EEM Script for CPU

[ssieger]

Hello,

from time to time we are experiencing high cpu-load. I try to find out who is responsible for that load and i've found that example of an eem-script:
event manager applet high_cpu_load
 event snmp oid 1.3.6.1.4.1.9.9.109.1.1.1.1.3.1 get-type exact entry-op ge entry-val 95 poll-interval 1
 action 0.0 syslog msg "High CPU DETECTED. Logging Information to disk0:high_cpu.txt"
 action 0.1 cli command "enable"
 action 0.2 cli command "show clock | append disk0:high_cpu.txt"
 action 1.2 cli command "term length 0"
 action 1.3 cli command "show process cpu sorted | append disk0:high_cpu.txt"
 action 1.4 cli command "term length 50"

So far so good, but im not well versed in eem scripting. Can anybody help to modify this script so that it is able to filter out the 0.00% lines from the show process cpu sorted command? The command "show process cpu sorted | e 0.00& | append disk0:high_cpu.txt" doesn't work. Further it would be nice to have a single txt-file for every event like high_cpu_1.txt, high_cpu_2.txt an so on, instead of one big file.

Thanks in advance,

Stephan

[Joe Clarke]

This is an age-old issue with EEM.  Only very recent versions of IOS can do direct file writes to enable this in applets.  Else you'll have to switch to Tcl.  What version of IOS do you have?

[ssieger]

IP-Services 151-2.SY4

[Joe Clarke]

I believe that is EEM 3.0, so you will have to switch to Tcl to do what you want.  Change your CLI command to "show proc cpu sort | exc 0.00", then take this applet to http://www.marcuscom.com/convert_applet/ .  In the resulting Tcl policy, add the following after you run each command:

append output $_cli_result

Then at the end of the script:

set fd [open "disk0:high_cpu.txt" a]

puts $fd $output

close $fd

[ssieger]

Hi Joseph,

that ist the output i get:

namespace import ::cisco::eem::*
namespace import ::cisco::lib::*

array set arr_einfo [event_reqinfo]

if [catch {cli_open} result] {
    error $result $errorInfo
} else {
    array set cli1 $result
}

action_syslog msg "High CPU DETECTED. Logging Information to disk0:high_cpu.txt"
if [catch {cli_exec $cli1(fd) "enable"} _cli_result] {
    error $_cli_result $errorInfo
}

if [catch {cli_exec $cli1(fd) "show clock"} _cli_result] {
    error $_cli_result $errorInfo
}

if [catch {cli_exec $cli1(fd) "term length 0"} _cli_result] {
    error $_cli_result $errorInfo
}

if [catch {cli_exec $cli1(fd) "sh processes cpu sorted | exclude 0.00"} _cli_result] {
    error $_cli_result $errorInfo
}

if [catch {cli_exec $cli1(fd) "term length 50"} _cli_result] {
    error $_cli_result $errorInfo
}

# Close open cli before exit.
catch {cli_close $cli1(fd) $cli1(tty_id)} result

As aforementioned i have no experiences with tcl scripts, so where exactely have i to put in the additional lines?