[c-nsp] MPLS over MLPPP Bug

[Antonio Soares]

I'm trying to find a matching bug id for this situation:

The LDP session over MLPPP goes down if the MLPPP bundle is congested and at
least one of the ML members goes down.

I found something similar but that only affects the 2941:

+++++++++++++++++++++++++++++++++
CSCtc31618 Bug Details 
2941: LDP session over MLPPP stays down even after congestion is over

Symptom:
LDP session over MLPPP stays down even if the congestion over MLPPP is over.


Conditions:
When user configures static routes to create the LSP for MPLS over MLPPP,
and MLPPP is congested, LDP session goes down, but if congestion is removed
or there is no congestion, still LDP stays down. 

Workaround:
Workaround is to clear the MLPPP interface using "clear multilink < MLP no>
+++++++++++++++++++++++++++++++++

In my case, the LDP session recovers after a few seconds:

R1:
*Mar 16 16:11:36.407: %LDP-5-NBRCHG: LDP Neighbor 2.2.2.2:0 (1) is DOWN
(Received error notification from peer: Holddown time expired)
*Mar 16 16:11:45.191: %LDP-5-NBRCHG: LDP Neighbor 2.2.2.2:0 (2) is UP

R2:
*Mar 16 16:11:36.555: %LDP-5-NBRCHG: LDP Neighbor 1.1.1.1:0 (1) is DOWN
(Discovery Hello Hold Timer expired)
*Mar 16 16:11:45.339: %LDP-5-NBRCHG: LDP Neighbor 1.1.1.1:0 (1) is UP

It's easy to reproduce the problem. I was able to do it with dynamips. Both
routers running 12.4.24T.


Thanks.

Regards,

Antonio Soares, CCIE #18473 (R&S/SP)
amsoares at netcabo.pt
http://www.ccie18473.net