Help - MPLS/VPLS flapping

Hi Support, I have been goin through the forum for some time but have not been able to solve this problem.

Any assistance or guidance will be appreciated.

We have network with 1 core router [1072] and multiple CCR-1009 routers deployed at sites which does the OSPF routing and MPLS/VPLS for pppoe users.

OSPF we are using PTP from 1072.

The OSPF routers are connected to CSS326 switches with vlans for pppoe users. Problem is at the PPPOE Server Router [1036] all VPLS links flap randomly together. It drops for 3-5 seconds and then all users get connected. But causing packet drops to customers and inconvinience if on calls or IPTV.

I have also gone through the forum and modified the mtu configs, but there is no change.. every hour 3-4 times it flaps.

Attached is the diagram, if someone can shed some light.
vpls_error.jpg

Three things to check for flapping VPLS

  1. Duplication of MAC addresses on VPLS interfaces from copy/paste
  2. Loop Detect is set on some or all routers - disable it
  3. Bridging loop in one or more VPLS instances

Thank you for your prompt response. I check all your suggestions and disabled loop protection on all physical & vlan interfaces connected to the OSPF routers and VPLS.

But it’s still flapping. All our routers are on OS 6.44.6? Do you recommend we upgrade or downgrade?

Because every 50-60 mins all VPLS links drop 5-8 timeouts and then reconnects, where all the physical routers are connected on fiber with no packet loss.

The latest long term version (6.47.10) is stable and contains a number of OSPF and MPLS bug fixes since the 6.44 series.

That’s what we typically put on client networks.

which platform are you using to connect the remote sites to your core-is it RF or fibre? If it is RF that is a 5Ghz-based, do check the quality of the link.

Hi,
I’m having a similar issue if it’s not the same.

Was a resolution ever found?

Perhaps good idea to start by confirm that the OSPF network is stable, run a continuous ping on the OSPF network and see if ping drops when the VPLS tunnels are flapping.

upgrade at 6.48.6 look vpls forwarding table when drop connection

one more thing u can try to detect the issue.
Do tracert (once u experiencing drop) to the remote ip-add (loop) in the vpls.

Check the path , make sure it goes right direction, sometimes can be routed to the end users, and may cost the issue.

@IPANetEngineer
what do u mean by that?

Bridging loop in one or more VPLS instances

more than one vpls to the one end point, is that right?