[ previous ] [ next ] [ threads ]
 
 From:  "Michael Joy" <mdjoy at phy dot olemiss dot edu>
 To:  "'Manuel Kasper'" <mk at neon1 dot net>
 Cc:  <m0n0wall at lists dot m0n0 dot ch>
 Subject:  RE: [m0n0wall] Lockups since upgrade to 1.2b3
 Date:  Sat, 18 Dec 2004 04:37:08 -0600
Well, I know for certain that I'm running the fixed wrap board. I inquired
with pcengines to make sure I received a fixed board as the first one I
received was DOA (smoked on a bench test). The bios on this board is less
than 2 months old, and it's only one revision old. It's not the newer 1D-3
that just came out. The board is only 3 months old, and came straight from
Switzerland. And if it was some instability inherent to the board, I would
have had problems with other builds most wouldn't I?

I'll try changing the udp ack timeout and see what happens. Perhaps the
buffer is filling up with udp ack's that are failing to timeout in a timely
manner, causing something to lock hard?

I know as far as the sip stuff is concerned, I run a packet8 DTA for my
telephone line, and I didn't have problems with it before with the udp
timeout at 24.

-----Original Message-----
From: Manuel Kasper [mailto:mk at neon1 dot net] 
Sent: Saturday, December 18, 2004 4:30 AM
To: Michael D. Joy
Cc: m0n0wall at lists dot m0n0 dot ch
Subject: Re: [m0n0wall] Lockups since upgrade to 1.2b3

On 17.12.2004 22:18 -0600, Michael D. Joy wrote:

> I'm running a Wrap 1C-3 and am also experiencing hard locks after
> 1.2b3 upgrade. I have since reverted to 1.2b2. Oddly enough, the

Hmm, the only change from 1.2b2 to 1.2b3 that could possibly cause
this (even though that would imply big problems with ipfilter), as
far as I can imagine, is the UDP ack timeout (24 -> 240 seconds). You
can try the following command on exec.php to set it back to 24 (until
the next reboot):

/sbin/sysctl net.inet.ipf.fr_udpacktimeout=24

I can't see why 1.2b2 should work when 1.2b3 doesn't - both versions
even use the exact same kernel image. As long as I can't reproduce it
here (been running it on a net4501 ever since it was released) and
there are no other indicators as to what's going wrong (console
messages before it dies etc.), it'll be pretty hard to fix. However,
two out of three people who reported this bug are running WRAPs - see
<http://m0n0.ch/wall/list/?action=show_msg&actionargs[]=109&actionargs[]=84>
.

- Manuel