Руководство По Устранению Ошибки для Cisco Cisco IOS XE 3.5E

provides plenty of information in order to succesfully resolve such a situation.

Use this command in order to trigger a log to be generated at a specific threshold:

The log generated with the command looks like this:

The log generated provides this information:

The total CPU utilization at the time of the trigger. This is identified by Total CPU
Utilzation(total/Intr) :50/0 in this example.

• 

Top processes − these are listed in the format of PID/CPU%. So in this example, these are:

    8622/25 − 8622 is PID for IOSd and 25 implies that this process is using 25% CPU.

    5753/12 − 5733 is PID for FED and 12 implies that this process is using 12% CPU.

• 

The EEM script is shown here:

event syslog pattern "%CPUMEM−5−RISING_THRESHOLD"

 action 0.1 syslog msg "high CPU detected"

 action 0.2 cli command "enable"

 action 0.3 cli command "show process cpu sorted | append nvram:<filename>.txt"

 action 0.4 cli command "show process cpu detailed process <process name|process ID>

 sorted | nvram:<filename>.txt"

 action 0.5 cli command "show platform punt statistics port−asic 0 cpuq −1

 direction rx | append nvram:<filename>.txt"

 action 0.6 cli command "show platform punt statistics port−asic 1 cpuq −1

 direction rx | append nvram:<filename>.txt"

 action 0.7 cli command "conf t"

 action 0.8 cli command "no event manager applet highcpu"

Note: The process cpu threshold command does not currently work in the 3.2.X train. Another point to
remember is that this command looks at the average CPU utilzation among the four cores and generates a log
when this average reaches the percentage that has been defined in the command.

• 

• 

• 

Updated: Apr 04, 2014

Document ID: 117594