RES_BLAST

[Mark Andrews]

> Hello danny and mark,
> Thanks for all your answers.
> 
> As of this moment, I cant really confirm whether RES_BLAST is really
> implemented in the bind9 library. My limited skills, as far as I can
> muster,  produced a "grep" result that RES_BLAST is only referenced to, when
> the bind library makes the decision to execute the RES_ROTATE action.
> 
> It seems that, RES_BLAST is not configurable also from "resolv.conf" file's
> "options" settings. Being a newbie and all, I don't know if implementing a
> simulaneous DNS lookup/query to multiple nameservers would prove to be
> disastrous to a network, that is why bind9 seems to have no interface to set
> the RES_BLAST option. (Again I may be wrong here.)
> 
> I agree with you totally on the behaviour of some unix resolvers when the
> first name-server listed becomes unreachable. Aside from this, I intend to
> take a look at how RES_BLAST will handle situations where nameserver 1
> returns an NXDOMAIN and nameserver 2 responds with an authoritative result.

	All servers listes in resolv.conf are expected to return
	consistant results.  RES_BLAST returns the *first* answer
	whether it be NXDOMAIN, NODATA or records.

	B.T.W. NXDOMAIN *is* an authoritative result

> As for creating my own name server, is there a way for me to mimic the
> effect of RES_BLAST using an IP forwarder. What I mean is, I setup an ip
> forwarder, delcare two root zones (if this is even possible), within my
> named.conf file so that the daemon is forced to lookup the forward list of
> both root zones? Is it even possible to have two root zones? As for lwresd,
> I am not yet familiar with it and will look into it also.
> 
> I hope I make sense.
> 
> Thank you all so much for the assistance,
> GovZ
> 
> On 10/24/07, Danny Thomas <d.thomas at its.uq.edu.au> wrote:
> >
> > >       RES_BLAST is available in lib/bind
> >
> > It would be great if you could make RES_BLAST the default in resolv.conf,
> > as it would overcome the bad behaviour of the (unix) resolver when the
> > first name-server listed becomes unreachable. Mind you, I don't know
> > whether resolv.conf directives would come from bind, or whether it
> > is code managed by the OS people importing bind code in the library.
> >
> > I know there are other approaches which do keep track of how quickly
> > name-servers respond, and which soon learn when a name-server is unreach-
> > able, but these involve daemons like a caching name-server or lwresd.
> > So resolving is affected if the daemon dies.
> >
> > Danny
> >
> > --
> >   d.thomas at its.uq.edu.au    Danny Thomas,
> >          +61-7-3365-8221    Software Infrastructure,
> > http://www.its.uq.edu.au    ITS, The University of Queensland
> >
> >
> >
> 
> 
> 
-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: Mark_Andrews at isc.org