Unanswered Question
Aug 6th, 2008

Dear all

I am experiencing high CPU on 7609 SUP720.

The process causing issue is not even showing on below output.

Any thoughts ?



XXXXX#sh processes cpu sort | exc 0.00

CPU utilization for five seconds: 98%/87%; one minute: 54%; five minutes: 25%

PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process

122 15347952 217411916 70 11.11% 5.94% 2.21% 0 IP Input

178 5656012 140664319 40 0.15% 0.04% 0.04% 0 MLSM Process

171 6116044 141139334 43 0.07% 0.08% 0.10% 0 PIM Process

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4 (1 ratings)
Joseph W. Doherty Thu, 08/07/2008 - 04:11

The interrupt % is what appears to be consuming most of your CPU, the 87% for the last five seconds. Interrupt CPU isn't shown as a separate process. The delta between your total CPU, 98%, and the interrupt CPU, 87%, which is 11% appears to be mainly within the IP Input process.

You might want to review http://www.cisco.com/en/US/products/hw/routers/ps359/products_tech_note09186a00801c2af0.shtml and http://www.cisco.com/en/US/products/hw/switches/ps708/products_tech_note09186a00804916e0.shtml#utilities. In the latter, review the information about packets "punted to the cpu".

cisco_lad2004 Fri, 08/08/2008 - 00:48

Thanks for the pointers !

this was ian interesting one to nail and I would like to share teh steps:

1-I identified that Interface X was experiencing high number of throatles...so I have to lookup at traffic load. traffic was about 50% on a 10GE.

2- I then used debug "netdr capture interface X" when teh cpu hit the roof (this command does not impact CPU).

3- I spotted an obvious huge ammont of multicast traffic switches via the interface. This traffic was new multicast range we were not aware of...so there was no RP mapping present to handle it correctly.

RP deployed and CPU went down.....now who has an axe, am making my way to the server guys !!



This Discussion