Re: LDP Error Message

yuening liao · ‎02-13-2011

Hello Everyone,

i ran the LDP on OSPF within BGP, both ospf and bgp connections were up. the loopback interfaces between two peers were also pingable each other. but LDP neighbor came never up. With Debug mpls ldp transport connection was following log output showed

*Mar 1 00:16:57.271: ldp: Opening ldp conn; adj 0x6514BEC8, 10.1.1.3 <-> 10.1.1.1; with normal priority
*Mar 1 00:16:57.279: ldp: MD5 setup for neighbor 10.1.1.1; password changed to [nil]
*Mar 1 00:16:57.283: ldp: Registered TCB with LDP TCB database tcb 0x64885614 [key 1017], total 2
*Mar 1 00:16:57.391: ldp: Conn failed (TCP activity)!; adj 0x6514BEC8, 172.14.8.5, tcb state 0x0
*Mar 1 00:16:57.395: ldp: Unregistered from LDP TCB database tcb 0x64885614 [key 1017], total 1
*Mar 1 00:16:57.395: ldp: Closing ldp conn 10.1.1.3:59945 <-> 10.1.1.1:646, adj 0x6514BEC8

anybody familiar with this case?

thank you very much in advance!

Reza Sharifi · ‎02-13-2011

Hi,

It seems you have authentication enabled, if yes it has to be enabled on both sides.

disable the authentication and see if the neighbor comes up

post "sh mpls ldp dis"

also, have a look at this doc on how to configure LDP

http://www.cisco.com/en/US/docs/ios/12_4t/12_4t2/ftldp41.html#wp1354703

HTH

Reza

yuening liao · ‎02-13-2011

Hi Reza,

thank you for your reply,

R1(config)#do sh mpls ldp dis
Local LDP Identifier:
    10.1.1.1:0
    Discovery Sources:
    Interfaces:
        Ethernet0/0 (ldp): xmit
        Ethernet0/2 (ldp): xmit

R3(config)#do sh mpls ldp dis
Local LDP Identifier:
    10.1.1.3:0
    Discovery Sources:
    Interfaces:
        Ethernet0/0 (ldp): xmit
        Ethernet0/2 (ldp): xmit
R3(config)#

and i find no configuration but following commands concerning ldp on both devices:

ip cef

mpls label protocol ldp

interface Ethernet0/2
mpls ip

mpls ldp router-id Loopback0

R3#p 10.1.1.1

Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 10.1.1.1, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 32/41/60 ms
R3#

Reza Sharifi · ‎02-13-2011

Hi,

Are both loopback addresses reachable?

did you enable mpls globally by using "ip mpls"?

also delete this command and see if the neighbor come up

mpls ldp router-id Loopback0

HTH

Reza

yuening liao · ‎02-15-2011

Hi,

I deleted the command "mpls ldp router-id Loopback0" enabled mpls ip both globally and on interfaces between those two direktly connected routers.

two loopback interface are perfectly reachable each other, but still no ldp neighbor comes up......

below is the output after entering "debug mpls ldp transport event" and ""debug mpls ldp transport connection"

R3.

*Mar 1 00:08:58.727: ldp: Send ldp hello; Ethernet0/2, src/dst 172.14.8.6/224.0.0.2, inst_id 0
*Mar 1 00:09:01.815: ldp: Rcvd ldp hello; Ethernet0/2, from 172.14.8.5 (10.1.1.1:0), intf_id 0, opt 0xC
*Mar 1 00:09:01.819: ldp: ldp Hello from 172.14.8.5 (10.1.1.1:0) to 224.0.0.2, opt 0xC
*Mar 1 00:09:01.823: ldp: New adj 0x6482DE40 for 10.1.1.1:0, Ethernet0/2
*Mar 1 00:09:01.823: ldp: adj_addr/xport_addr 172.14.8.5/10.1.1.1
*Mar 1 00:09:01.827: ldp: local idb = Ethernet0/2, holdtime = 15000, peer 172.14.8.5 holdtime = 15000
*Mar 1 00:09:01.827: ldp: Link intvl min cnt = 2, intvl = 5000, idb = Ethernet0/2
*Mar 1 00:09:01.831: ldp: Opening ldp conn; adj 0x6482DE40, 10.1.1.3 <-> 10.1.1.1; with normal priority
*Mar 1 00:09:01.835: ldp: Found adj 0x6482DE40 for 10.1.1.1 (Hello xport addr opt)
*Mar 1 00:09:01.839: ldp: MD5 setup for neighbor 10.1.1.1; password changed to [nil]
*Mar 1 00:09:01.843: ldp: Registered TCB with LDP TCB database tcb 0x64885BC0 [key 541], total 1
*Mar 1 00:09:01.895: ldp: Conn failed (TCP activity)!; adj 0x6482DE40, 172.14.8.5, tcb state 0x0
*Mar 1 00:09:01.899: ldp: Unregistered from LDP TCB database tcb 0x64885BC0 [key 541], total 0
*Mar 1 00:09:01.903: ldp: Closing ldp conn 10.1.1.3:12976 <-> 10.1.1.1:646, adj 0x6482DE40
*Mar 1 00:09:01.907: ldp: Adj 0x6482DE40; state set to closed

R1.

*Mar 1 00:07:55.531: ldp: Rcvd ldp hello; Ethernet0/2, from 172.14.8.6 (10.1.1.3:0), intf_id 0, opt 0xC
*Mar 1 00:07:55.803: ldp: Send ldp hello; Ethernet0/2, src/dst 172.14.8.5/224.0.0.2, inst_id 0
*Mar 1 00:08:00.431: ldp: Rcvd ldp hello; Ethernet0/2, from 172.14.8.6 (10.1.1.3:0), intf_id 0, opt 0xC
*Mar 1 00:08:00.559: ldp: Send ldp hello; Ethernet0/2, src/dst 172.14.8.5/224.0.0.2, inst_id 0

the debug output of R3 seems related to LDP authentication. but actually no commad of LDP Auth were enabled.

Peter Paluch · ‎02-15-2011

Hello,

Is it possible that the R1 uses some kind of ACL or traffic filtering that would prevent it from accepting the LDP TCP connection from R3? Can you use the debug ip icmp command to see if there are any unreachables being sent?

Best regards,

Peter

yuening liao · ‎02-15-2011

Hello,

not at all.

Only those neccesary protocols BGP, OSPF, LDP run on the two peers. neighborhood of all running protocols but ldp come rapidly up once powered on. no more special policy were made.

and there are the debug ip icmp output:

R3(config-if)#do deb ip ic
ICMP packet debugging is on
R3(config-if)#do p 10.1.1.1 so 10.1.1.3

Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 10.1.1.1, timeout is 2 seconds:
Packet sent with a source address of 10.1.1.3
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 20/43/68 ms
R3(config-if)#
*Mar 1 00:03:31.655: ICMP: echo reply rcvd, src 10.1.1.1, dst 10.1.1.3
*Mar 1 00:03:31.719: ICMP: echo reply rcvd, src 10.1.1.1, dst 10.1.1.3
*Mar 1 00:03:31.759: ICMP: echo reply rcvd, src 10.1.1.1, dst 10.1.1.3
*Mar 1 00:03:31.787: ICMP: echo reply rcvd, src 10.1.1.1, dst 10.1.1.3
*Mar 1 00:03:31.811: ICMP: echo reply rcvd, src 10.1.1.1, dst 10.1.1.3
R3(config-if)#

by the way, i run this on GNS3. Could it be a IOS bug or some kind error of compatibility?

Luc De Ghein · ‎02-15-2011

Hi,

What does "show mpls ldp discovery detail" and "show mpls interface detail" show, from both sides?

Thanks,

Luc

Peter Paluch · ‎02-15-2011

Hi,

I was not exactly interesting in pings but rather to seeing whether it is possible that the R1 is refusing the connections - in such case, an ICMP Unreachable would be generated, probably with Host unreachable code/type (in the case of an ACL in place).

If you are using GNS3, you can then capture the traffic on the link between R3 and R1 and check everything that is being received and sent. Does the packet capture reveal anything significant? Do you see the R3 or R1 establishing a TCP connection to the other router to the LDP service? Is the TCP establishement successful?

Best regards,

Peter

yuening liao · ‎02-15-2011

Dear all,

R1(config-if)#do sh mpls ldp dis
Local LDP Identifier:
    10.1.1.1:0
    Discovery Sources:
    Interfaces:
        Ethernet0/2 (ldp): xmit
R1(config-if)#
*Mar 1 00:00:51.307: %OSPF-5-ADJCHG: Process 1, Nbr 10.1.1.3 on Ethernet0/2 from LOADING to FULL, Loading Done
*Mar 1 00:01:17.547: %BGP-5-ADJCHANGE: neighbor 10.1.1.3 Up
R1#

R3(config)#do sh mpls
*Mar 1 00:00:51.915: %OSPF-5-ADJCHG: Process 1, Nbr 10.1.1.1 on Ethernet0/2 from LOADING to FULL, Loading Doneldp dis
Local LDP Identifier:
    10.1.1.3:0
    Discovery Sources:
    Interfaces:
        Ethernet0/2 (ldp): xmit
R3(config)#
*Mar 1 00:01:18.215: %BGP-5-ADJCHANGE: neighbor 10.1.1.1 Up

i haven't captuer traffic yet, but i'm sure that no acl to block tcp traffic is configured and the tcp connection is fine, otherwise the bgp can't come up.

but there is a very suspected phenomena on R1 that it often carshs down once powered on, rapidly and reaptly display those error message like that until you powered off the device:

%ALIGN-1-FATAL: Corrupted program counter 00:01:31 UTC Fri Mar 1 2002
pc=0xBFC011A0 , ra=0xBFC011A0 , sp=0x63A29A98

%ALIGN-1-FATAL: Corrupted program counter 00:01:31 UTC Fri Mar 1 2002
pc=0xBFC011A0 , ra=0xBFC011A0 , sp=0x63A28F28

thank you.

Luc De Ghein · ‎02-15-2011

Hi,

The "show mpls ldp disc" command shows that you are not receiving LDP link hello's from the

neighboring router, on both sides.It only shows "xmit", but not "recv".

These packets are UDP packets. Are you blocking them?

Once this is fixed, make sure that you have a host route (/32) to the LDP router ID of the

neighboring router.

Thanks,

Luc

shivjain · ‎02-19-2011

You can see only xmit in discovery command which means neighbouring router is not sending any hello ldp packets. Plz chk on neighbouring router whether the mpls is enabled or not.

Regards

Shivlu Jain

Sent from Cisco Technical Support iPhone App