKNOWLEDGEMENTS - through 4.8.3 Many thanks to the users at U. C. Berkeley for falling into many of the holes involved with integrating BIND into the system so that others would be spared the trauma. I would also like to extend gratitude to Jim McGinness and Digital Equipment Corporation for permitting me to spend most of my time on this project. Ralph Campbell, Doug Kingston, Craig Partridge, Smoot Carl-Mitchell, Mike Muuss and everyone else on the DARPA Internet who has contributed to the development of BIND. To the members of the original BIND project, Douglas Terry, Mark Painter, David Riggle and Songnian Zhou. Name Server Operations Guide for BIND SMM:10-43 Anne Hughes, Jim Bloom and Kirk McKusick and the many others who have reviewed this paper giving considerable advice. This work was sponsored by the Defense Advanced Research Projects Agency (DoD), Arpa Order No. 4871 moni- tored by the Naval Electronics Systems Command under con- tract No. N00039-84-C-0089. The views and conclusions con- tained in this document are those of the authors and should not be interpreted as representing official policies, either expressed or implied, of the Defense Research Projects Agency, of the US Government, or of Digital Equipment Cor- poration. SMM:10-44 Name Server Operations Guide for BIND REFERENCES [Birrell] Birrell, A. D., Levin, R., Needham, R. M., and Schroeder, M.D., "Grapevine: An Exercise in Dis- tributed Computing." In Comm. A.C.M. 25, 4:260-274 April 1982. [RFC819] Su, Z. Postel, J., "The Domain Naming Convention for Internet User Applications." Internet Request For Comment 819 Network Information Center, SRI International, Menlo Park, California. August 1982. [RFC974] Partridge, C., "Mail Routing and The Domain Sys- tem." Internet Request For Comment 974 Network Information Center, SRI International, Menlo Park, California. February 1986. [RFC1032] Stahl, M., "Domain Administrators Guide" Internet Request For Comment 1032 Network Information Center, SRI International, Menlo Park, California. November 1987. [RFC1033] Lottor, M., "Domain Administrators Guide" Internet Request For Comment 1033 Network Information Center, SRI International, Menlo Park, California. November 1987. [RFC1034] Mockapetris, P., "Domain Names - Concept and Facilities." Internet Request For Comment 1034 Network Information Center, SRI International, Menlo Park, California. November 1987. [RFC1035] Mockapetris, P., "Domain Names - Implementation and Specification." Internet Request For Comment 1035 Network Information Center, SRI Interna- tional, Menlo Park, California. November 1987. [RFC1101] Mockapetris, P., "DNS Encoding of Network Names and Other Types." Internet Request For Comment 1101 Network Information Center, SRI Interna- tional, Menlo Park, California. April 1989. [RFC1123] R. Braden, Editor, "Requirements for Internet Hosts -- Application and Support" Internet Request For Comment 1123 Network Information Center, SRI International, Menlo Park, California. October 1989. [RFC1183] Everhart, C., Mamakos, L., Ullmann, R., and Mocka- petris, P., "New DNS RR Definitions" Internet Request For Comment 1183 Network Information Name Server Operations Guide for BIND SMM:10-45 Center, SRI International, Menlo Park, California. October 1990. [RFC1327] Hardcastle-Kille, S., "Mapping between X.400(1988) / ISO 10021 and RFC 822" Internet Request For Com- ment 1327 Network Information Center, SRI Interna- tional, Menlo Park, California. May 1992. [RFC1664] Allocchio, C., Bonito, A., Cole, B., Giordano, S., Hagens, R., "Using the Internet DNS to Distribute RFC1327 Mail Address Mapping Tables" Internet Request For Comment 1664 Network Information Center, SRI International, Menlo Park, California. August 1994. [RFC1713] Romao, A., "Tools for DNS debugging" Internet Request For Comment 1713, also FYI27 Network Information Center, SRI International, Menlo Park, California. November 1994. [Terry] Terry, D. B., Painter, M., Riggle, D. W., and Zhou, S., The Berkeley Internet Name Domain Server. Proceedings USENIX Summer Conference, Salt Lake City, Utah. June 1984, pages 23-31. [Zhou] Zhou, S., The Design and Implementation of the Berkeley Internet Name Domain (BIND) Servers. UCB/CSD 84/177. University of California, Berke- ley, Computer Science Division. May 1984. [Mockapetris] Mockapetris, P., Dunlap, K, Development of the Domain Name System ACM Computer Communications Review 18, 4:123-133. Proceedings ACM SIGCOMM '88 Symposium, August 1988. [Liu] Liu, C., Albitz, P., DNS and BIND O'Reilly & Asso- ciates, Sebastopol, CA, 502 pages, ISBN 0-937175- 82-X 1992