Comment 82 for bug 959037

On 15/06/12 08:04, Thomas Hood wrote:
> Alkis: This relies on the assumption that NM's configuration text can be
> dropped in alongside whatever other configuration text is present and
> that dnsmasq will still work properly. This assumption is, er,
> questionable.

There was an attempt, some time ago, to provide a way to allow something
like libvirt to add its DHCP configuration to a system dnsmasq
configuration without interfering with the existing config. It's
basically a way to specify an interface and subnet for DHCP in a config
line which overrides other access control, so for instance if the
system dnsmasq config limits it to certain interfaces, then the
interface specified by libvirt would be added to that set.

To my knowledge this facility has never actually been used.

>
> And this is also one answer to my question in #72. The "dnsmasq
> cascade" may waste resources but it has maintenance advantages. One
> dnsmasq process is under the control of NM. The other is under the
> control of the admin. They communicate with each other via a well
> defined protocol, RFC 1035.

This is a good argument, I think.

Simon.