S8016 is connected with NE40. Ping packets and packet loss rate is 10%. The packet loss changes with the service.
1. Check the status of the GE interface of S8016 and NE40. The status of the interface is forced 1000M and full. There is no error packet.
2. Check the log and alarm. They are normal and no packet loss occurs on the system bucket. CPU usage is normal.
3. Perform eacl statistics traffic testing. Configure eacl on S8016 and NE40 and enable debbug ip icmp.
A. Pingg �c 20 (lost 4%) on S8016. NE40 receives 20 icmp request packets, while S8016 receives19 icmp echo packets. Based on the debugging information, NE40 receives 20 icmp request packets and sends 19icmp echo packets. One packet is lost when NE40 sends icmp packets to CPU.
B. Ping �c 20 (lost 3%) on NE40. S8016 receives 20 icmp request packets and NE40 also receives 20 icmp echo packets. Based on the debugging information, S8016 receives 20 icmp request packets and sends 20 icmp echo packets. S8016 is normal.
C. Observe port traffic. The traffic on 4*GE reaches 2G at peak hour.
D. Check the status of NP register. Packet loss occurs because the traffic is too heavy. In theory, the traffic can reach 2.5G. With regard to the cost and transmission loss, the traffic cannot reach 2.5G.
1. The interface does not match the interface of the transmission device. CRC and error packets occur.
2. S8016 and NE40 suffer from network attack. As a result, the CPU is too high and packet loss occurs in system bucket.
3. The traffic on GE interface is too heavy and reaches the limit.