[c-nsp] EoMPLS between C7206 and C3845

Justin Shore justin at justinshore.com
Thu Sep 11 11:28:50 EDT 2008


Is that an EtherSwitch "Network" module or "Service" module?  They are 
very different beasts.  I'd imagine that you were using the Network 
module and that the problem could have been avoided with a Service module.

http://tinyurl.com/2ok8ox

The Service module literally acts as an independent switch that happens 
to be mounted inside the ISR chassis.  I don't have a solution for your 
EoMPLS problem when using the Network module unfortunately.  Maybe 
someone from Cisco can chime in on that one.

Justin

Junaid wrote:
> Hi,
> 
> I have narrowed the problem. Now EoMPLS is working between the two
> routers - the change is that instead of connecting CE2 to the
> EtherSwitch module of C3845, I have connected it on an external 2950
> switch which is then dot1q trunked to C3845. The problem appears when
> I connect the host on the EtherSwitch port. The configuration on the
> routing portion of C3845 is exactly same in both cases and the config
> on the 2950 and EtherSwitch is similar. Does anyone has any experience
> of running EoMPLS on C3845 with a host on an EtherSwitch module port?
> Is there any special consideration that needs to be catered for in
> such a scenario?
> 
> Will appreciate any help.
> 
> Regards,
> Junaid
> 
> 
> On Thu, Aug 7, 2008 at 2:15 AM, Junaid <junaid.x86 at gmail.com> wrote:
>> Hi,
>>
>> I am trying to make EoMPLS (VLAN mode) to work between a 7206VXR
>> (NPE400) running c7200-jk9s-mz.123-21.bin and a 3845 running
>> c3845-advipservicesk9-mz.124-15.T.bin. These two PE routers are
>> connected back-to-back via FastEthernet. The customers are connected
>> via a switch connected to each PE:
>>
>> CE1 --- Switch --- PE1 --- PE2 --- Switch --- CE2
>>
>> The control place comes up without any issue:
>>
>> C7200-PE1#sh mpls l2transport vc de
>> Local interface: Fa0/0.3 up, line protocol up, Eth VLAN 3 up
>>  Destination address: XXXXX (loopback ip of PE2), VC ID: 100, VC status: up
>>    Next hop: XXXXXX (ip of PE2's interface connected with PE1)
>>    Output interface: Fa3/0, imposed label stack {234}
>>  Create time: 04:55:52, last status change time: 04:22:07
>>  Signaling protocol: LDP, peer XXXXX (loopback ip of PE2):0 up
>>    MPLS VC labels: local 2207, remote 234
>>    Group ID: local 0, remote 0
>>    MTU: local 1500, remote 1500
>>    Remote interface description: MPLS TEST
>>  Sequencing: receive disabled, send disabled
>>  VC statistics:
>>    packet totals: receive 658, send 558
>>    byte totals:   receive 61117, send 57759
>>    packet drops:  receive 0, send 0
>>
>>
>> C3845-PE2#sh mpls l2transport vc de
>> Local interface: Gi4/0.3 up, line protocol up, Eth VLAN 3 up
>>  Destination address: XXXXX (loopback ip of PE1), VC ID: 100, VC status: up
>>    Next hop: XXXXXX (ip of PE1's interface connected with PE2)
>>    Output interface: Gi0/0, imposed label stack {2207}
>>  Create time: 05:06:06, last status change time: 04:42:00
>>  Signaling protocol: LDP, peer XXXXX (loopback ip of PE1):0 up
>>    MPLS VC labels: local 234, remote 2207
>>    Group ID: local 0, remote 0
>>    MTU: local 1500, remote 1500
>>    Remote interface description: MPLS test
>>  Sequencing: receive disabled, send disabled
>>  VC statistics:
>>    packet totals: receive 807, send 697
>>    byte totals:   receive 81235, send 63925
>>    packet drops:  receive 0, seq error 0, send 0
>>
>>
>> But the data plane is having severe issue. I cannot ping end-to-end
>> from the CEs. It seems that when I ping CE1 from CE2 (i.e. from the CE
>> connected to 3845), ARP works and I am able to send a ping packet to
>> CE1. But CE1 never receives it. On the other side, CE2  does not get
>> replies to its own ARP requests. Once I statically bind the mac
>> address of CE2 on CE1, CE1 sends an ICMP packet to CE2 and CE2 replies
>> to it but CE1 never receives the reply. It seem that the communication
>> is one way, from CE1 (one behind C7206) to CE2 (one behind C3845) and
>> not the other way round. I replaced C3845 with C7206 and there was not
>> issue in the data plane.
>>
>> My question is with the IOS I used for C3845, is EoMPLS not supported
>> on it? As per Cisco's documentation, EoMPLS is supported on the IOS I
>> used for C3845. Any one any experience in running EoMPLS on C3845?
>>
>> Another thing I noted was in the following output from C3845, it shows
>> MRU=0 and also there was no outgoing interface attached:
>>
>> C3845-PE2#sh mpls forwarding-table labels 234 detail
>> Local  Outgoing    Prefix            Bytes tag  Outgoing   Next Hop
>> tag    tag or VC   or Tunnel Id      switched   interface
>> 234                l2ckt(100)        50732      none       point2point
>>        MAC/Encaps=0/0, MRU=0, Tag Stack{}
>>        No output feature configured
>>
>> While on C7206, the output was as it should be:
>>
>> C7200-PE1#sh mpls forwarding-table labels 2207 detail
>> Local  Outgoing    Prefix            Bytes tag  Outgoing   Next Hop
>> tag    tag or VC   or Tunnel Id      switched   interface
>> 2207   Untagged    l2ckt(100)        55853      Fa0/0.3    point2point
>>        MAC/Encaps=0/0, MRU=1500, Tag Stack{}
>>        No output feature configured
>>
>>
>> Any explanations/solutions?
>>
>>
>>
>> Regards,
>>
>> Junaid
>>
> _______________________________________________
> 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/


More information about the cisco-nsp mailing list