MPLS OSPF IGP and BGP with VSS CE

[Peter Valdes]

Hi,

When converging from VSS Active switch to Standby switch, it took around 30+ seconds to access the internal routes 10.X.X.X/24 via BGP. We are using OSPF for IGP MPLS backbone and BGP.

My setup:

PPEx Route-Reflector1

||

MPLS PPE1 <----> CE1 VSS Active     ---

                                                              |    |

                                                              |    }10.X.X.X/24

                                                              |    |

MPLS PPE2 <----> CE2 VSS Standby ---

||

PPEy Route-Reflector2

Attached is the CE config and also the loggs during failover.

Failover is tested by applying the "redundancy force-switchover" on the active VSS.

LDP to BGP neighbors come up which I think is causing the long failover. OSPF works fine as I can see that the loopback stays up with minimum packet loss.

Any alternatives to shortening this failover wait time?

[Moh Mogh]

Hello,

I see that you have nsf enabled ospf, but have you configured NSF/SSO (graceful-restart) for LDP and BGP?

I hope this links helps:
http://www.cisco.com/c/en/us/td/docs/ios/mpls/configuration/guide/12_2sy/mp_12_2sy_book/mp_ldp_grace_nsfsso.html

Regards,
Mohammad

[Peter Valdes]

Hi Mohammad,

I will read up on the link you posted and let you know.

Thanks for the info.

Regards

Peter