Re: [v6ops] SLAAC renum: Problem Statement & Operational workarounds

Philip Homburg <> Wed, 30 October 2019 10:38 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id 8384E12010F for <>; Wed, 30 Oct 2019 03:38:19 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -1.499
X-Spam-Status: No, score=-1.499 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, KHOP_HELO_FCRDNS=0.399, SPF_HELO_NONE=0.001, SPF_NONE=0.001] autolearn=no autolearn_force=no
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id n_MNxbKXWwDc for <>; Wed, 30 Oct 2019 03:38:18 -0700 (PDT)
Received: from ( [IPv6:2001:888:1044:10:2a0:c9ff:fe9f:17a9]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by (Postfix) with ESMTPS id D876D120013 for <>; Wed, 30 Oct 2019 03:38:17 -0700 (PDT)
Received: from (localhost [::ffff:]) by with esmtp (TLS version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384) (Smail #157) id m1iPlMZ-0000J5C; Wed, 30 Oct 2019 11:38:03 +0100
Message-Id: <>
From: Philip Homburg <>
References: <> <> <> <> <> <> <> <>
In-reply-to: Your message of "Tue, 29 Oct 2019 10:20:26 -0700 ." <>
Date: Wed, 30 Oct 2019 11:37:57 +0100
Archived-At: <>
Subject: Re: [v6ops] SLAAC renum: Problem Statement & Operational workarounds
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: v6ops discussion list <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Wed, 30 Oct 2019 10:38:19 -0000

> > For IPv4 + NAT, if you flash renumber the upstream address then existing
> > connections will be stuck.
> Not exactly They do fairly quickly get sent TCP RSTs in most cases.

That's not my experience. I hardly ever see a CPE generate a RST when a flow
doesn't exist. 

> OTOH, leaving the connection hung until it times out at the
> application or host TCP stack level creates huge delays in service
> restoration.

This is exactly what happens today in many cases. Of course, where developers
get annoyed by this behaviour, they implement shorter timeout at the 
application level. I.e., a typical webbrowser doesn't wait for the host to
report an error on a TCP connection.

> Because the IPv4 host fairly quickly receives an RST packet when
> the outbound packet hits a non-existent state table entry or a
> state table entry marked as dead.

My experience is different. In any case, if we want the CPE to send a RST
when it upstream prefix changes, then of course that can be coded as well.
I very much doubt that we really want to go there. Changing the host
would be better.

> Unless the CPE knows to send out an RA with a preferred lifetime
> of 0 for the now deprecated prefix, then you are mistaken in the
> actual systemic behavior.

The whole point of this discussion is to write a draft that requires the CPE
to do so. Probably with a companion of recommended host changes.