Problem with BGP signaled VPLS

RouterOS Forwarding Protocols
1

I’m running into issues with BGP signaled VPLS failing to re-establish following a reboot (by both graceful reboot and pulling the plug). OSPF, BGP, and LDP fully establish, and MPLS traffic will flow following the reboot. The VPLS dynamic interface shows up and running on the router that rebooted, but do not on the remote PEs that terminate the other end. This came to light after setting up a client who has had some power issues at some of their sites. I’ve reproduced this on test equipment and it consistently fails to re-establish after a reboot.

My test environment is using one of our core routers (Juniper MX) for the P so that I have full routing tables from OSPF and access to our route reflectors. Aside from that, it’s pretty simple (PE1 → P → PE2).

This happens with both RB1100s and CloudCores. RouterOS versions between 6.7 and 6.19.

Has anybody else experienced this?

2

I have also experienced this, but only in select cases. I truly have not spent time diagnosing the problem as it’s only happened 1 or 2 times and I do not have consistent failure. I can just disable and enable the VPLS tunnel on the PE and it comes back up. Curious as to what others have to say.