[Dnsmasq-discuss] dnsmasq and RA

Gene Czarcinski gene at czarc.net
Thu Nov 1 13:48:48 GMT 2012


On 11/01/2012 08:59 AM, Gene Czarcinski wrote:
> On 10/31/2012 04:41 PM, Gene Czarcinski wrote:
>> Instead of running both radvd and dnsmasq, I would like to run only 
>> dnsmasq and have it do what is now being done by radvd.
>>
>> This is for libvirt so there are some specifics which could help.
>>
>> 1. There will be one and only one interface for an instance of dnsmasq.
>>
>> 2. Currently (at least with some patches I have created) run one ipv4 
>> dhcp subnetwork or one ipv6 dhcp supnetwork or both one ipv4 network 
>> and one ipv6 network.
>>
>> 3.  If a dhcp-range is specified for the ipv6 subnetwork then I need 
>> the AdvManagedFlag on which (I believe) is done with the additional 
>> dsnmasq parameter "enable-ra".
>>
>> 4.  But, I may have an IPv6 subnetwork but dhcp-range has not been 
>> specified.  In this case I want the AdvManagedFlag off. This one is 
>> the situation that I am not clear on.
>>
>> 5.  None of the other dnsmasq capabilities with respect to RA will be 
>> used.
>>
>> Comments/suggestions?
>>
>> If this requires a little enhancement patch, I am willing.
>>
>>
> OK, I believe I have answered my own question but I would still like 
> confirmation.
>
> I ran a test with no radvd running ... only dnsmasq.  With enable-ra 
> and dhcp-range specified for ipv6 addresses, the client side gets a 
> dhcp6 address and RA default route.  With enable-ra and NO dhcp-range 
> specified, the client gets a SLAAC address and RA default route.
>
> With enable-ra and dhcp-range,I do notice a lot of RTR-ADVERT messages 
> ... like one every couple of seconds.  Is there a way to suppress this 
> message and/or increase the time between the events or at least the 
> messages [that is without disabling dhcp logging] ?
>
> Maybe it would be useful is this message was only sent to the syslog 
> once every ten or twenty events.
>
One additional question  With enable-ra and dhcp-range, then dnsmasq 
supplies the RDNS info in the RA packet but it does not if dchpv6 is not 
specified.  It seems like this should be reversed.

Gene



More information about the Dnsmasq-discuss mailing list