Subject: Re: [resolv.conf] domain vs. search
To: VaX#n8 <vax@ccwf.cc.utexas.edu>
From: Alan Barrett <barrett@daisy.ee.und.ac.za>
List: current-users
Date: 12/20/1994 10:07:04
> In my case, I set my hostname to slip-xx-yy.ots.utexas.edu and
> the domain name in /etc/resolv.conf to cc.utexas.edu.... the
> idea being that I could "telnet ccwf", for example, and get
> "ccwf.cc.utexas.edu".  This worked fine, but "telnet dimebox.cs" did
> not get me "dimebox.cs.utexas.edu" - that is, it was not trying each
> successive parent domain, only the one listed.

See RFC 1537 for a discussion of some security problems with implicit
searches.  The current recommendation is that you should have to specify
the search list explicitly.

> One more question: the distinction between bind/resolver and other
> things is kind of blurry to me.

The name BIND is often used to refer to the package containing a
particular implementation of both a domain name server and a resolver
(or client). "named" is the server (configured via /etc/named.boot),
and the resolver (really a set of library functions) is the client
(configured via /etc/resolv.conf).  Programs call library functions
that implement the resolver; the resolver sends request packets to one
or more servers; the servers send response packets to the resolver; the
resolver returns the answers to the program.

> Where can I get a description of how names are converted into IP
> addresses in a monolithic or sequential document?

The BIND Operations Guide (BOG), included in the standard BIND
distribution, is a good start.  BIND-4.9.3 is currently in late
beta testing, and the announcement for BETA14 should be in
comp.protocols.tcp-ip.domains right now.

--apb (Alan Barrett)