High CPU use in 3550

Unanswered Question
Oct 6th, 2008
User Badges:

Hi All,


I'm facing the following problem and before I contact Cisco TAC, I want to drop a line here.


I've got a C3550-IPBASE-M running IOS Version 12.2(25)SEE2. It's got two L3 VLAN interfaces. All Interfaces are placed in on. I've got one static route and 1 ACL's with 5 rule.


Basically, the setup is a border router connected to our internetlink. Traffic is about 40Mbps sustained, 100Mbps peak.


I've got a second 3750 connected to the same ISP. It's doing the same. Both 3750 use HSRP for redundancy.


Now, CPU usage shoots up to 90-99%.


I'm not using IPv6


No Logging is enabled


No individual process is using more than 1% of processor.


I am using the no ip access list command. But CPU usage is same.


System mtu is 1500


No IPSEC/GRE tunnel interface.


Using ip cef.


Anyone please help about that matter.


  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
RicheeJJJ_2 Mon, 10/06/2008 - 04:40
User Badges:

do show proc cpu sorted


This will tell you which process are using the most memory starting with the highest first. Post which process is using up all this cpu and we can help you more.

RicheeJJJ_2 Mon, 10/06/2008 - 06:55
User Badges:

Do this:

sh proc cpu | exclude 0.00%__0.00%__0.00%


Use a calculator, add them all up, and see if it's 99%. If it's not 99% then yes it's a TAC case. Because your cisco router can't do basic math and who knows what else it isn't doing right.

shahjahansiraj Mon, 10/06/2008 - 07:23
User Badges:

VOIP(config)#do sh process cpu | exclude 0.00

CPU utilization for five seconds: 95%/94%; one minute: 96%; five minutes: 96%

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

29 126056688 201933850 624 0.47% 0.75% 0.92% 0 Vegas LED Proces

37 289839368 26181425 11070 0.47% 0.47% 0.49% 0 Vegas Statistics

38 13868684 118710584 116 0.15% 0.08% 0.09% 0 HMATM Learn proc

genusys(config)#

nate-miller Wed, 10/08/2008 - 06:04
User Badges:
  • Bronze, 100 points or more

the 95%/94% indicates that 95% of your processor is being used, and 94% of it is interrupt driven, rather than process driven. You won't see these interrupts in the process list.



If you've got it hooked up to an ISP, how many routes are you getting on it?


Look at the command "sh sdm prefer"

This shows how the TCAM on a 3750 has been allocated. By default, a 3750 can only store 2000 IPV4 routes. It uses enough TCAM to store 6000 MAC addresses.


By using the config command "sdm prefer routing", you can juggle the TCAM around so it can handle more routes.


sh ip route summary


This will tell you how many subnets you're actually learning. If these numbers don't match up with the SDM numbers, I would apply the new SDM profile and reboot- and see what happens.


If this isnt' the case, there's more you can look at- but this is what caused my network to choke the first time I saw this.






shahjahansiraj Wed, 10/08/2008 - 07:49
User Badges:

This is the result of sh ip route summary command



VOIP#sh ip route summary

IP routing table name is Default-IP-Routing-Table(0)

IP routing table maximum-paths is 32

Route Source Networks Subnets Overhead Memory (bytes)

connected 0 4 256 608

static 1 0 64 152

internal 3 3516

Total 4 4 320 4276

genusys#

netwalkr1 Wed, 10/08/2008 - 06:27
User Badges:

I saw a similar scenario once before. The customer had a Metro Ethernet connection from the SP. The default route next hop was set the the FE interface, not the next hop address. The ARP table would populate with all the destinations being on the Internet and destroy the router (1721). This *may* be due to proxy-arp being enabled by default on the older code running on the platform. I changed the default to the next hop IP and it resolved the problem.

Actions

This Discussion