:: Re: [DNG] how to clear DNS cache
Top Pagina
Delete this message
Reply to this message
Auteur: Rick Moen
Datum:  
Aan: dng
Onderwerp: Re: [DNG] how to clear DNS cache
Quoting Steve Litt (slitt@???):


[BIND9:]

> I didn't like it for that same reason, but always wondered if it were
> just a case of my drinking the djb kool aid.


Dan's many criticisms of BIND9 and of its BIND8 predecessor (a
from-scratch rewrite separating the two) were always well-founded in
fact even if so very slathered in assholish rhetoric, and surrounded by
dickish personal conduct, that people often resisted his truths just
because of his extremely offputting manner and conduct. (I never did,
FWIW.)

The very least of that assholish rhetoric is speaking as if BIND8 and
BIND9 were the very same codebase rather than a from-scratch rewrite
having occurred between them. It goes on from there, as you are
probably aware.


[PowerDNS;]

> Nix on that, for similar reasons as my keeping away from dbus-necessary
> software.


I of course concur. The pair of PowerDNS packages are aimed at huge
installations, primarily, e.g., my $FIRM where I did the migration to
PowerDNS from BIND9.


[Deadwood:]

> Sounds excellent. Maintained by one is less of a problem when there are
> so many alternatives.


Sam Trenholme freely admits sometimes when he took a while getting
around to fixing a problem in MaraDNS or Deadwood, telling people on his
mailing list that he's obliged to put work and other priorities first.
He says, quite reasonably, that if anyone wants faster response, he'll
gladly accept most forms of, y'know, money.

On the one hand, this means that maintenance is a bit thinly available.
On the other, it's somewhat likely true of many other open source
codebases including the competitors, and the only big difference is that
Sam is bluntly honest about it. (On the gripping hand, to use an old
Niven & Pournelle metaphor, I get the impression that Unbound is
-very- responsively maintained.)


[Unbound:]

> Sounds excellent.


IMO, as should be obvious, it is.


[Debian djbdns/dbndns:]

> As the author of runit, Gerrit Pape is my main man. If I continue using
> djbdns, I'd certainly get his version, *if* I could get a source version
> not tightly bound to Debian, and to dbus and systemd.


Project page: https://packages.debian.org/source/sid/djbdns
Please let me know how it goes.


[N-DJBDNS:]

> Ain't no redhat in this shop.


But you might look at the N-DJBDNS source tree if Gerrit Pape's Debian
djbdns/dbndns doesn't suit for some reason. Not every coder's work
sucks automatically just because you don't like his/her employer.


> If the others are small, this wouldn't be a priority with me.


All of dnscache, Unbound, and Deadwood qualify as small, fast, light.


> I don't find Jonathan de Boyne Pollard credible.


But that doesn't prevent verifiable factual assertions of his from being
correct, and am pretty certain that one was. I haven't verified it
myself, but it certainly _sounds_ like the sort of thing Dan would
ignore as 'Yes, my software doesn't resolve Akamai's incredibly
squirrely delegation, because I don't like the way Akamai does that and
consider it Bad and Wrong.' That would match his established pattern of
deciding to deliberately ignore RFCs he doesn't like that the
consensus-reality world regards as routine and saying 'I don't care.
They suck.'

I'm not arguing with Dan's judgement that certain RFCs suck, or even with
the particular ones he dislikes. I'm just saying that this pattern of
behaviour is part of what you get with djbware -- this being one of the
several points I made in my Web essay, and indeed one of the particular
points in it to which Dan specifically objected, when he made his
legal threat.



> I think we've all read those exchanges.


{shrug} I'm well-known within a tiny little circle of software people. ;->


> Geez, only local people have called me an idiot. I'd love to have
> Poettering call me an idiot.


I'll note in passing that Dan's assertion of my idiocy was based on
his deliberate _misrepresentation_ of my characterisation of his
then-licensing. I.e., his cartoon distortion of what I said might
indeed be seen as idiotic, but what I _said_ IMO wasn't.

I occasionally am obliged to clarify that, as here on spf-devel back in
2004: http://linuxmafia.com/~rick/faq/just-another-djb-groupie.html

-- 
Cheers,                 "The crows seemed to be calling his name, thought Caw."
Rick Moen                                     -- Deep Thoughts by Jack Handy 
rick@??? 
McQ! (4x80)