Hello everyone,
Today, I'm going to introduce the discussion about OSPF convergence in the new HCIE lab.
Topic
1. R1 advertises a route 10.1.1.1. A virtual link is created between R2 and R4, and the interface cost of R2 and R4 is changed to 1000. What problem will happen?
2. If R2 is a Huawei device and R4 is a Cisco device, what problems will occur in the interconnection?
Question 1
R1 advertises a route 10.1.1.1. A virtual link is created between R2 and R4, and the interface cost of R2 and R4 is changed to 1000. What problem will happen?
Answer
When accessing 10.1.1.1, a loop occurs between AR4, AR5, and AR6.
Establishing a Vlink between AR2 and AR4 can solve the problem of inconsecutive backbone areas. After the vlink is configured, 10.1.1.1 can be flooded in area 0 through the virtual link. As a result, R5 and R6 preferentially use type 1 LSAs destined for 10.1.1.1 in area 0. AR5 has two LSAs destined for 10.1.1.1. One is a Type 3 LSA in the non-backbone area sent from AR3 and the other is a Type 1 LSA in the backbone area sent from AR6. AR5 preferentially selects the Type 1 LSA in the backbone area. Access 10.1.1.1 through AR6. The LSA of AR6 is sent from AR4. AR6 accesses 10.1.1.1 through AR4. In this case, the LSDB of R4 has three LSAs destined for 10.1.1.1. One LSA is transmitted through the vlink and belongs to area 0, and the other is type 3 LSA generated by AR2 in area 1. One is the type-3 LSA generated by AR3 in area 1. According to the cost calculation, the cost of the Type 3 LSA generated by AR3 in area 1 is the smallest. In this scenario, AR4 does not preferentially use the Type 1 LSA in area 0 for route calculation but prefers the path with the lower cost. As a result, R4 accesses 10.1.1.1 through R5, causing a loop.
The reason is that the ABR (AR4) with a vlink preferentially selects the inter-area route with a lower cost. The route selection mechanism that is type 1 LSAs are better than type 3 LSAs is not always used. You can create another vlink on AR3 and AR5 to solve the loop problem.
Question 2
If R2 is a Huawei device and R4 is a Cisco device, what problems will occur in the interconnection?
Answer
1. Cisco v-links do not periodically send Hello packets after neighbor relationships are established. Huawei vlinks periodically send Hello packets to maintain neighbor relationships. Huawei vlinks may consider the peer neighbor to be invalid, causing the virtual link to go Down.
2. By default, Huawei devices do not compare interface MTUs, but Cisco devices compare interface MTUs by default. If the MTUs are not adjusted and the MIUs at both ends are inconsistent, therefore, when OSPF is implemented on a Huawei device and a Cisco device, OSPF is stuck in the exstart state and cannot enter the full state. It is recommended that the MTU check function be enabled on Huawei devices and the MTU settings of the interconnected interfaces on the two devices are the same.
That is all I want to share with you!
