ArtsAutosBooksBusinessEducationEntertainmentFamilyFashionFoodGamesGenderHealthHolidaysHomeHubPagesPersonal FinancePetsPoliticsReligionSportsTechnologyTravel

DNS CONFIGURATION FOR IPV6

Updated on April 22, 2012

Following section provides an in depth description of IPv6 configuration in BIND. All the details and examples are given assuming the reader has the basic knowledge on DNS, BIND and BIND configuration etc. In order to have the IPv6 capability in our DNS server we need to do several configuration changes. With the installation of BIND, all the DNS data will be located in /var/named folder. It contains the zone files and other related data.

Following basic changes and enhancements should be done in the /etc/named.conf boot file in order to operate over IPv6.

  • Configure the server to listen for incoming queries sent using IPv6.

listen-on-v6 { any; };

Server will listen on default port 53 for any IPv6 address.

listen-on-v6 port 2000 { 2001:df0:17:c00:1::2222; };

Server will listen on port 2000 for only the specified address.

If no listen-on-v6 option is specified, the server will not listen on any IPv6 address unless -6 is specified when named is invoked. If -6 option is specified, then named will listen on port 53 on all IPv6 interfaces by default.

  • Set the other options as same for IPv4 by only replacing the IPv4 addresses with preferred IPv6 addresses.

allow-query { 2001:df0:17:c00::/64; 10.8.106.0/24;};

allow-transfer { 2001:df0:17:c00::/64; };

allow-recursion { 2001:df0:17:c00::/64; };

  • zone transfers

!The zone transfer is initiated by the slave server. Example configuration is shown below for a slave server in order to fetch zone files for the zone “uom.lk” form the master server 2001:df0:17:e00::1.

zone "uom.lk" {

type slave;

masters { 2001:df0:17:e00::1; };

file " uom.lk.db";

transfer-source-v6 { 2001:df0:17:e00::4444; }; };

Following configuration option needs to be set in the slave server in order to fetch zones transferred inbound by the server. In the master server (here 2001:df0:17:e00::1), address of the slave server should be specified within the allow-transfer statement.

allow-transfer{2001:df0:17:d00::2222;};

  • Current ACL statements should be modified to include ipv6 addresses.


Forward resolution

For the locally administered domains, we need to add following type of zone details in the named.conf file. General configuration procedure for IPv4 is not affected by IPv6 for this process.

zone "ipv6.uom.lk” {

type master;

file "ipv6.uom.lk.db";

};

For each locally administrated zone, zones file need to be created in the /var/named/ directory as follows.

$TTL 86400

@ IN SOA ns.ipv6.uom.lk. admin.ipv6.uom.lk. (

20100212 ; Serial

604800 ; Refresh

86400 ; Retry 2419200 ; Expire

604800 ) ; Negative Cache TTL

;; DNS server

@ IN NS ns.ipv6.uom.lk. ;;dual stack nodes www IN A 192.168.8.200 IN AAAA 2001:df0:17:e00::4:2222 Cache IN A 192.168.8.201 IN AAAA 2001:df0:17:e00::4:3333 Lms IN A 192.168.8.202 IN AAAA 2001:df0:17:e00::4:4444 ;;ipv6 only test lab www6 IN AAAA 2001:df0:17:e00::1111 lms6 IN AAAA 2001:df0:17:e00::2222 pc1 IN AAAA 2001:df0:17:e00::3331 pc2 IN AAAA 2001:df0:17:e00::3332 pc3 IN AAAA 2001:df0:17:e00::3334

Reverse resolution

Configuring the BIND nameserver for reverse resolution is much troublesome task due to the longer IPv6 addresses. For each domain the network address should reversed digit-wise.

Configuration for the 2001:df0:17:e10::/64

;;for 2001:df0:17:e10::/64

zone "0.0.e.0.7.1.0.0.0.f.d.0.1.0.0.2.ip6.arpa" {

type master;

file "2001:df0:17:e10.db";

};

Corresponding reverse zone file will be as follows,

$TTL 86400 @ IN SOA ns.ipv6.uom.lk. admin.ipv6.uom.lk. ( 20100212 ; Serial 604800 ; Refresh 86400 ; Retry 2419200 ; Expire 604800 ) ; Negative Cache TTL @ IN NS ns.ipv6.uom.lk. $ORIGIN 8.248.192.in-addr.arpa 200 IN PTR www.ipv6.uom.lk. 201 IN PTR cache.ipv6.uom.lk. 202 IN PTR lms.ipv6.uom.lk.

;; test ipv6 lab $ORIGIN 0.0.e.0.7.1.0.0.0.f.d.0.1.0.0.2.ip6.arpa.

1.1.1.1.0.0.0.0.0.0.0.0.0.0.0.0 IN PTR www6.ipv6.uom.lk

. 2.2.2.2.0.0.0.0.0.0.0.0.0.0.0.0 IN PTR lms6.ipv6.uom.lk.

3.3.3.3.0.0.0.0.0.0.0.0.0.0.0.0 IN PTR pc1.ipv6.uom.lk.

pure ipv6 name resolution

For a pure IPv6 name resolution process, all the nodes from the root name servers to the final authoritative name server should be IPv6 supported. But in the current scenario this is not yet achieved for most of the cases.

IPv6 support in the root name servers is not fully implemented. Only 7 of the 13 root servers are ipv6 capable.

In the internet there are three types of DNS servers.

1. IPv4 only (only reachable by IPv4)

2. IPv6 only (only reachable by IPv6)

3. Dual stack (reachable by both IPv4 and IPv6)

According to the DNS IPv6 Transport Operational Guidelines presented in RFC3901, it suggests all the recursive name servers to be either IPv4 only or dual-stack. It also suggests that each DNS zone need to have at least one IPv4-reachable DNS server (which is the current situation). For the dual stack name resolution to be consistent this must be adhered. We cannot have IPv6 reachable server for each domain.

The reason for this suggestion is to avoid the fragmentation of the name space. Consider a pure IPv6 name resolution process by a recursive DNS server. It may get halted if it encounters a DNS server which can be reached only in IPv4. Same thing would happen for the pure IPv4 name resolution in the transition period.

So in most of the cases pure IPv6 name resolution is halted without completing the required resolution. So in the initial phase of the transition we need to adhere to the above best practice.


Comments

    0 of 8192 characters used
    Post Comment

    No comments yet.

    Click to Rate This Article