Re: [dnsext] Lame Server responses

[Tony Finch <[e-mail]>]

On Mon, 11 Oct 2010, Edward Lewis wrote:
> At 18:52 +0100 10/11/10, George Barwood wrote:
>
> > I agree with BIND, it seems to me that REFUSED is closest to then codes
> > defined by the standard.  I expect either SERVERFAIL or REFUSED will
> > work perfectly well.
>
> From reading the spec, neither really applies.  But I think they are the only
> two choices (from the existing pool). I.e., in REFUSED, the "eg" uses the word
> "wishes" which isn't the issue.

You could view it as not wishing to / refusing to return a referral.

> OTOH, SERVFAIL talks about name server error, which also isn't the
> issue.  Our bias was that it was a system error that caused any earnest
> query to go to a lame server, so we adopted SERVFAIL.

BIND's behaviour makes a useful distinction between lame (REFUSED) and
broken (SERVFAIL).

> > Can you think of any other situation that causes REFUSED to be returned (to
> > a normal query)?
>
> AXFR?  Depends on what's normal.  A server that will only answer queries with
> specific TSIG keys?

Or a recursive server that only provides service to specific networks.
Or a master server that does not permit dynamic updates.

Tony.
-- 
f.anthony.n.finch  <[e-mail]>  http://dotat.at/
HUMBER THAMES DOVER WIGHT PORTLAND: NORTH BACKING WEST OR NORTHWEST, 5 TO 7,
DECREASING 4 OR 5, OCCASIONALLY 6 LATER IN HUMBER AND THAMES. MODERATE OR
ROUGH. RAIN THEN FAIR. GOOD.