Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 19 Aug 2002 15:59:14 -0700
From:      Steve Francis <sfrancis@expertcity.com>
To:        Jesper Skriver <jesper@FreeBSD.org>
Cc:        Jeff Behl <jeff@expertcity.com>, net@freebsd.org
Subject:   Re: MTU not working?
Message-ID:  <3D617842.4020304@expertcity.com>
References:  <3D585310.8010709@expertcity.com> <20020819151623.GD48757@FreeBSD.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Allow me to respond for Jeff (we work at same place, and have both been
looking this issue)

ICMP's being  blocked are the most common explanation for this - but
this is not the case here.

tcpdump run on the server system shows the ICMP fragmention required -
DF bit set messages being received, and - the irrefutable proof that it
is not an ICMP filtering issue -the FreeBSD system DOES lower the MTU
for that host's cloned route  to the value specified in the ICMP (1420
in the snippet below).  New packets are segmented to sizes <= the new
MTU, but it continues resend the original packet over  and over, in the
original 1500 byte size.

So I still say FreeBSD has broken pMTU-D code.

This is reproducible at will, so we can collect whatever info anyone
wants.


The ICMP's

Jesper Skriver wrote:
 >
 > On Mon, Aug 12, 2002 at 05:30:08PM -0700, Jeff Behl wrote:
 > > We're seeing something odd that is seriously impacting our service for
 > > some clients.   Even though a the box seems to have the correct MTU in
 > > the routing table, it doesn't seem to follow it in all cases.
 > >
 > > dell350-12.snv#sysctl -a | grep mtu
 > > net.inet.tcp.path_mtu_discovery: 1
 > >
 > > dell350-12.snv#uname -a
 > > FreeBSD dell350-12.snv 4.4-RELEASE FreeBSD 4.4-RELEASE #0: Fri Aug  2
 > > 03:24:36 PDT 2002
 > > root@dell350-12.snv:/usr/src/sys/compile/COMMS44-2  i386
 > >
 > >
 > > dell350-12.snv#netstat -rnal | grep 10.4.1.134
 > > 10.4.1.134         63.xxx.224.129     UGHW        1     2268   1420 
   fxp0
 > >
 > >
 > > but tcpdump shows:
 > >
 > >
 > > 17:21:43.497275 63.xxx.224.154.80 > 10.4.1.134.2314: . 1:1461(1460) ack
 > > 248 win 17520 (DF)
 > > 17:21:51.497212 63.xxx.224.154.80 > 10.4.1.134.2314: . 1:1461(1460) ack
 > > 248 win 17520 (DF)
 > > 17:22:07.497065 63.xxx.224.154.80 > 10.4.1.134.2314: . 1:1461(1460) ack
 > > 248 win 17520 (DF)
 > >
 > >
 > > obviosly the packet isn't making it back to the client as it's too big.
 > > any ideas?  I don't see any bug reports on it.  We believe the same
 > > situation exsits with a 4.6 server but I haven't checked that thorougly
 > > enough yet.
 >
 > It's most likely a firewall or router filter, that block the
 > ICMP messages coming back to your server.
 >
 > /Jesper
 >
 > --
 > Jesper Skriver, jesper(at)skriver(dot)dk  -  CCIE #5456
 > Senior network engineer @ AS3292, TDC Tele Danmark
 >
 > One Unix to rule them all, One Resolver to find them,
 > One IP to bring them all and in the zone to bind them.
 >
 > To Unsubscribe: send mail to majordomo@FreeBSD.org
 > with "unsubscribe freebsd-net" in the body of the message


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-net" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3D617842.4020304>