[c-nsp] cisco 3650 & high packet losts

[Sergey Velikanov [Intelsoft]]

Hello

Yesterday evening I've run multicast in my network (30 channels ~ 100Mbit),
and I've discovered that performance of my cisco 3560 decreased

#sh ver
Cisco IOS Software, C3560 Software (C3560-ADVIPSERVICESK9-M),
cisco WS-C3560-48TS (PowerPC405) processor (revision C0) with 118784K/12280K bytes of memory.

my config is

interface Loopback0
  ip address 192.168.9.1 255.255.255.255
  ip pim sparse-mode
end

(clients ports from fa0/1 to fa0/30)
interface FastEthernet0/1
  description vlan 101
  switchport access vlan 101
  switchport mode access
   speed 10
  storm-control broadcast level pps 10k
  storm-control multicast level pps 20k
  storm-control unicast level pps 20k
  no cdp enable
end

interface Vlan101
  description customer1
  ip address 192.168.46.9 255.255.255.248
  ip access-group ipprotocols in
  ip pim sparse-mode
end

[skip fa0/2 .. fa0/30 & vlan102...vlan130]

# Multicast Interface
interface GigabitEthernet0/3
  description vlan 151
  switchport access vlan 151
  switchport mode access
end

interface Vlan151
  description digitaltv
  ip address 10.0.0.6 255.255.255.252
  ip access-group ipprotocols in
  ip pim sparse-mode
end

ip pim rp-address 192.168.9.1

mulricast works fine but

#ping 218.45.143.10 repeat 100  (customer 3)
   Type escape sequence to abort.
    Sending 100, 100-byte ICMP Echos to 218.45.143.10, timeout is 2 seconds:
    !!.!!!!!!!..!!!!!.!!!!!..!!..!.!!.!!!.!!!.!!!!!!!!.!!!!!!!!...!!!!!!!!.!!.!!!!!!!!!!!..!!.!!!.!..!!.
    Success rate is 75 percent (75/100), round-trip min/avg/max = 1/2/9 ms

#ping 192.168.29.195 repeat 100
    Type escape sequence to abort.
     Sending 100, 100-byte ICMP Echos to 192.168.29.195, timeout is 2 seconds:
      !!!!.!!.!!!.!!!!.!.!!!!!.!!!!!!!!!!!!!!!.!!..!.!!!!..!!!!!.!!!!!.!!!!!
      !!!!!!.!!!!!!!!.!!.!!.!.!!!.!.
      Success rate is 79 percent (79/100), round-trip min/avg/max = 1/8/67 ms


ok

#conf t
(config)#int vlan103
(config-if)#no ip pim sparse-mode

(config)#int gi0/3
(config-if)#shut
(config-if)#no shut


#ping 192.168.29.195 repeat 100
Type escape sequence to abort.
Sending 100, 100-byte ICMP Echos to 192.168.29.195, timeout is 2 seconds:
!!!!!.!!.!!!!!!.!....!.!.!.!!!!!!.!!.!!!!!!!.!!!!!!!!!!!!.!.!!.!!.!!!!
!.!!..!!!.!!!!!.!!!!!..!.!!..!
Success rate is 73 percent (73/100), round-trip min/avg/max = 1/2/9 ms

#ping 218.45.143.10 repeat 150
Type escape sequence to abort.
Sending 150, 100-byte ICMP Echos to 218.45.143.10, timeout is 2 seconds:
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!
Success rate is 100 percent (150/150), round-trip min/avg/max = 1/1/9 ms

I've try it five times, as soon as I disable pim on customers interface ping success rate become 100%

#show processes cpu history

                          3333311111
      5777775555555555555555555666665555544444555556666666666555
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)


      333141     4 4  33      1111113111111111111111111111111
      5004406666945956777666790211111222231223221211222222321666
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 minute (last 60 minutes)
               * = maximum CPU%   # = average CPU%

#sh processes memory
Total: 102882168, Used: 38380560, Free: 64501608


where is bottleneck in my cisco?