NE80 router and the switch of one company are connected through Gigabit Ethernet fiber. They both operate OSPF protocol. They can be pinged, but they cannot learn the route of the peer. Execute display ospf peer command at NE80 and it is found that OSPF neighbor state flaps from Exchange to Exchange Start.
Delete ACL configuration on the switch and the problem is solved.
From OSPF neighbor state, the problem lies on the exchange process of DD packet. Turn on debugging switch of OSPF Packet and OSPF Event at NE80. It is found that HELLO packet is interacted normally and Two-Way relation is established. In the subsequent exchange of DD packet, compare their Router ID. NE80 regards itself as master router and sends OSPF DD packet to the peer. But the switch ignores it. It sends initialized negotiated DD packet to NE80. So OSPF neighbor state flaps from Exchange to Exchange Start.
There are possible two reasons:
1. NE80 does not send DD packet.
2. NE80 sends DD packet, but the switch does not receive it correctly. OSPF module does not receive it.
3. Check NE80 configuration and statistics and it is normal.
4. Check the switch configuration and it is found that GE enables ACL.
5. Check rules of ACL and it is found that the configured ACL denies IP packet with destination address as GE interface address (It permits ICMP packet to pass.). DD packet sent by NE80 is filtered completely, but OSPF HELLO packet is sent with multicast and it is not influenced.