From owner-freebsd-bugs  Thu Jun  6 20:55:18 1996
Return-Path: owner-bugs
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id UAA26511
          for bugs-outgoing; Thu, 6 Jun 1996 20:55:18 -0700 (PDT)
Received: from enteract.com (root@enteract.com [206.54.252.1])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id UAA26473;
          Thu, 6 Jun 1996 20:55:09 -0700 (PDT)
Received: (from tqbf@localhost) by enteract.com (8.7.5/8.7.6) id WAA06848; Thu, 6 Jun 1996 22:55:07 -0500 (CDT)
From: Thomas Ptacek <tqbf@rdist.org>
Message-Id: <199606070355.WAA06848@enteract.com>
Subject: ifconfig alias issues
To: bugs@freebsd.org
Date: Thu, 6 Jun 1996 22:55:06 -0500 (CDT)
Cc: hackers@freebsd.org
Reply-To: tqbf@enteract.com
X-Mailer: ELM [version 2.4 PL24 ME8a]
Content-Type: text
Sender: owner-bugs@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

Ok. Same problem. We've got a machine with a couple hundred interface
aliases on a single interface. Periodically, one or two specific aliases
on it will break; the interface will no longer respond to packets 
addressed to that alias, and there will be extra host routes added towards
the interface.

Deleting the routes and the alias, then re-adding the alias, solves the 
problem. However, do this once or twice (and 2 or 3 always break at 
once), and all (every one of 'em) the aliases go away.

We add the aliases from a script. Re-running the script and re-adding the 
aliases has twice now fixed the problem for weeks at a time.

I've observed that this problem only occurs at or around boot-up time (I 
check immediately after reboots now). I've also noticed that this only 
seems to happen immediately after the route cache expire times get 
shorter. I'm wondering if route expirations could be causing the problem; 
that would explain (to me) why the system stabilizes completely after 
adding the aliases again (all the routes get re-added) and the problem 
goes away for a long time (because the route cache expire time only goes 
down once or twice, and always right after the system comes up).

Any ideas? This is happening on a Pentium 133 with an SMC ISA Ethernet 
NIC and 64M RAM. MAXUSERS is currently 64, NMBCLUSTERS is 2048 (we're not 
running out of mbufs or anything tho.) rtmaxcache is 400, up from 200, and
I'm trying it at 600, although this seems like a band-aid for the problem 
(if route expires really are the problem), since changes in the route 
cache expiration shouldn't (from what I can tell) kill all the interfaces.

Gratzi.

----------------
Thomas Ptacek at EnterAct, L.L.C., Chicago, IL [tqbf@enteract.com]
----------------
main(){while(1)fork();}