PPP over FR using virtual-template !

Unanswered Question
May 22nd, 2009

hi all, can some one tell me why i cant ping my own IP when using PPP over FR through virtual access interfaces ? is there any document that addresses this issue ?

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
rakesh.hegde Mon, 05/25/2009 - 10:49


This is a known issue. You can unnumber the virtual template or use a Multilink interface as a work around.

Have you had a chance to google this ?



illusion_rox Tue, 05/26/2009 - 03:42

Dear Rakesh, i have googled it as well as searched a lot on cisco site as well but wasnt able to find any explanation to this behaviour :(. When i do debug from pinging myself, i can see packets going to virtual-access just like they do for remote IPs, but i cant get the reply,

Do you think i should search in some tac repository ? i have my company account so if you can tell me the link may be there i will find answer to this behaviour :(

Giuseppe Larosa Tue, 05/26/2009 - 13:46

Hello Ovais,

as explained in the document linked by Rakesh pinging your own ip address works differently on wan interfaces then on ethernet interfaces.

LAN interfaces ping themselves without sending a packet on the wire.

Actually if you take an unplugged ethernet of fastethernet you configure

int eth0

ip address

no keepalive

you can then ping it even if it is unplugged I've used this trick on lab tests some years ago.

With serial interface, ATM, FR the things are different:

pinging your own ip address means the packet is sent out on wires: the icmp echo request reach the remote end DTE that "reflects back" the frame to local router interface.

I saw this clearly the first time I did Netflow tests: by comparing interface counters, SNMP interface counters and Netflow data all was multiplied by 2 when pinging the ip address of an ATM interface from itself.

On all this trip when you do

PPP over FR you cannot solve this by simply mapping on FR your own ip address as it can be done on a FR only multipoint interface.

Because you now have:

IP/PPP/FR you cannot map ip over FR.

to be noted that if you can ping the other side all is well.

By the way you have alternate ways to verify that connectivity is good:

you can check the virtual-template to see if the ipcp phase is in open state meaning a successful IPCP negotiation has been performed on the logical PPP channel.

Hope to help


rakesh.hegde Tue, 05/26/2009 - 14:09

Hi Giuseppe,

I dont think the inability to ping has got anything to do with DLCI. We would see the same behaviour if a multipoint interface is used for frame relay , but mapings for local ip doesnt make sense in this situation since we dont have an ip assigned . I do agree with you that this is due to double L2 encapsulation.

Am I missing something ?


illusion_rox Tue, 05/26/2009 - 18:44

Dear Sir, thanks alot for the feedback. Actually i was preparing for my CCIE so i thought i should try my level best to know why certain things dont work as they should. I know there are workarounds for this, all i wanted to get some logic behind as to why virtual-access usage prevents the local ping ?


This Discussion