Core Switch having high CPU Load

[Ejaz Ahmed]

Hi Experts,

We have two core switches in our network (Cisco WS-4503). HSRP is working as Active-Standby mode. Recently it has been found that core 1 (which is in active mode) shown high CPU load (60% - 100%). Also found that the below mentioned process takes more CPU. Our network consist of around 30 Vlans.

We have already checked for the solution in Cisco website( http://www.cisco.com/en/US/products/hw/switches/ps663/products_tech_note09186a00804cef15.shtml)  but couldn't find exact reason. Please see the below outputs for core 1 switch.

Core1# sh processes cpu sorted

CPU utilization for five seconds: 55%/3%; one minute: 65%; five minutes: 71%
PID Runtime(ms)   Invoked      uSecs   5Sec   1Min   5Min TTY Process
  30   1606933521086983644        147 24.79% 29.93% 35.85%   0 Cat4k Mgmt LoPri
  29  25669008801834660174       1399 13.83% 12.52%  9.03%   0 Cat4k Mgmt HiPri
  55  28000131363911735150          0 10.00% 16.25% 19.94%   0 IP Input
  49   365077356 759894567        480  1.11%  1.05%  1.04%   0 Spanning Tree

Core1# sh platform cpu packet statistics

Packets Received by Packet Queue

Queue                  Total           5 sec avg 1 min avg 5 min avg 1 hour avg
---------------------- --------------- --------- --------- --------- ----------
Control                     1434456125        65        55        45         34
Host Learning                  1021913         0         0         0          0
L3 Fwd High                        149         0         0         0          0
L3 Fwd Medium                        3         0         0         0          0
L3 Fwd Low                 17485670284       635       675       508        700
L2 Fwd Medium                      255         0         0         0          0
L2 Fwd Low                   393455266        18        11         8          4
L3 Rx High                         494         0         0         0          0
L3 Rx Low                     46447447         2         0         0          0
RPF Failure                        134         0         0         0          0
ACL fwd(snooping)            383954350        16        10         7          0

Please provide your valuable suggestion on this issue.

Regards and Thanks

Ejaz

[bbordereau]

hello did you follow all the command provided in the cisco doc ? especially the show platform health

one action you can try is to use the following debug

Switch#debug platform packet all receive buffer

this debug is not cpu intensive it uses an asic so it is done i hw

i used several times for troubleshoot high cpu on c4k and never see more high cpu usage with it

this debug can help you to isolate if there is a same mac addr or ip addr coming often and "disturbing the cpu"

i do not know if it is possible for you but do you have the issue with one core (the other one is isolated from the network) to see if the problem follow the hsrp status

when you swap to the slave that becomes master : issue or not ?

if it the case the root cause is perhaps elsewhere on the network

one action also is to try to unplug several port step by step to see if when you unplug a port the high cpu drops down immediately

Bertrand

Sent from Cisco Technical Support iPad App

[mahmoodmkl]

Hi,

Can you verify whether fast-switching is enabled under your SVI's.

Thanks

[David Santos]

Have you used a SPAN session to monitor your traffic?

Is proxy-arp enabled?