[Dnsmasq-discuss] Using IPv6 DHCP

Simon Kelley simon at thekelleys.org.uk
Tue Sep 25 12:45:58 BST 2012


  Doh!

I didn't read your config carefully enough. You're declaring the 
dhcp-range as "ra-stateless", which tells dnsmasq to advertise that 
subnet via router-advertisements, and set the bits which tell the 
clients to use DHCPv6 for configuration, but not address allocation.

Clients are expected to send DHCP information-requests for DNS servers, 
NTP servers, etc, etc, but won't get IP addresses allocated via DHCPv6. 
Hence dnsmasq is refusing to allocate and address.

Remove ra-stateless from your config and it will all work.


(On the v6-only question, if you don't supply a DHCPv4 address range, 
then dnsmasq won't reply to DHCPv4 requests.

Cheers,

Simon.

On 25/09/12 10:41, Sean Boran wrote:
> Hi,
>
> On 25 September 2012 10:27, Simon Kelley <simon at thekelleys.org.uk> wrote:
>> On 25/09/12 09:10, Sean Boran wrote:
>>>
>>> Hi,
>>>
>>> I'm doing sdome DHCPv6 tests and have a few questions :-)
>>>
>>> First is it is possible to disable v4 respond, and only respond to v6
>>> requests? When testing I dont want to disturb a v4 dhcp already
>>> running.  (e.g. should I set listen-address to my IPv6 address?)
>>
>>
>> If you're happy to use the very latest code, then grab 2.64test5 and add
>> dhcp-ignore=!DHCPv6
>>
>> to your configuration.
>
> I also disable "enable-ra", since ra should not be needed, when using dhcp?
>
>
> Sep 25 11:33:55 sisdns6 dnsmasq[9577]: started, version 2.64test5 cachesize 150
> Sep 25 11:33:55 sisdns6 dnsmasq[9577]: compile time options: IPv6
> GNU-getopt no-DBus no-i18n no-IDN DHCP DHCPv6 no-Lua TFTP no-conntrack
> Sep 25 11:33:55 sisdns6 dnsmasq-dhcp[9577]: DHCPv6 stateless on
> 2001:918:ffcb:199::10
> Sep 25 11:33:55 sisdns6 dnsmasq-dhcp[9577]: SLAAC on
> 2001:918:ffcb:199:: prefix valid 1d
>
>
>>>
>>> Secondly, I'm having issues assigning v6 addresses.
>>>
>>> compiled dnsmasq-2.63, config:
>>> --
>>> domain=mydomain.net
>>> dhcp-range=2001:918:ffcb:199::10, 2001:918:ffcb:199::20, ra-stateless, 24h
                                                              ^^^^^^^^^^^
>>> dhcp-sequential-ip
>>> enable-ra
>>> dhcp-option=option6:dns-server,[::]
>>> log-dhcp
>>> --
>>>
>>
>> Does the relevant interface have an address on the 2001:918:ffcb:199::
>> subnet, and prefix length 64?
>
> Yes.
>
>   ifconfig
> eth0      Link encap:Ethernet  HWaddr 00:50:56:83:3a:d4
>            inet addr:10.90.199.11  Bcast:10.90.199.255  Mask:255.255.255.0
>            inet6 addr: 2001:918:ffcb:199:fc54:9480:e886:e639/64 Scope:Global
>            inet6 addr: 2001:918:ffcb:199:250:56ff:fe83:3ad4/64 Scope:Global
>            inet6 addr: fe80::250:56ff:fe83:3ad4/64 Scope:Link
>
>
>>
>> Please could you try without the dhcp-sequential-ip option. That's unlikely
>> to have been well tested for DHCPv6, and smells suspicious to me.
>>
>> Cheers,
>>
>> Simon.
>
> Thanks done that, but no change really.
>
> Sep 25 11:38:46 sisdns6 dnsmasq-dhcp[937]: 7022133 available DHCP
> range: 2001:918:ffcb:199::10 -- 2001:918:ffcb:199::20
> Sep 25 11:38:46 sisdns6 dnsmasq-dhcp[937]: 7022133 DHCPSOLICIT(eth0)
> 00:01:00:01:00:50:10:5a:de:08
> Sep 25 11:38:46 sisdns6 dnsmasq-dhcp[937]: 7022133 requested options:
> 24:domain-search, 39:FQDN, 23:dns-server,
> Sep 25 11:38:46 sisdns6 dnsmasq-dhcp[937]: 7022133 requested options:
> 31:sntp-server
> Sep 25 11:38:46 sisdns6 dnsmasq-dhcp[937]: 7022133 tags: DHCPv6, eth0
> Sep 25 11:38:46 sisdns6 dnsmasq-dhcp[937]: 7022133 sent size: 10
> option:  1 client-id  00:01:00:01:00:50:10:5a:de:08
> Sep 25 11:38:46 sisdns6 dnsmasq-dhcp[937]: 7022133 sent size: 14
> option:  2 server-id  00:01:00:01:17:f2:f6:90:00:50:56:83:3a:d4
> Sep 25 11:38:46 sisdns6 dnsmasq-dhcp[937]: 7022133 sent size: 40
> option:  3 ia-na  IAID=1451433281 T1=0 T2=0
> Sep 25 11:38:46 sisdns6 dnsmasq-dhcp[937]: 7022133 nest size: 24
> option: 13 status  2 No addresses available
> Sep 25 11:38:46 sisdns6 dnsmasq-dhcp[937]: 7022133 sent size: 16
> option: 23 dns-server  2001:918:ffcb:199:fc54:9480:e886:e639
>
> Regards,
>
> Sean
>




More information about the Dnsmasq-discuss mailing list