This article describes the high CPU behavior that is expected when any route-related commands are executed with a large number of routes installed in the routing table.
Routing Engine (RE) CPU spikes are observed with a large number of routes in the routing table when any route-related commands with the " | no-more " option are executed, which can cause a delay in the Command Line Interface (CLI). From the system processes extensive output, the following processes may show high CPU utilization depending on how one is accessing the device.
| no-more
system processes extensive
mgd, clid, telnetd, sshd lab@R1> show system processes extensive last pid: 14359; load averages: 0.43, 0.63, 0.55 up 10+04:18:39 13:03:16 154 processes: 4 running, 136 sleeping, 14 waiting Mem: 707M Active, 307M Inact, 140M Wired, 650M Cache, 201M Buf, 29G Free Swap: 8192M Total, 8192M Free PID USERNAME THR PRI NICE SIZE RES STATE TIME WCPU COMMAND 10 root 1 171 52 0K 16K RUN 165.0H 41.99% idle 13856 labroot 1 103 0 26488K 20840K RUN 0:05 16.31% cli 19 root 1 -68 -187 0K 16K RUN 22.1H 11.82% irq11: em0 em1 em2* 13857 root 1 4 0 40792K 32760K sbwait 0:03 8.45% mgd 13851 root 1 99 0 2028K 1532K select 0:03 7.91% telnetd 1790 root 1 4 0 433M 348M kqread 40.5H 3.08% rpd lab@R1> show system processes extensive last pid: 14371; load averages: 1.33, 0.89, 0.66 up 10+04:20:19 13:04:56 154 processes: 4 running, 136 sleeping, 14 waiting Mem: 707M Active, 307M Inact, 140M Wired, 650M Cache, 201M Buf, 29G Free Swap: 8192M Total, 8192M Free PID USERNAME THR PRI NICE SIZE RES STATE TIME WCPU COMMAND 13856 labroot 1 111 0 26488K 20840K RUN 0:39 31.69% cli 13851 root 1 106 0 2028K 1532K select 0:23 18.60% telnetd 13857 root 1 4 0 40792K 32728K sbwait 0:22 18.55% mgd 19 root 1 -68 -187 0K 16K RUN 22.1H 11.67% irq11: em0 em1 em2* 1790 root 1 4 0 433M 348M kqread 40.5H 6.25% rpd
It is also observed that the CPU idle usage is zero percent.
lab@R1> show chassis routing-engine | no-more Routing Engine status: Slot 0: Current state Master Election priority Master (default) Temperature 47 degrees C / 116 degrees F CPU temperature 44 degrees C / 111 degrees F DRAM 32733 MB (32768 MB installed) Memory utilization 7 percent CPU utilization: User 42 percent Background 0 percent Kernel 35 percent Interrupt 23 percent Idle 0 percent Model RE-S-1800x4 Serial ID 9013126320 Start time 2017-05-20 08:45:06 IST Uptime 10 days, 4 hours, 20 minutes, 17 seconds Last reboot reason Router rebooted after a normal shutdown. Load averages: 1 minute 5 minute 15 minute 1.60 1.01 0.71 lab@R1> show chassis routing-engine | no-more Routing Engine status: Slot 0: Current state Master Election priority Master (default) Temperature 47 degrees C / 116 degrees F CPU temperature 44 degrees C / 111 degrees F DRAM 32733 MB (32768 MB installed) Memory utilization 7 percent CPU utilization: User 43 percent Background 0 percent Kernel 39 percent Interrupt 18 percent Idle 0 percent Model RE-S-1800x4 Serial ID 9013126320 Start time 2017-05-20 08:45:06 IST Uptime 10 days, 4 hours, 20 minutes, 22 seconds Last reboot reason Router rebooted after a normal shutdown. Load averages: 1 minute 5 minute 15 minute 1.63 1.03 0.72
The router in this example has about 500,000 active routes in the routing table as below:
lab@R1> show route summary Autonomous system number: 100 Router ID: 100.100.100.100 inet.0: 500019 destinations, 500019 routes (500019 active, 0 holddown, 0 hidden) Direct: 6 routes, 6 active Local: 6 routes, 6 active OSPF: 1 routes, 1 active BGP: 500000 routes, 500000 active Static: 3 routes, 3 active IGMP: 1 routes, 1 active PIM: 2 routes, 2 active inet.1: 3 destinations, 3 routes (3 active, 0 holddown, 0 hidden) Multicast: 3 routes, 3 active inet6.0: 2 destinations, 2 routes (2 active, 0 holddown, 0 hidden) PIM: 2 routes, 2 active inet6.1: 1 destinations, 1 routes (1 active, 0 holddown, 0 hidden) Multicast: 1 routes, 1 active
The presence of 500,000 active routes in the routing table will cause MGD to fetch data from RPD and display the routes in CLI, causing the respective CPUs to spike for the time period the route output is displayed.
Once the routes are displayed and the command stops, CPUs return to normal as seen below:
lab@R1> show chassis routing-engine Routing Engine status: Slot 0: Current state Master Election priority Master (default) Temperature 46 degrees C / 114 degrees F CPU temperature 44 degrees C / 111 degrees F DRAM 32733 MB (32768 MB installed) Memory utilization 7 percent CPU utilization: User 0 percent Background 0 percent Kernel 2 percent Interrupt 15 percent Idle 83 percent Model RE-S-1800x4 Serial ID 9013126320 Start time 2017-05-20 08:45:06 IST Uptime 10 days, 4 hours, 25 minutes, 43 seconds Last reboot reason Router rebooted after a normal shutdown. Load averages: 1 minute 5 minute 15 minute 0.05 0.61 0.65 lab@R1> show system processes extensive last pid: 14416; load averages: 0.04, 0.58, 0.64 up 10+04:26:31 13:11:08 154 processes: 3 running, 137 sleeping, 14 waiting Mem: 707M Active, 307M Inact, 140M Wired, 650M Cache, 201M Buf, 29G Free Swap: 8192M Total, 8192M Free PID USERNAME THR PRI NICE SIZE RES STATE TIME WCPU COMMAND 10 root 1 171 52 0K 16K RUN 165.1H 81.10% idle 19 root 1 -68 -187 0K 16K RUN 22.1H 10.55% irq11: em0 em1 em2* 11 root 1 -40 -159 0K 16K WAIT 571:40 2.69% swi2: netisr 0 1680 root 2 8 -88 120M 16856K nanslp 296:01 0.93% chassisd 1790 root 1 4 0 433M 348M kqread 40.5H 0.00% rpd 14 root 1 -16 0 0K 16K - 53:35 0.00% yarrow
This behavior is expected when a large number of routes exist in the routing table and the no-more option is used to display the routes. The CPU is expected to go high when any route-related outputs as follows are executed with the no-more option:
no-more
show route show route extensive/detail show route active-paths
The recommendation is to avoid usage of the no-more option wherever applicable to ensure that the CPU is stable. However, there is no user impact due to this CPU utilization since routers will assign their resources well when they have higher priority process.
2020-08-11: Article reviewed for accuracy; no changes required; article valid and relevant2022-08-16: Article reviewed for accuracy; no changes required; article valid and relevant.