NE80 creates ROUTE-POLICY and then it does not respond. About 45 seconds later it recovers. During the period, the forwarding of the service is normal. This new ROUTE-POLICY is not imparted by any route protocol. When changing created and non-imparted ROUTE-POLICY, the same problem occurs. Operate BGP at NE80 and receive INTERNET route. The list of route table is above 10,000. Version: VRP 3.10-2226。
%Aug 3 05:44:48 2005 NE80-A-FZ1 SHELL/6/CLI:'route-policy apply_Telecom_main permit node 10' from 18.104.22.168 by cncfznoc %Aug 3 05:45:05 2005 NE80-A-FZ1 SHELL/6/CLI:'if-match as-path 134' from 22.214.171.124 by cncfznoc %Aug 3 05:45:53 2005 NE80-A-FZ1 SHELL/6/CLI:'DIS cpu' from 126.96.36.199 by cncfznoc cpu utilization:[NE80-A-FZ1-route-policy]DIS cpu CPU Performance: 0 seconds ago ||=================== || 99%
1. Configure ROUTE-POLICY. After the device recovers check CPU utilization and it reaches 99%, ROUT process reaches 99%.
2. Check route table and the list reaches above 10,000. Route list is much and refresh results in slow response.
NE80 will refresh route table when it realizes ROUTE-POLICY. Even there is no route protocol imparted, it will refresh and re-compute (route process reaches 99%). If there are many routes, SRU re-computes and results in high CPU utilization when configuring policy. The LPU forwarding table does not join the computation and occupy little resource. So it does not influence forwarding performance and forwarding service.
Configure ROUTE-POLICY for many route lists, avoid service rush.