traceroute double hop

[kent.plummer]

Can anyone explain the following traceroute from the net into our network...

IP's are not the actual. Note double hop 5 & 6 - which is the hop after the last MPLS label pop.

Performing trace to 111.50.208.137...

1 FastEth8-1-0.sb1.optus.net.au (192.168.34.2) 3.385 ms 2.941 ms 2.783 ms

2 61.88.240.109 (61.88.240.109) 18.069 ms 18.118 ms 18.133 ms

3 61.88.240.109 (61.88.240.109) 18.109 ms 18.044 ms 17.965 ms

4 uecommlimited.14vrc76f06.optus.net.au (59.154.4.106) 18.446 ms 18.768 ms 18.574 ms

5 v10.pe1.dc1-vars.ourdomain.com (111.50.213.126) 20.346 ms 20.379 ms 20.175 ms

6 v10.pe1.dc1-vars.ourdomain.com (111.50.213.126) 20.260 ms 20.158 ms 20.293 ms

7 111.50.208.137 (111.50.208.137) 19.663 ms 18.757 ms 18.357 ms

We have "no mpls ip propagate-ttl forwarded" on all P's and PE's in our network. When I turn ttl propagation back on the traffic is only hitting this device once.

[Harold Ritter · Accepted solution]

Kent,

Is the device at # 4 a 6500 with a sup720? If so, it could be due to CSCef16357.

Regards

[sschulak]

on hop 4 look at entry for prefix 111.50.208.137/? in rib and lfib. then compare the same prefix on hop 5 in rib and lfib to see if prefix length is different.

I ran into a simular situation but in your case this router probaly is going to a FW

with weird icmp rules

show ip route vrf ??? 111.50.208.137

sho mpls for vrf ??? 111.50.208.137 /? det

[kent.plummer]

Harold is correct in that we have run into bug CSCef16357.

I confirmed with the TAC that all PFC3 based SUPs are affected by this bug.

There is no fix and it seems there is none planned.