Cisco Support Community
Showing results for 
Search instead for 
Did you mean: 
New Member

MPLS VRFs hanging routes

Hi all,

We've a cell-based MPLS network (based on BPX 8600/LSC 7200 acting as the P and MGXs with RPMs acting as the PEs and connected with E3s to the BPX).

On those PEs...we're running MPLS VPNs for our customers and there're 2 PEs acting as Route Reflectors for all the other PEs for reflecting the MP-BGP routes for the VRFs.

The problem is that with any RPM reloads or any interface flapping or without any reason....all of a sudden we found that a VRF customer that has for example 2 of them connected to POPX and the other branch connected to POPY complaining that there's no connectivity bet the 2 branches although when issuing the command " sh ip route vrf Customer AAA " on the PE of POPX we found that the IBGP routes of the other branch are present in its VRF routing table.....but still the 2 branches cannot ping each other.

The same problem may be repeated for all VRF customers connected bet those 2 POPs and aren't solved except when issuing the command on the PE of POP X "clear ip route (lpbk add of the PE in POPY)"

After that command....everything is OK and the 2 branches can ping each other without problems.

After some investigation...we found that this problem is due to an LSC bug....the suspected bugs were CSCea21665 and CSCea74222 and the workaround for those bugs are "clear ip route (Remote PE lpbk add)"

As listed in those bugs also that the fix for them is in IOS 12.2(15)T05 and we upgraded our LSC from ver 12.2(8)T4 to the latest


Unfortunately we found that the problem is not yet solved and still the same syptoms appers for the VRFs.....and that mean that upgrading the IOS ver for the LSc is not enough and there's a step yet missing for avoiding that fatal problem.

So has anyone faced this problem before ??? and if yes what were the steps done to avoid it other than the famous workaround "clear ip route (Remote PE lpbk add)"???

New Member

Re: MPLS VRFs hanging routes


I red your problem, because I'm interested on all the WAN switching staff.

Look at bug CSCea21665 on CCO, the fix is not integrated in 12.2 main line, so you have to go to one of the following minimum IOS 12.2(15)T05, 12.2(17.6)S, 12.3(1.9), 12.3(1.9)T, 12.0(25.3)S01, 12.2(11)T09, 12.2(15)ZK, 12.3(2.3)B, 12.2(15)ZK01.

Look at Bug CSCea74222, it's fixed in

12.2(15)T03, 12.3(1.5), 12.3(1.5)T, 12.2(17.3)S, 12.2(15)ZK, 12.3(2.3)B

From that two bugs, do not use 12.2 main line, the fix is not integrated.

Don't use 12.3, it's to new ;-))

I would recommend 12.2(15)T05 or higher, that means 12.2(15)T07

Than you shouldn't see the problem again.



CreatePlease to create content