cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
426
Views
0
Helpful
0
Replies

3850 IGMP input high CPU

Hello,

I have a Cisco 3850 stack with 2 switches, 3.7.0E IOS, and they are experiencing high CPU peak.

Peaks are cyclic and take place one a minute during 15 seconds more or less, I have checked processes:

                    111111111111111
               6666600000000000000066666
    211111333336666600000000000000077777222222222233333111112222
100                 ***************
 90                 ***************
 80                 ***************
 70            *************************
 60            *************************
 50            *************************
 40            *************************
 30            *************************
 20            *************************
 10            *************************
   0....5....1....1....2....2....3....3....4....4....5....5....
             0    5    0    5    0    5    0    5    0    5
               CPU% per second (last 60 seconds)

PCC-COM-SWT#show processes cpu sorted | exclude 0.00
Core 0: CPU utilization for five seconds: 6%; one minute: 8%;  five minutes: 21%
Core 1: CPU utilization for five seconds: 68%; one minute: 20%;  five minutes: 15%
Core 2: CPU utilization for five seconds: 1%; one minute: 20%;  five minutes: 10%
Core 3: CPU utilization for five seconds: 1%; one minute: 6%;  five minutes: 5%
PID    Runtime(ms) Invoked  uSecs  5Sec     1Min     5Min     TTY   Process
10639  1905347     13642207 1147   17.09    8.93     10.23    0     iosd
5943   813160      35212827 51     1.89     1.47     1.53     1088  fed
5945   1943420     41293167 47     0.19     0.07     0.08     0     stack-mgr
6498   1863200     12307674 151    0.10     0.25     0.26     0     ffm
5944   1032440     58072118 17     0.05     0.07     0.06     0     platform_mgr
6501   1006550     1134373  887    0.05     0.05     0.05     0     cpumemd
10634  2831480     23240118 12     0.05     0.12     0.13     0     wcm

PCC-COM-SWT#show processes cpu detailed process iosd | exclude 0.00
Core 0: CPU utilization for five seconds: 3%; one minute: 8%; five minutes: 20%
Core 1: CPU utilization for five seconds: 99%; one minute: 32%; five minutes: 18%
Core 2: CPU utilization for five seconds: 3%; one minute: 17%; five minutes: 10%
Core 3: CPU utilization for five seconds: 3%; one minute: 7%; five minutes: 6%
PID     T C  TID     Runtime(ms) Invoked   uSecs  5Sec   1Min 5Min TTY   Process
                                                  (%)    (%)  (%)
10639   L            1910407     136422258 1147   24.8   10.2 10.4 0     iosd
10639   L 1  10639   2044974     121471752 0      24.5   9.89 10.2 0     iosd
10639   L 0  11578   3849380     10222258  0      0.34   0.27 0.23 0     iosd.fastpath
7       I            1915240     150788    0      0.66   0.33 0.33 0       Check heaps
22      I            1223640     2932338   0      0.77   0.33 0.22 0       CMI IOSd task
289     I            1334423     4678902   0      96.7   37.3 37.7 0       IGMP Input
358     I            2153310     56211     0      0.77   0.44 0.44 0       QoS stats process

I checked IGMP traffic and even though there isn´t so much IGMP traffic received, between 40 and 50 groups, each time he sends a queries the CPU goes up until all reports are processed. I have already checked but I found no issues/bugs related

Any ideas on what may be going on?

Thanks!

0 Replies 0