Spontaneous high Routing Engine CPU utilization

While processing updates, the Routing Engine (RE) CPU may show a utilization spike for a couple of seconds. This is normal behavior and can be safely ignored.

Sometimes the Routing Engine (RE) CPU may spike for a couple of seconds. This spontaneous RE CPU spike is generally observed during processing updates. This may include RPD updates like route updates (which cannot be seen in log messages), kernel tasks, SNMP/MIB processing, etc. While processing such updates, additional CPU resources would be utilized which can trigger a momentary CPU spike for few seconds before it returns to a normal state.

While processing the updates related to RPD, kernel, SNMP, etc., additional CPU resources may be called upon, which can trigger a momentary CPU spike. If the CPU utilization continues to remain high for long time, then this a possible concern which needs to be analyzed further. If the RE CPU continues to remains high, check which process is being excessively utilized by executing the following command:

> show system processes extensive                                   
last pid: 97074;  load averages:  0.05,  0.02,  0.00  up 118+04:04:06    15:50:39
120 processes: 3 running, 100 sleeping, 17 waiting

Mem: 267M Active, 94M Inact, 50M Wired, 375M Cache, 69M Buf, 707M Free
Swap: 2048M Total, 2048M Free


  PID USERNAME  THR PRI NICE   SIZE    RES STATE    TIME   WCPU COMMAND
   11 root        1 171   52     0K    12K RUN    2748.8 95.85% idle
 1504 root        1  96    0   112M 11952K RUN     65.9H  0.98% chassisd
   12 root        1 -20 -139     0K    12K WAIT   222:44  0.00% swi7: clock sio
 1530 root        1  96    0  6408K  2976K select 137:19  0.00% license-check
 1621 root        1  96    0 11912K  5324K select 116:45  0.00% l2ald
 1643 root        1  96    0 23328K 12788K select 114:20  0.00% mib2d
 1525 root        1  96    0 11580K  3048K select  80:10  0.00% shm-rtsdbd
 1641 root        1  96    0 18928K  7120K select  70:17  0.00% pfed
   14 root        1 -40 -159     0K    12K WAIT    49:15  0.00% swi2: netisr 0
   22 root        1 -80 -199     0K    12K WAIT    46:40  0.00% irq9: cbb1 fxp0
 1217 root        1   8    0   680K   392K nanslp  42:34  0.00% gstatd
 1501 root        1  96    0  1776K  1164K select  38:57  0.00% bslockd
 1644 root        1  96    0 19480K 16236K select  29:06  0.00% snmpd
   39 root        1 -16    0     0K    12K psleep  27:46  0.00% vmkmemdaemon
 1645 root        1  96    0  6780K  2956K select  23:16  0.00% alarmd

This is expected behavior and this sort of sudden spike for few seconds will not cause any problem during normal operations. It can be ignored.

Related