In the last episode (Sep 18), David Preece said:
> I tried this on -questions but to no avail. It strikes me as something that 
> is either dead easy, or can't be done. Does anyone have any ideas?

If "route get <clientip>" shows an MTU of 576, "route delete
<clientip>" should clear it.
 
> I have a network consisting of:
> 
> Server (mtu1500)<->(mtu1500)gateway(mtu576)<->Client (mtu1500)
> 
> The server and gateway are both FreeBSD4.x machines and the client is
> NT4, hence changing MTU is a bit of a mystery, but I digress. So, the
> first time the client connects to the server and the server attempts
> to send back a packet of over 576 bytes, the gateway objects and
> sends back an ICMP 3(4) saying that the packet can't be forwarded
> until the MTU is reduced to 576. The server tries again with the new
> MTU and all is well.
> 
> The kernel on the server also has the good sense to remember that
> there's an MTU of 576 bytes on that route and doesn't try anything
> larger in future. This would be great *except* that I'm trying to
> turn the whole thing into a factory for ICMP cannot fragment
> messages. Is there a simple way to flush the route->mtu table on the
> server?

-- 
        Dan Nelson
        [EMAIL PROTECTED]

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Reply via email to