[Dnsmasq-discuss] Sending RA without dhcp-range

Shixiong Shang sparkofwisdom.cloud at gmail.com
Fri Jan 17 14:21:25 GMT 2014


Hi, Simon and Albert:

Thank both of you for quick response!

According to the man page online, the mode like “ra-only”, and “ra-stateless” are only used as part of “dhcp-range” option. However, this will also turn dnsmasq to DHCPv6 server, which I try to avoid. Would you please help me clarify how I can send RA on specific interfaces with the right flags? The only thing I can think of is leveraging “—ra-param”, but it doesn’t have place to put in mode keywords…..

Thanks again!

Shixiong







On Jan 17, 2014, at 5:49 AM, Simon Kelley <simon at thekelleys.org.uk> wrote:

> On 17/01/14 06:10, Albert ARIBAUD wrote:
>> Le 17/01/2014 05:30, Shixiong Shang a écrit :
>>> Hi, experts:
>>> 
>>> Is it possible that dnsmasq can send out RA with either A=0, MO=1
>>> (i.e. for DHCPv6 Stateful), or AO=1, M=0 (i.e. for DHCPv6 Stateless)
>>> WITHOUT dhcp-range? What I am trying to achieve here is using dnsmasq
>>> to send out RA to trigger DHCPv6 Stateful or DHCPv6 Stateless mode
>>> without acting as DHCPv6 server. Based on my reading, I am under the
>>> impression that “—enable-ra” must be used with “—dhcp-range”.
>>> Otherwise, dnsmasq will not send RA…..
>> 
>> You are right about having to add enable-ra.
>> 
>> However there is a syntax of dhcp-range which will not cause DHCPv6.
>> 
>> As per 'man dnsmasq' (search for 'dhcp-range'), the possible modes for
>> IPv6 ranges include 'ra-only', which will previsely enable RAs without
>> enabling any DHCP.
>> 
> 
> Neither of you are completely right.
> 
> enable-ra causes RA with A=0 and M = 1  O = 1 on every interface that's doing DHCP, just to instruct the clients to do stateful DHCP. It's intended as a simple option for systems which don't use RA at all.
> 
> You don't need to set enable-ra at all to do RA on specific interfaces with other flag combinations, ie
> 
> ra-only A=1 M=0 O=0
> slaac   A=1 M=1 O=1
> ra-stateless A=1 M=0 O=1
> 
> Cheers,
> 
> Simon.
> 



More information about the Dnsmasq-discuss mailing list