Questions[radvd](server)

I've been trying to get IPv6 PD with dhcpcd working on my home server along with ppp and radvd to allow for SLAAC to work and eventually replace my ISP-provided router. Unfortunately, the prefix which dhcpcd receives on the ppp0 interface via RAs only gets configured as an address on that interface, instead of also being delegated to the lan interface on which radvd is operating.

What am I missing?

/etc/dhcpcd.conf:

# Ignore changes on docker interfaces
denyinterfaces veth*

duid

# Persist interface configuration when dhcpcd exits.
persistent

option domain_name_servers, domain_name, domain_search, host_name
option interface_mtu

require dhcp_server_identifier

slaac private

noipv6rs
waitip 6
ipv6only

#Prefix Delegation
interface ppp0
option rapid_commit
ipv6rs
iaid 1
ia_pd 1/::/64 lan/0/64

#don't use ISP DNS servers
nohook resolv.conf

/etc/radvd.conf:

interface lan
{
    AdvSendAdvert on;
    prefix ::/64
    {
    AdvOnLink on;
         AdvAutonomous on;
         AdvRouterAddr on;
    };
};

The lan and wan.7 (my ISP requires VLAN 7, so this is what ppp binds to) interfaces are managed by systemd-networkd (DHCP is not handled by it though), but ppp0 gets created dynamically by ppp.

I have a freshly installed Windows Server 2008 R2 SP1 virtual machine which is completely unable to reach any IPv6 web pages, despite apparently having proper IPv6 connectivity. In addition, other Linux VMs cannot reach IPv6 web sites either.

This setup has previously worked, with full IPv6 connectivity in the virtual machines, and has stopped working without obvious reason.

All of my VMs are bridged to the physical Ethernet, and receive announcements from radvd on the host machine. IPv6 works correctly on the host machine, which is also the IPv6 router. Wireshark shows that the host machine is sending back an ICMPv6 Destination Unreachable (Administratively prohibited) after receiving the HTTP SYN packet.

Internet Explorer reports that it cannot display the web page, and Google Chrome only says Oops! Chrome could not connect to the web page, without an error number.

I am even able to ping the local gateway and Google's IPv6 addresses and do IPv6 DNS lookups.

PS C:\Users\Administrator> ping -6 fe80::6e62:6dff:fed1:dfad

Pinging fe80::6e62:6dff:fed1:dfad with 32 bytes of data:
Reply from fe80::6e62:6dff:fed1:dfad: time<1ms
Reply from fe80::6e62:6dff:fed1:dfad: time<1ms
Reply from fe80::6e62:6dff:fed1:dfad: time<1ms
Reply from fe80::6e62:6dff:fed1:dfad: time<1ms

Ping statistics for fe80::6e62:6dff:fed1:dfad:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 0ms, Maximum = 0ms, Average = 0ms

PS C:\Users\Administrator> ping -6 www.google.com

Pinging www.l.google.com [2001:4860:800a::67] with 32 bytes of data:
Reply from 2001:4860:800a::67: time=43ms
Reply from 2001:4860:800a::67: time=42ms
Reply from 2001:4860:800a::67: time=46ms
Reply from 2001:4860:800a::67: time=42ms

Ping statistics for 2001:4860:800a::67:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 42ms, Maximum = 46ms, Average = 43ms

My virtual machine's configuration looks like this:

PS C:\Users\Administrator> ipconfig /all

Windows IP Configuration

   Host Name . . . . . . . . . . . . : WIN-CRLO5NIQB72
   Primary Dns Suffix  . . . . . . . :
   Node Type . . . . . . . . . . . . : Hybrid
   IP Routing Enabled. . . . . . . . : No
   WINS Proxy Enabled. . . . . . . . : No
   DNS Suffix Search List. . . . . . : local

Ethernet adapter Local Area Connection 2:

   Connection-specific DNS Suffix  . : local
   Description . . . . . . . . . . . : Red Hat VirtIO Ethernet Adapter
   Physical Address. . . . . . . . . : 52-54-00-DD-DF-3E
   DHCP Enabled. . . . . . . . . . . : Yes
   Autoconfiguration Enabled . . . . : Yes
   IPv6 Address. . . . . . . . . . . : 2001:db8:1600:80bf:5054:ff:fedd:df3e(Preferred)
   Link-local IPv6 Address . . . . . : fe80::5054:ff:fedd:df3e%13(Preferred)
   IPv4 Address. . . . . . . . . . . : 192.168.12.146(Preferred)
   Subnet Mask . . . . . . . . . . . : 255.255.255.0
   Lease Obtained. . . . . . . . . . : Monday, July 09, 2012 1:59:42 PM
   Lease Expires . . . . . . . . . . : Tuesday, July 10, 2012 1:59:42 PM
   Default Gateway . . . . . . . . . : fe80::6e62:6dff:fed1:dfad%13
                                       192.168.12.1
   DHCP Server . . . . . . . . . . . : 192.168.12.1
   DNS Servers . . . . . . . . . . . : 2001:4860:4860::8888
                                       2001:4860:4860::8844
                                       192.168.12.1
   NetBIOS over Tcpip. . . . . . . . : Enabled

Tunnel adapter isatap.local:

   Media State . . . . . . . . . . . : Media disconnected
   Connection-specific DNS Suffix  . : local
   Description . . . . . . . . . . . : Microsoft ISATAP Adapter
   Physical Address. . . . . . . . . : 00-00-00-00-00-00-00-E0
   DHCP Enabled. . . . . . . . . . . : No
   Autoconfiguration Enabled . . . . : Yes

Tunnel adapter Teredo Tunneling Pseudo-Interface:

   Connection-specific DNS Suffix  . :
   Description . . . . . . . . . . . : Microsoft Teredo Tunneling Adapter
   Physical Address. . . . . . . . . : 00-00-00-00-00-00-00-E0
   DHCP Enabled. . . . . . . . . . . : No
   Autoconfiguration Enabled . . . . : Yes
   IPv6 Address. . . . . . . . . . . : 2001:0:4137:9e76:10d1:317d:3f57:f36d(Preferred)
   Link-local IPv6 Address . . . . . : fe80::10d1:317d:3f57:f36d%12(Preferred)
   Default Gateway . . . . . . . . . :
   NetBIOS over Tcpip. . . . . . . . : Disabled

PS C:\Users\Administrator> netsh interface ipv6 show route

Publish  Type      Met  Prefix                    Idx  Gateway/Interface Name
-------  --------  ---  ------------------------  ---  ------------------------
No       Manual    256  ::/0                       13  fe80::6e62:6dff:fed1:dfad
No       Manual    256  ::1/128                     1  Loopback Pseudo-Interface 1
No       Manual    8    2001::/32                  12  Teredo Tunneling Pseudo-Interface
No       Manual    256  2001:0:4137:9e76:10d1:317d:3f57:f36d/128   12  Teredo Tunneling Pseudo-Interface
No       Manual    8    2001:db8:1600:80bf::/64   13  Local Area Connection 2
No       Manual    256  2001:db8:1600:80bf:5054:ff:fedd:df3e/128   13  Local Area Connection 2
No       Manual    256  fe80::/64                  13  Local Area Connection 2
No       Manual    256  fe80::/64                  12  Teredo Tunneling Pseudo-Interface
No       Manual    256  fe80::5efe:192.168.12.146/128   11  isatap.local
No       Manual    256  fe80::10d1:317d:3f57:f36d/128   12  Teredo Tunneling Pseudo-Interface
No       Manual    256  fe80::5054:ff:fedd:df3e/128   13  Local Area Connection 2
No       Manual    256  ff00::/8                    1  Loopback Pseudo-Interface 1
No       Manual    256  ff00::/8                   13  Local Area Connection 2
No       Manual    256  ff00::/8                   12  Teredo Tunneling Pseudo-Interface

PS C:\Users\Administrator> netsh interface ipv6 show prefixpolicies
Querying active state...

Precedence  Label  Prefix
----------  -----  --------------------------------
        50      0  ::1/128
        40      1  ::/0
        30      2  2002::/16
        20      3  ::/96
        10      4  ::ffff:0:0/96
         5      5  2001::/32

So far in the VM I have tried:

netsh interface ipv6 set global randomizeidentifiers=disabled

No change.

Disabling the Teredo adapter: No change. And it somehow got re-enabled.

Using the Microsoft Fix-It to prefer IPv6 over IPv4: No change.

So far on the host I have tried:

Checked for IPv6 forwarding sysctl:

net.ipv6.conf.all.forwarding = 1
net.ipv6.conf.br0.forwarding = 1
net.ipv6.conf.default.forwarding = 1
net.ipv6.conf.em1.forwarding = 1
net.ipv6.conf.lo.forwarding = 1
net.ipv6.conf.sit0.forwarding = 1
net.ipv6.conf.sixxs.forwarding = 1
net.ipv6.conf.virbr0.forwarding = 1
net.ipv6.conf.virbr0-nic.forwarding = 1
net.ipv6.conf.vnet0.forwarding = 1
net.ipv6.conf.vnet1.forwarding = 1
net.ipv6.conf.vnet2.forwarding = 1

Restarted radvd: No change.

I'm setting up some servers that will run virtual machines using IPv6. The network topology is something like this, where the br0 interfaces are virtual:

             2001:db8:fedc:aaaa::/64
   ---+----------------------------------+---
      |                                  | 
      | eth0: 2001:db8:fedc:aaaa::1      | eth0: 2001:db8:fedc:aaaa::2
 +----+----+                        +----+----+
 | server1 |                        | server2 | 
 +----+----+                        +----+----+
      | br0: 2001:db8:fedc:abcd::1       | br0: 2001:db8:fedc:cdef::1
      |                                  | 
      | VM network 1:                    | VM network 2:
      | 2001:db8:fedc:abcd::/64          | 2001:db8:fedc:cdef::/64
 +----+----+                        +----+----+
 |    |    |                        |    |    | 
vm1  vm2  vm3                      vm4  vm5  vm6

I'm trying to avoid using the network's default gateway to route the packets from a VM in a server to one in another server, because with many hosts it would become a bottleneck.

The idea is then to have each server advertise a route to its own VM network to the other servers. So in the example above, server1 would advertise a route to 2001:db8:fedc:abcd::/64 and server2 to 2001:db8:fedc:cdef::/64.

I have this in server1's radvd.conf:

interface eth0 {
  AdvSendAdvert on;
  prefix 2001:db8:fedc:abcd::/64 { };
  route  2001:db8:fedc:abcd::/64 { };
};

And this is server2's:

interface eth0 {
  AdvSendAdvert on;
  prefix 2001:db8:fedc:cdef::/64 { };
  route  2001:db8:fedc:cdef::/64 { };
};

I can see this information being received from one of the servers if I run "rdisc6 eth0" on the other one, but for some reason the route isn't being added to its routing table.

What am I missing here?

My ISP has just started IPv4/IPv6 dual stack service. In order to connect IPv4/IPv6 internet, almost all users usually use a CPE leased by the ISP. But I want to use my Linux router instead of such CPE, because my router has so many roles(a lot of iptables rules and 3 Ethernet interfaces, http reverse proxy, http cache, memcached, smtp/pop server and DNS proxy). I want to let it to be an IPv4/IPv6 dual stack router.

My router

• CentOS6.0 i686
• eth0 for WAN
• eth1,2 for LAN
• ISC dhcp(version 4.1.1) installed by "# yum install dhcp" from CentOS-updates repo
• radvd(version 1.6) installed by "# yum install radvd" from CentOS-base repo
• [EDIT] The reason why I use ISC dhclient is that my router has to send a DHCPv6 option 16 (Vendor class) in its solicit and request messages, in order to get vendor-specific-information option which includes some ISP-specific informations(SIP phone number, firmware server address). I know how to set option 16 in ISC dhclient, but don't know it in WIDE-dhcpv6. And I can not find such settings as "id-assoc" in ISC dhclient.

My goal

• /48 prefix is delegated to DHCPv6-PD(DHCPv6 Prefix Delegation) client(dhclient) on eth0.
• An IPv6 default route is set toward to ISP. I have to consider ISP's DHCPv6 server's link-local address as a default route.
• One /64 subnet and one global address (which are in the delegated prefix) are assigned to each LAN I/F(eth1/eth2).
• radvd on eth1 and eth2 announces RA based on the assigned /64 subnet.
• dhcpd6 on eth1 and eth2 announces additional network information(name-servers, domain-searchlist and sip-servers-addresses) received from ISP's DHCPv6 server.

My current configuration

/etc/sysconfig/network

NETWORKING=yes
HOSTNAME=xxx.yyy.zzz

# Enable IPv6 routing and stop accept_ra/autoconf.
IPV6FORWARDING=yes

/etc/sysconfig/network-scripts/ifcfg-eth0

DEVICE="eth0"
ONBOOT=yes
TYPE=Ethernet
BOOTPROTO=none

# Enable DHCPv6-PD client.
DHCPV6C=yes
DHCPV6C_OPTIONS="-P -cf /etc/dhcp/dhclient6.conf"

# Disable make_resolv_conf function in /sbin/dhclient-script.
PEERDNS=no

/etc/dhcp/dhclient6.conf

script "/sbin/dhclient-script";
interface "eth0" {
        send dhcp6.reconf-accept;
        also request dhcp6.vendor-opts, dhcp6.sip-servers-names, dhcp6.sip-servers-addresses, dhcp6.sntp-servers;
}

Result

After "# service network restart", DHCPv6-PD client seems to finish successfully.

lease6 {
  interface "eth0";
  ia-pd xx:xx:xx:xx {
    starts 1312464004;
    renew 7200;
    rebind 10800;
    iaprefix 24xx:xxxx:xxxx::/48 {
      starts 1312464004;
      preferred-life 12600;
      max-life 14400;
    }
  }
  option dhcp6.client-id xx:xx:xx:xx:xx:xx:xx:xx:xx:xx;
  option dhcp6.server-id xx:xx:xx:xx:xx:xx:xx:xx:xx:xx;
  option dhcp6.reconf-accept ;
  option dhcp6.sip-servers-addresses 24xx:xxxx:xxxx:xxxx::1;
  option dhcp6.name-servers 24xx:xxxx:xxxx:xxxx::3,24xx:xxxx:xxxx:b::3;
  option dhcp6.domain-search "xxx.jp.", "yyy.jp.";
  option dhcp6.sntp-servers 24xx:xxxx:xxxx:xxxx::a,24xx:xxxx:xxxx:b::b;
}

Question

DHCPv6-PD client is delegated /48 prefix successfully. But after that,

• How to set an IPv6 default route in the router's routing table ?
• How to assign a global IPv6 address and /64 subnet to each LAN I/F(eth1 and eth2) based on the delegated prefix ?
• How to trigger radvd to announce RA on each LAN I/F(eth1 and eth2) ?
• How to trigger DHCPv6 server(dhcpd6) on each LAN I/F to announce additional network information(name-servers, domain-searchlist and sip-servers-addresses) received from ISP's DHCPv6 server ?

My guess

According to "$ man dhclient" and "$ man dhclient-script", dhclient-script seems to be used to configure such settings, and it offers some hooks. But now, it does nothing for DHCPv6-PD client.

Yes, I may have to write some scripts suitable for my DHCPv6-PD environment, though I am not familiar with IPv6 network configurations. Anyone let me know the above settings in dhclient-script ? [EDIT] I want to know the practical dhclient-script.

And I worry, some special consideration should be taken. For example, because each IPv6 address has preferred/valid lifetime, when I assign a global IPv6 address to a LAN interface based on the delegated prefix, I have to assign not only an IPv6 address but also its lifetimes which must be derived from the delegated prefix's preferred/valid lifetimes ?

Also, radvd and DHCPv6 server on LAN interfaces may have to be re-configured when DHCPv6 re-configuration happens ? Which hook should be used ?

[EDIT] When DHCPv6 re-configuration happens, how to announce both new and obsolete information ? radvd automatically announce the previous prefix is obsolete? DHCPv6 server automatically announce the previous name-servers are obsolete ?

IPv6 and RFC 2462 provide a means for hosts to configure their own IP addresses via stateless address autoconfiguration. To me this seems like the bee's knees, and it leaves me wondering why someone would want to go through the trouble of configuring a DHCPv6 server instead. I don't manage networks as a profession, so I'm guessing that there are some obvious simple reasons why one would want to support DHCPv6 that don't occur to me. Could someone please elaborate on what these reasons might be?