[c-nsp] L2TPv3 with MTU difference

[Rens]

Hi Ahmad,

When I do the same test but lower the MTU again on the physical interface
towards the L2TPv3 tunnel (MTU 1500), so that they are identical again, the
1518 frames pass again with the testers. (Fragmentation)

So I'm more thinking when the MTU is different between the 2 routers, it
drops all the frames that are too big? Shouldn't the sender start to
fragment?

Here some tests:

7200#sh int GigabitEthernet 0/2.801
GigabitEthernet0/2.801 is up, line protocol is up 
  Hardware is BCM1250 Internal MAC, address is 0006.d6ca.481a (bia
0006.d6ca.481a)
  Description: #backbone: VLAN801-dot1q 
  Internet address is 80.91.151.1/27
  MTU 1500 bytes, BW 1000000 Kbit, DLY 10 usec, 
     reliability 255/255, txload 2/255, rxload 2/255
  Encapsulation 802.1Q Virtual LAN, Vlan ID  801.
  ARP type: ARPA, ARP Timeout 04:00:00
  Last clearing of "show interface" counters never

r1841#sh int Fa0/0.801
FastEthernet0/0.801 is up, line protocol is up 
  Hardware is Gt96k FE, address is 0023.0447.4ce4 (bia 0023.0447.4ce4)
  Description: #backbone: VLAN801-dot1q [100M]
  Internet address is 80.91.151.3/27
  MTU 1500 bytes, BW 100000 Kbit, DLY 100 usec, 
     reliability 255/255, txload 25/255, rxload 13/255
  Encapsulation 802.1Q Virtual LAN, Vlan ID  801.
  ARP type: ARPA, ARP Timeout 04:00:00
  Last clearing of "show interface" counters never
r1841#

r1841#ping 80.91.151.1 size 1518

Type escape sequence to abort.
Sending 5, 1518-byte ICMP Echos to 80.91.151.1, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 1/3/4 ms
r1841#

When I now higher the MTU on 7200 to 1600, the pings aren't being
fragmented:

r1841#ping 80.91.151.1 size 1518

Type escape sequence to abort.
Sending 5, 1518-byte ICMP Echos to 80.91.151.1, timeout is 2 seconds:
.....
Success rate is 0 percent (0/5)
r1841#



-----Original Message-----
From: Cheikh-Moussa Ahmad [mailto:acm at axians.de] 
Sent: lundi 4 mai 2009 9:45
To: Rens
Subject: AW: [c-nsp] L2TPv3 with MTU difference

Hi Rens,

1518 is too big for a FastEthernet Interface. Did you tried 1480 ?
20 Bytes for udp and l2tp header should you removed from the max frame
size.

Regards, 
 Ahmad


> -----Ursprüngliche Nachricht-----
> Von: cisco-nsp-bounces at puck.nether.net [mailto:cisco-nsp-
> bounces at puck.nether.net] Im Auftrag von Rens
> Gesendet: Montag, 4. Mai 2009 09:19
> An: cisco-nsp at puck.nether.net
> Betreff: Re: [c-nsp] L2TPv3 with MTU difference
> 
> Nobody that can help me with this?
> 
> -----Original Message-----
> From: cisco-nsp-bounces at puck.nether.net
> [mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of Rens
> Sent: vendredi 17 avril 2009 14:43
> To: cisco-nsp at puck.nether.net
> Subject: [c-nsp] L2TPv3 with MTU difference
> 
> Hi,
> 
> 
> 
> I have an OSPF broadcast configured with several routers.
> 
> 
> 
> Some of the routers have a higher MTU then others so I use ip ospf mtu
> ignore on all the neighbours. (to compensate with the fragmentation at
> higher bandwidths)
> 
> 
> 
> I have routers with mtu 1600 and others have the default 1500 because
> of
> FastEthernet interfaces
> 
> 
> 
> I have a L2TPv3 tunnel that runs over this IP network, when I configure
> a
> tunnel between a router that has 1600 & 1500 mtu I can't pass any
> frames of
> 1518
> 
> 
> 
> When doing 1518, the tester that is connected to the router that does
> 1600
> is receiving them, but the tester that is connected to the router that
> does
> 1500 isn't receiving anything.
> 
> 
> 
> When I lower it to 1280 it works again.
> 
> 
> 
> All help welcome
> 
> 
> 
> Regards,
> 
> 
> 
> Rens
> 
> 
> 
> 
> 
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
> 
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/


Sitz der NK Networks & Services GmbH: Von-der-Wettern-Straße 15, 51149 Köln
Registergericht: Amtsgericht Köln, Registernummer HRB 30805
Geschäftsführer: Tonis Rüsche