From owner-IPv6-jp@jp.freebsd.org  Thu Jun 29 18:39:51 2000
Received: (from daemon@localhost)
	by castle.jp.freebsd.org (8.9.3+3.2W/8.7.3) id SAA61573;
	Thu, 29 Jun 2000 18:39:51 +0900 (JST)
	(envelope-from owner-IPv6-jp@jp.FreeBSD.org)
Received: from ns5.sony.co.jp (ns5.Sony.CO.JP [202.238.80.5])
	by castle.jp.freebsd.org (8.9.3+3.2W/8.7.3) with ESMTP id SAA61568
	for <IPv6-jp@jp.freebsd.org>; Thu, 29 Jun 2000 18:39:50 +0900 (JST)
	(envelope-from mistral@imasy.or.jp)
Received: from mail3.sony.co.jp (gatekeeper15.Sony.CO.JP [202.238.80.28])
	by ns5.sony.co.jp (R8) with ESMTP id SAA05863
	for <IPv6-jp@jp.freebsd.org>; Thu, 29 Jun 2000 18:39:50 +0900 (JST)
Received: from mistral.wind.prv ([43.1.172.41])
	by mail3.sony.co.jp (R8) with ESMTP id SAA27696
	for <IPv6-jp@jp.freebsd.org>; Thu, 29 Jun 2000 18:39:48 +0900 (JST)
Message-Id: <200006290939.SAA27696@mail3.sony.co.jp>
Received: (from yohta@localhost)
	by mistral.wind.prv (8.9.3/3.7Wpl2-990626) id SAA05240;
	Thu, 29 Jun 2000 18:39:49 +0900 (JST)
Date: Thu, 29 Jun 2000 18:39:49 +0900 (JST)
From: mistral@imasy.or.jp (Yoshihiko SARUMARU)
To: IPv6-jp@jp.freebsd.org
In-Reply-To: Your message of "Thu, 29 Jun 2000 17:22:47 +0900".
	<200006290822.e5T8Mmx00864@plum.ssr.bisd.hitachi.co.jp>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
X-Mailer: mnews [version 1.22PL4] 2000-05/28(Sun)
Reply-To: IPv6-jp@jp.freebsd.org
Precedence: list
X-Distribute: distribute version 2.1 (Alpha) patchlevel 24e+000315
X-Sequence: IPv6-jp 809
Subject: [IPv6-jp 809] Re: =?ISO-2022-JP?B?GyRCJTUhPCVQJCwbKEI=?= IPv6
	=?ISO-2022-JP?B?GyRCJEAkSENZJCQhKRsoQg==?= 
Errors-To: owner-IPv6-jp@jp.freebsd.org
Sender: owner-IPv6-jp@jp.freebsd.org
X-Originator: mistral@imasy.or.jp

ume@bisd.hitachi.co.jp wrote:

> >>>>> On Thu, 29 Jun 2000 17:01:50 +0900 (JST)
> >>>>> mistral@imasy.or.jp (Yoshihiko SARUMARU) said:
> 
> mistral> It still becomes slow in about 50/50 probability. But I could
> mistral> know that if both server (leafnode+) and client (mnews) are IPv6 
> mistral> speakable, then the problem don't happen even if with -DINET6 kernel.
> 
> Does it take about 60 seconds?  I suspect timeout code in nntputil.c
> and connect() doesn't fail right now.
> Do you have IPv6 route to the host mnews is running?

No, it requires about 30 - 40 seconds (average), and in half of
cases it consumes only 1 second to list active.
In my environment, both server and client are running on the
same host, so it may use loopback interface and route to
loopback interace (::1) is available.

And I may forgot to post here, even if -DINET6 option is
not defined, slowness is happen on KAME (+ PAO) kernel.
# I have forget whether I tested on plain KAME kernel, but I may 
# test it and the result is slow.

--
Yoshihiko SARUMARU
mail: mistral@imasy.or.jp	web: http://www.imasy.or.jp/~mistral/
