50% packet loss on GRE tunnel

[Europacom.net]

Good Morning

We have a site were we have a 887 cisco hsiung a GRE tunnel

When pinging anything froom the tunnel source to the tunnel dest on the local lan we get a  50% packet loss

ping

Protocol [ip]:

Target IP address: 10.9.93.22

Repeat count [5]: 100

Datagram size [100]:

Timeout in seconds [2]:

Extended commands [n]:

Sweep range of sizes [n]:

Type escape sequence to abort.

Sending 100, 100-byte ICMP Echos to 10.9.93.22, timeout is 2 seconds:

!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.

!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.

Success rate is 50 percent (50/100), round-trip min/avg/max = 48/51/96 ms

debug ip icmp detail output

001640: *Sep 20 04:55:01.158 PCTime: FIBipv4-packet-proc: route packet from Vlan1 src 10.9.93.22 dst 172.22.0.1

001641: *Sep 20 04:55:01.158 PCTime: FIBfwd-proc: Default:172.22.0.0/16 proces level forwarding

001642: *Sep 20 04:55:01.158 PCTime: FIBfwd-proc: depth 0 first_idx 0 paths 1 long 0(0)

001643: *Sep 20 04:55:01.158 PCTime: FIBfwd-proc: try path 0 (of 1) v4-ap-Tunnel1 first short ext 0(-1)

001644: *Sep 20 04:55:01.158 PCTime: FIBfwd-proc: v4-ap-Tunnel1 valid

001645: *Sep 20 04:55:01.158 PCTime: FIBfwd-proc: Tunnel1 no nh type 3  - deag

001646: *Sep 20 04:55:01.162 PCTime: FIBfwd-proc: ip_pak_table 0 ip_nh_table 65535 if Tunnel1 nh none deag 1 via fib 0 path type attached prefix

001647: *Sep 20 04:55:01.162 PCTime: FIBfwd-proc: Default:172.22.0.0/16 not enough info to forward via fib (Tunnel1 none)

001648: *Sep 20 04:55:01.162 PCTime: FIBipv4-packet-proc: packet routing failed

001649: *Sep 20 04:55:01.162 PCTime: IP: tableid=0, s=10.9.93.22 (Vlan1), d=172.22.0.1 (Tunnel1), routed via RIB

001650: *Sep 20 04:55:01.162 PCTime: IP: s=10.9.93.22 (Vlan1), d=172.22.0.1 (Tunnel1), len 100, output feature

001651: *Sep 20 04:55:01.162 PCTime:     ICMP type=0, code=0, TCP Adjust MSS(40), rtype 1, forus FALSE, sendself FALSE, mtu 0, fwdchk FALSE

001652: *Sep 20 04:55:01.162 PCTime: IP: s=10.9.93.22 (Vlan1), d=172.22.0.1 (Tunnel1), len 100, output feature

001653: *Sep 20 04:55:01.162 PCTime:     ICMP type=0, code=0, Post-Ingress-NetFlow(52), rtype 1, forus FALSE, sendself FALSE, mtu 0, fwdchk FALSE

001654: *Sep 20 04:55:01.162 PCTime: IP: s=10.9.93.22 (Vlan1), d=172.22.0.1 (Tunnel1), len 100, sending full packet

001655: *Sep 20 04:55:01.162 PCTime:     ICMP type=0, code=0

this one has us puzzled.. any ideas would be really helpful.

replicated the setup exatly in our lab  and it works 100% with exactly the same config.

Andre

[cadet alain]

Hi,

could you post your config.

Regards

Alain

Don't forget to rate helpful posts.

[Richard Burts]

Andre

Some things in your post have me puzzled. You show the ping command with this

Target IP address: 10.9.93.22

but in the debug output Target IP address: 10.9.93.22 appears as the source and the destination is 172.22.0.1.

Can you provide clarification about the relationship of the command that you show and the debug?

Another thing that I notice is that the debug seems to show attempts to forward two packets to the same destination

001647: *Sep 20 04:55:01.162 PCTime: FIBfwd-proc: Default:172.22.0.0/16 not enough info to forward via fib (Tunnel1 none)

001654: *Sep 20 04:55:01.162 PCTime: IP: s=10.9.93.22 (Vlan1), d=172.22.0.1 (Tunnel1), len 100, sending full packet

So why does one fail and the other succeed with the same destination address. It seems that something in going on with the routing logic.

This makes me quite agree with Alain that it would be helpful if you would post the configuration. I wold also like to see the output of show ip route - or if the routing table is too large then the output of show ip route 10.9.93.22 and of show ip route 172.22.0.1.

HTH

Rick