1 ) Check cpu utilization to figure out whether it is due to process or interrupt traffic.
show proc cpu | e 0.00 or show proc cpu sorted
2) If it is due to io-base then use the below command to check which process/component like "k5cpumain" review etc.. is causing high cpu.
show proc cpu details pid <> sorted
3) check which component is utilizing more resources than expected for example in TARGET and ACTUAL CPU column.
show platform health
4) check which CPU queue status/packet drops in the queue and find out in which queue packet is punting to CPU in the average column 5 sec /5 minut/1hours/. CPU has 64 queues.
show platform cpu packet statistics -------> Looks for the kind of queue and packet punting to the CPU like "Adj same If " , "SA miss ", "K5CPU main review."
Repeat the command multiple times and check counters if it is incrementing.
5) There are two option to see which packets are going to CPU in detail.
Below are the different tools to use to capture the packet
Tool 1 : Use "debug platform cpu packet receive buffer " and show platform cpu packets buffer. And look for the same "Event id " as same queue where packets are punting to the CPU. For example , if packets are punting to CPU in Adj same If queue as per show platform cpu packets statistic then look for the same queue under EVENT ID in show platform cpu packets buffer.
Tool 2 : Second option is to span the packets punting to the CPU queue by using external packet capture. Use the below comment to configure the SPAN.
Switch(config)#monitor session 1 source cpu queue all rx
Switch(config)#monitor session 1 destination interface <destination interface number>
Switch(config)#end
Tool 3 : Identify the Interface That Sends Traffic to the CPU—Cisco IOS Software Release 12.2(20)EW and Later
Reason for high CPU : SA miss due to host flapping between interfaces.
No comments:
Post a Comment