BIND 8.3.0 RC2 is now available for public testing
Nate Campi
nate at wired.com
Fri Jan 4 18:47:14 UTC 2002
On Fri, Jan 04, 2002 at 07:52:14AM -0800, Bill Manning wrote:
>
> On Thu, Jan 03, 2002 at 11:05:58PM -0500, William Stacey wrote:
> > I know this must sound dumb, but with 9.2.0 out, why do the still build new
> > 8.x.x versions?
> >
> > --
> > William Stacey
>
> ) its faster
> ) its smaller
For resolver performance BIND 8 is light years ahead of BIND 9. I've
posted to this list before about how the EDNS0 checks BIND 9 does
against remote servers brings BIND 9 to a crawl. Use queryperf with
debugging off to see just how slow it is, then turn on debugging to
see what's making it so slow - you'll see what I mean. I've found BIND 9
to be totally useless as a resolver, using real life tests (queryperf
and internal web log reporting apps that can do as many queries/sec as
the resolver can handle).
For authoritative data BIND 9 seems faster than any other DNS server
I've found, but most of us have many more resolvers than "content" DNS
servers (I have 13 "content" DNS servers, actually, though I'm moving to
just five).
--
Nate Campi | Terra Lycos DNS | WiReD UNIX Operations
Dump everything on Osama.
# dump 0f osama /
More information about the bind-users
mailing list