[c-nsp] 3750 12.2(44)SE1 CPU 5% weirdness

Paul paul at gtcomm.net
Sun May 11 19:31:15 EDT 2008


CPU utilization for five seconds: 6%/0%; one minute: 6%; five minutes: 5%
 PID Runtime(ms)   Invoked      uSecs   5Sec   1Min   5Min TTY Process
 362        2609      1734       1504  0.95%  0.18%  0.04%   1 Virtual Exec
   9     5221403  25084858        208  0.63%  0.27%  0.24%   0 ARP Input
 184      958738   3069313        312  0.15%  0.10%  0.06%   0 IP Input

It's always at 5%, even with no traffic at all, maybe it's just the way 
it's supposed to be with the
new software, which is fine with me I just want to know about it :>


3750 with 12.2.44SE1
PING 10.9.0.1 (10.9.0.1): 56 data bytes
64 bytes from 10.9.0.1: icmp_seq=0 ttl=255 time=1.482 ms
64 bytes from 10.9.0.1: icmp_seq=1 ttl=255 time=1.434 ms
64 bytes from 10.9.0.1: icmp_seq=2 ttl=255 time=0.660 ms
64 bytes from 10.9.0.1: icmp_seq=3 ttl=255 time=1.744 ms
64 bytes from 10.9.0.1: icmp_seq=4 ttl=255 time=2.071 ms
64 bytes from 10.9.0.1: icmp_seq=5 ttl=255 time=0.588 ms
64 bytes from 10.9.0.1: icmp_seq=6 ttl=255 time=2.019 ms
64 bytes from 10.9.0.1: icmp_seq=7 ttl=255 time=2.264 ms
64 bytes from 10.9.0.1: icmp_seq=8 ttl=255 time=2.389 ms


3750 with 12.2.25
PING 10.3.100.1 (10.3.100.1): 56 data bytes
64 bytes from 10.3.100.1: icmp_seq=0 ttl=255 time=0.923 ms
64 bytes from 10.3.100.1: icmp_seq=1 ttl=255 time=0.982 ms
64 bytes from 10.3.100.1: icmp_seq=2 ttl=255 time=0.895 ms
64 bytes from 10.3.100.1: icmp_seq=3 ttl=255 time=0.958 ms
64 bytes from 10.3.100.1: icmp_seq=4 ttl=255 time=0.961 ms
64 bytes from 10.3.100.1: icmp_seq=5 ttl=255 time=0.927 ms
64 bytes from 10.3.100.1: icmp_seq=6 ttl=255 time=3.469 ms
64 bytes from 10.3.100.1: icmp_seq=7 ttl=255 time=0.946 ms
64 bytes from 10.3.100.1: icmp_seq=8 ttl=255 time=0.936 ms
64 bytes from 10.3.100.1: icmp_seq=9 ttl=255 time=0.974 ms
64 bytes from 10.3.100.1: icmp_seq=10 ttl=255 time=0.941 ms
64 bytes from 10.3.100.1: icmp_seq=11 ttl=255 time=1.214 ms
64 bytes from 10.3.100.1: icmp_seq=12 ttl=255 time=0.902 ms

Kind of weird..

I guess the only issue to me is that the traceroute-through times looks 
weird..
You see a traceroute like this:

traceroute to cnn.com (64.236.29.120), 30 hops max, 38 byte packets
 1  router1 (x.x.x.x)  0.231 ms  0.202 ms  0.164 ms
 2  3750with12244SE1 (x.x.x.x)  2.328 ms  1.190 ms  2.047 ms
 3  edge-router  (x.x.x.x)  0.284 ms  0.282 ms  0.405 ms

looks a bit odd that way :) the previous software was usually at least 
under 1ms

I'm not complaining because the forwarding obviously works at much lower 
latency or the edge router wouldn't be so low, but customers will 
undoubtedly complain over 1.5ms because they don't understand what is 
happening.

Paul

Wink wrote:
> "show proc cpu" please?
>
> Is it really affecting the ping times?  Doesn't seem like it would 
> affect ping times at all.
>
> Paul wrote:
>> Anyone out there have 3750 running 12.2(44)SE1 ?
>> Strange issue with the CPU sitting at 5% no matter what is going on, 
>> zero traffic or lots of traffic.
>> Simple config, very few routes, 2 etherchannels, nothing major.
>>
>> Just curious.. It's not affecting anything except the ping time when 
>> you ping the switch directly.
>>
>>
>> _______________________________________________
>> cisco-nsp mailing list  cisco-nsp at puck.nether.net
>> https://puck.nether.net/mailman/listinfo/cisco-nsp
>> archive at http://puck.nether.net/pipermail/cisco-nsp/
>>
>>
>>   
>



More information about the cisco-nsp mailing list