| | |
| | |
Information: NE40-1 and NE40E-3 are A plane and NE40-2 and NE40E-4 are B plane. IGP protocol is ISIS. Configure MPLS VPN and VPN FRR.
Problem: When the link between PE and P of A plane is down, the service does not rapidly switch over B plane within 500ms but about 1S.
1. Check VPN FRR configuration and it is normal. Check routing table of four PEs and it is found that BkNextHop and BkLabel are normal (NE40E-1 BkNextHop is loopback address of NE40-4.). It indicates that VPN FRR is correctly configured and it is normal.
2. Check routing table and it is found that the next-hop of loopback address of NE40E-4 is NE40Ea (P). After the link from PE to P is down, BkNextHop is valid, but the next-hop is still P node and there is loop. After there is route convergence, the service recovers.
3. Check cost planning of ISIS. It is found that the link between PE and P is 1000, the link between P and P is 5 and the link between PE and PE is 4. There is equal-cost route from NE40E-1 to NE40-4 and their cost is 1003. Change weight of equal-cost route with weight of ISIS to avoid loop.
5. Change weight of ISIS and ensure that loopback address of remote PE of B plane can learn local PE through B plane and the problem is solved.
1. VPN FRR configuration problem.
2. Route planning is not proper and there is loop.
For the planning and practice of VPN FRR, avoid loop as much as possible. It should set two unequal-cost route with the cost of IGP when deploying. If there is equal-cost route, adjust it with weight of ISIS.