Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Fragmentation issues with site-to-site VPN

44 views
Skip to first unread message

Dan24

unread,
Jul 2, 2009, 3:01:11 PM7/2/09
to
Hi,

I'm trying to set up a tunnel between a Checkpoint Safe@Office 500
device to a Cisco 837 router. I've used SDM to set up the tunnel and
everything seemed to work and ping shows that the connection is
stable. But I then noticed the connection is very slow and several
services complained of connection failures. Running the test in SDM, I
get the following output:
"A ping with data size of this VPN interface MTU size and 'Do not
Fragment' bit set to the other end VPN device is failing. This may
happen if there is a lesser MTU network which drops the 'Do not
fragment' packets."
SDM also suggests I can use crypto ipsec df-bit clear to resolve this.
I tried lowering the MTU on the Cisco (issued ip mtu 1400 on the
Dialer0 interface) and also tried lowering the MTU on the Checkpoint
device but no still no success.

Although I'm a software developer I'm not very familiar with CLI so
please don't be brief if you suggest making any changes on the
Cisco ;)
Here's the current config:
http://pastebin.com/m6785596

Thanks in advance,

Dan

Thrill5

unread,
Jul 5, 2009, 10:51:11 PM7/5/09
to
MTU issues on tunnels can be tricky to diagnose but you have already gotten
past this hurdle. The problem is that MTU on a TCP connection is
negotiated with the end-points of the connections and is usually 1500 bytes
(the MTU of and Ethernet network) without any regard of the MTU along the
path. If the MTU of a path in the middle of the connection is less than
this, then packets are dropped if the df bit (do-not fragment bit) is set.
The end-points can figure out that the MTU of the path in the is lower than
the end-points using MTU path discovery, but this needs to be supported (or
enabled) on the endpoints. A tunnel puts and IP packet within an IP packet,
so the MTU of the inner packet is 20 bytes smaller than the MTU of the outer
packet. Lowering the MTU on interface doesn't fix the problem. Enabling
"df-bit clear" will fix the problem because it will fragment the packets
even if the do-not-fragment bit is set.

"Dan24" <dlib...@gmail.com> wrote in message
news:694f2362-7169-4d6c...@i6g2000yqj.googlegroups.com...

0 new messages