[c-nsp] Varying CPU usage on 881
Pelle
perc69 at gmail.com
Mon Sep 29 10:20:04 EDT 2008
Hi Rodeney.
> That graph is total usage so there may be another process that
> contributes that takes longer on the 881 that doesn't show up
> on the 1841. ie: checkheaps, etc...
>
> Watch 'sh proc cpu sorted' during one of those spikes and see
> if you see another process contributing.
That was my original theory as well, but there are no other
significant processes:
881_router#sh proc cpu | ex 0.00
CPU utilization for five seconds: 23%/18%; one minute: 14%; five minutes: 5%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
51 1104 514 2147 2.21% 0.69% 0.23% 6 Virtual Exec
60 4208 632 6658 3.19% 2.67% 1.06% 0 COLLECT STAT COU
881_router#sh proc cpu | ex 0.00
CPU utilization for five seconds: 23%/18%; one minute: 14%; five minutes: 5%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
51 1136 519 2188 2.21% 0.69% 0.23% 6 Virtual Exec
60 4260 639 6666 3.19% 2.67% 1.06% 0 COLLECT STAT COU
881_router#sh proc cpu | ex 0.00
CPU utilization for five seconds: 43%/40%; one minute: 16%; five minutes: 6%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
51 1168 524 2229 1.31% 0.74% 0.25% 6 Virtual Exec
60 4292 644 6664 1.88% 2.61% 1.07% 0 COLLECT STAT COU
881_router#sh proc cpu | ex 0.00
CPU utilization for five seconds: 43%/40%; one minute: 16%; five minutes: 6%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
51 1200 529 2268 1.31% 0.74% 0.25% 6 Virtual Exec
60 4332 650 6664 1.88% 2.61% 1.07% 0 COLLECT STAT COU
881_router#sh proc cpu | ex 0.00
CPU utilization for five seconds: 43%/40%; one minute: 16%; five minutes: 6%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
51 1228 534 2299 1.31% 0.74% 0.25% 6 Virtual Exec
60 4372 656 6664 1.88% 2.61% 1.07% 0 COLLECT STAT COU
881_router#sh proc cpu | ex 0.00
CPU utilization for five seconds: 43%/40%; one minute: 16%; five minutes: 6%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
51 1260 539 2337 1.31% 0.74% 0.25% 6 Virtual Exec
60 4404 661 6662 1.88% 2.61% 1.07% 0 COLLECT STAT COU
881_router#sh proc cpu | ex 0.00
CPU utilization for five seconds: 27%/23%; one minute: 17%; five minutes: 6%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
51 1292 544 2375 1.63% 0.82% 0.27% 6 Virtual Exec
60 4484 673 6662 2.78% 2.62% 1.10% 0 COLLECT STAT COU
881_router#sh proc cpu | ex 0.00
CPU utilization for five seconds: 27%/23%; one minute: 17%; five minutes: 6%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
51 1324 549 2411 1.63% 0.82% 0.27% 6 Virtual Exec
60 4532 680 6664 2.78% 2.62% 1.10% 0 COLLECT STAT COU
881_router#sh proc cpu | ex 0.00
CPU utilization for five seconds: 27%/23%; one minute: 17%; five minutes: 6%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
51 1360 554 2454 1.63% 0.82% 0.27% 6 Virtual Exec
60 4572 686 6664 2.78% 2.62% 1.10% 0 COLLECT STAT COU
881_router#sh proc cpu | ex 0.00
CPU utilization for five seconds: 21%/17%; one minute: 18%; five minutes: 6%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
51 1396 559 2497 1.72% 0.89% 0.30% 6 Virtual Exec
60 4612 692 6664 2.62% 2.62% 1.12% 0 COLLECT STAT COU
(The captures are taken quite rapidly.)
It's the sudden peak from 18% to 40% and back down to 17% for the
_interrupt based traffic_ that puzzles me.
> Also, I don't think it applies here given the pps number you gave
> before.
>
> I usually see that when the rate is divisible by 4.
The same wavy CPU load occurs whatever traffic I throw at the router.
--
Pelle
More information about the cisco-nsp
mailing list