IPv6 - Maila

Note: Once you enable IPv6 on node it is automatically configured for auto-configuration. To disable auto-configuration on IPv6 enabled node, we need to use '##IPV6_AUTOCONF=no##' option in either '##/etc/sysconfig/network##' (global) or in interface configuration file (per interface).

27

28

29

30

== Configuring temporary static IPv6 addresses ==

31

32

To configure temporary static IPv6 addresses which will cease to exist if machine or network is restarted, use:

33

34

{{{ifconfig eth0 add <ipv6_address>/<netmask>

35

}}}

36

37

command. Note that we can add any number of IP addresses using this technique as in IPv6 same interface can be assigned more than one IP address.

== Removing IPv6 address assigned to a interface ==

42

43

Address prefixes learned via router will slowly get deferred and their valid lifetime would decrease, and eventually they will get removed automatically. However if we have added some address (temporary or permanent) statically then to remove it we can use:

44

45

{{{ifconfig eth0 del <ipv6_address>/<netmask>

46

}}}

47

48

* Note that if we remove some address that was learned via router, then it is possible that the address gets added again on next router advertizement message. To really ensure that address does not get added again we have to disable auto-configuration ('##sysctl net.ipv6.conf.{all|<interface>}.autoconf = 0##') or ignore all router advertisements ('##sysctl net.ipv6.conf.{all|<interface>}.accept_ra = 0 ##').

== Configuring permanent static IPv6 addresses ==

53

54

To configure permanent static IPv6 addresses on a machine which would remain even if network service is restarted:

55

56

* First enable IPv6 on machine

57

* Then edit file '##/etc/sysconfig/network-scripts/ifcfg-eth<n>##' and add lines

*:

{{{

}}}

*

*:

*:: IPV6ADDR=<ipv6_address1_without_quotes>

66

*:: IPV6ADDR_SECONDARIES="<ipv6_address2> <ipv6_address3> <ipv6_address4>"

67

*:

68

69

Note that multiple IPv6 addresses can be defined in single IPV6ADDR_SECONDARIES option each separated from one another using space.

== Disabling auto-configuration of addresses via router discovery/advertisements ==

74

75

By default nodes learn prefixes through router solicitation and advertisement messages and configure an address for each prefix learned via such messages. This also allows hosts to configure routes through which they can send packets to nodes which are not on same link.

76

77

78

=== Disabling auto-configuration on all interfaces ===

79

80

To disable auto-configuration on complete node (all interfaces), edit file '##/etc/sysconfig/network##' and add option

{{{IPV6_AUTOCONF=no

}}}

We can also use sysctl to disable auto-configuration using:

86

87

{{{sysctl net.ipv6.conf.all.autoconf=0

}}}

=== Disabling auto-configuration on particular interface ===

92

93

To disable auto-configuration on particular interface, edit interface configuration file '##/etc/sysconfig/network-scripts/ifcfg-eth<n>##' and add line:

{{{IPV6_AUTOCONF=no

}}}

We can also disable auto-configuration on particular interface using sysctl as:

99

100

{{{sysctl net.ipv6.conf.<interface_name>.autoconf=0

}}}

= Basic network service test commands =

107

108

== Pinging hosts using their link-local addresses ==

109

110

To ping a host using its link-local address we can use:

111

112

{{{ping6 -I eth<n> <link_local_address_of_remote_node>

113

}}}

114

115

Note that we need to specify interface to be used as node will have one IPv6 link local address (fe80::/10) on each interface. Hence there is no way for node to know which interface should be used to send ping requests as it has IP address in that range on all interfaces.

116

117

118

== Pinging all nodes or all routers multicast address ==

119

120

To ping all nodes on a link we can use:

121

122

{{{ping6 -I eth<n> ff02::1

123

}}}

124

125

To ping all nodes on a routers we can use:

126

127

{{{ping6 -I eth<n> ff02::2

128

}}}

129

130

Note that same as in case of pinging link local addresses, here also we need to explicitly specify which interface should be used to send the ping request to multicast addresses as multicast requests can be sent via any interface and there wont be any default route configured to send multicasts.

= Informational commands =

136

137

== Seeing neighbour hardware address cache ==

138

139

We can see learned MAC addresses through ICMPv6 using command:

{{{ip neigh show

}}}

This command will show MAC addresses learned via both IPv4 and IPv6. To just see mac addresses learned via IPv4 we can use:

{{{arp -a -n

}}}

and to see MAC addresses learned via IPv6 we can use

{{{ip -6 neigh show

}}}

== Seeing IPv6 address valid and preferred lifetimes ==

157

158

We can see list of IP addresses assigned to a node (or to a particular interface of a node) using

159

160

{{{ip addr show [<interface>]

161

}}}

162

163

instead of using '##ifconfig [<interface>]##'. The advantage of using '##ip addr show##' command over '##ifconfig##' command is that it will also show valid and preferred lifetime of all IPv6 addresses assigned to the node (or particular interface of node) which is not available in '##ifconfig##' output.

== Seeing IPv6 routing table ==

We can use command

{{{ip -6 route show

}}}

to see IPv6 routing table.

175

176

If you wnat to use route command then use:

{{{route -6 -n

}}}

to see IPv6 routes.

== Resolving IPv6 addresses using nslookup ==

185

186

By default '##nslookup##' only sends and accepts IPv4 DNS queries and replies. In order to make '##nslookup##' also request and process IPv6 queries and responses we can use:

187

188

{{{nslookup -type=any <domain_name> [<DNS_server_IP>]

189

}}}

190

191

format so that even IPv6 responses are displayed.

192

193

Instead of nslookup you can also use:

194