High CPU usage on WS-C4506-E

[laparo]

Hi dears,

I have a monitoring tool (Cacti) that when the C4500 has high cpu usage not graphic via SNMP. I mean the graph looks like cutted.

Could you pls take me a hand on tshoot that?

Thanks in advance.

-RR

[marce1000]

  -  FYI : https://www.cisco.com/c/en/us/td/docs/switches/lan/catalyst4500/troubleshooting/cpu_util.html

   M.

[laparo]

Hi dear, could you take me a hand pls analysing these outputs in order to solve the high cpu usage?

The low values ​​that don't influence on CPU consumption aren't included below.

Thanks in advance.

# show processes cpu detailed

Core 0: CPU utilization for five seconds: 100%; one minute: 69%; five minutes: 57%
Core 1: CPU utilization for five seconds: 98%; one minute: 66%; five minutes: 61%
PID T C TID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
(%) (%) (%)
5048 L 1616028 9735191 165 47.2 6.70 12.1 0 licensed
5048 L 1 6834 586183 8700993 0 9.06 1.23 0.41 0 licensed
5048 L 1 7228 13613 1457 0 38.1 5.97 10.2 0 eicore_ipc
5057 L 154682 9522425 16 2.07 5.15 2.74 0 cli_agent
5057 L 1 21959 6385 1625 0 1.88 41.3 47.0 0 cli_agent
5090 L 1580238 261378714 232 49.3 53.5 52.8 0 iosd
5090 L 0 5090 397699 165956555 0 44.3 48.0 48.8 0 iosd
5090 L 0 7452 3290996 921502583 0 5.03 5.49 3.95 34816 iosd.fastpath
91 I 3791946 394464513 0 7.22 7.00 6.88 0 Cat4k Mgmt HiPri
92 I 537867 508317551 0 43.0 45.1 46.9 0 Cat4k Mgmt LoPri
211 I 3015863 400951063 0 30.2 31.9 33.0 0 IP Input

#show platform health

%CPU %CPU RunTimeMax Priority Average %CPU Total
Target Actual Target Actual Fg Bg 5Sec Min Hour CPU
GalChassisVp 3.00 0.09 20 105 100 500 0 0 0 133:11
S2w-JobEventSchedule 10.00 0.00 10 0 100 500 0 0 0 0:00
Stub-JobEventSchedul 10.00 0.00 10 0 100 500 0 0 0 0:00
IrmL2LispMan Review 15.00 0.00 15 0 100 500 0 0 0 0:00
K5CpuMan Review 30.00 53.64 30 65 100 500 65 60 49 7688:36
K5AclCamMan stale en 1.00 0.00 10 5 100 500 0 0 0 0:00
K5FlowStatsFifoSearc 25.00 0.00 2 0 100 500 0 0 0 0:00
-------------
%CPU Totals 486.69 71.16

[MHM Cisco World]

5090 L 0 7452 3290996 921502583 0 5.03 5.49 3.95 34816 iosd.fastpath <<-

K5CpuMan Review 30.00 53.64 30 65 100 500 65 60 49 7688:36 <<-

there are huge traffic punt to CPU 

MHM

[Flavio Miranda]

@laparo 

If you get the OID used by the Cacti and run snmpwalk manually from the Linux, do you get result from the 4500?

[laparo]

Hi dear,

Please, could you explain to me in more detail how exactly it would be done?

Just to clarify, I don't have access to the server where the Cacti is located.

Thanks in advance.

[Flavio Miranda]

On this case I dont think the idea I gave will help you.  It should be checked on the Cacti server to make sure if can pull the SNMP information from the switch. If it does, the problem is not on the switch. 

[laparo]

Hi dear,

I've checked it (SNMP walk) with Cacti's admis and works fine!

Thanks.

[Flavio Miranda]

Great. This means the switch is providing the information Cacti needs. Your problem must be related to Browser, Java, javascript or something related.

[Leo Laohoo]

Post the complete output to the following commands: 

sh version
sh module
sh process cpu sorted

[laparo]

Hi dear,

Attached are the requested show commands.

Any idea on this would be welcome.

Thanks in advance.

[Leo Laohoo]

High CPU (both CPU) are due to IOSd process. 

Reboot the VSS or upgrade to, say 03.11.11E.

[laparo]

Hi dear,

Thanks for the quick answer.

I've rebooted the VSS some time before.

However following your advice I'll upgrade soon.

Best.

[laparo]

Dear all,

In the meantime, what can I do to mitigate the high CPU usage?

I ask the above because to perform the update I need authorization and it will take some time.

I have seen some CPU control policies (COPP) but honestly I have never done it and I don't know how it behaves.

Thanks in advance.