[c-nsp] More EoMPLS / c7200 gripes
David Freedman
david.freedman at uk.clara.net
Thu Aug 31 07:59:26 EDT 2006
Ok, so , I've more EoMPLS implementation gripes with c7200
NPE-G1 box running 12.3 (mainline), want to use a PA-FE-TX for
customer's attachment circuit.
Come across the following:
router# sh mpls l2transport hw-capability interface f3/0
<snip>
Transport type Eth VLAN
Core functionality:
MPLS label disposition supported
Control word processing supported
Sequence number processing not supported
Edge functionality:
MPLS label imposition supported
Control word processing supported
Sequence number processing not supported
Transport type Ethernet
Core functionality:
MPLS label disposition supported
Control word processing supported
Sequence number processing not supported
Edge functionality:
Not supported <==== Huh?
</snip>
This is telling me essentially that I can ONLY use the PA-FE-TX
for a VLAN mode attachment circuit and not port mode!
Does anybody know the technical reason for this?
It seems the wrong way around from a logical point of view, surely if
there is a restriction in terms of the PA capability it would be the
opposite (i.e VLAN mode is too much work and only port mode should be
supported?)
To add more confusion to this, the G1 ports themselves (g0/1,2,3) seemed
to have the same restriction!
router# sh mpls l2transport hw-capability interface g0/1 | begin type
Ethernet
Transport type Ethernet
Core functionality:
MPLS label disposition supported
Control word processing supported
Sequence number processing not supported
Edge functionality:
Not supported <==== Again
So anyway, we still need a port mode xconnect, or rather, the ability
for the customer to pass in untagged traffic and receive it untagged the
other end , so came up with the following hack on both PE routers:
!
interface FastEthernet X/Y
description To Customer
!
!
interface FastEthernet X/Y.100
encapsulation dot1q 100 native
xconnect 192.168.0.1 100 encapsulation mpls
!
So, xconnect comes up, but no traffic passes.
I expected this to be the case.
I noted from the output of "show mpls l2transport vc 100 detail"
that traffic was being received and none sent, such:
router#show mpls l2transport vc 100 detail
Local interface: F3/0.100 up, line protocol up, Eth VLAN 100 up
Destination address: 192.168.10.1, VC ID: 100, VC status: up
Next hop: 10.10.10.12
Output interface: Gi0/1, imposed label stack {84 340}
Create time: 11:03:20, last status change time: 11:03:15
Signaling protocol: LDP, peer 192.168.10.1:0 up
MPLS VC labels: local 26, remote 340
Group ID: local 0, remote 0
MTU: local 1500, remote 1500
Remote interface description: Its not going to work, is it
Sequencing: receive disabled, send disabled
VC statistics:
packet totals: receive 2410, send 0
byte totals: receive 23583, send 0
packet drops: receive 0, send 0
Started all the l2vpn debugs I could think of and nothing yielded
anything useful.
Whilst I accept that this platform is not really geared to Ethernet edge
functionality such as this (and that I would probably have more luck
burning the inbuilt G1 ports up to do so) I'm just left feeling a bit
bewildered as to why such functionality can be configured, but not work.
I did, eventually find the following caveat:
"–The native VLAN of a trunk must not be configured as an EoMPLS VLAN."
buried deep inside a Catalyst MPLS/L2VPN configuration document here:
http://www.cisco.com/en/US/products/hw/switches/ps708/products_configuration_guide_chapter09186a008029d19e.html
So am guessing that this caveat applies globally to all EoMPLS
implementations?
Dave.
More information about the cisco-nsp
mailing list