cpu utalization urgent help

Unanswered Question
Jul 16th, 2009

guys we are having a big problem and i cant get my head around it.....currently we have two router sin data centre A and B A is cisco 3825 router and B is 2801 router with 64 flash and 256 ram....now both running hsrp on its lan B is a redundant router.....A is connected to mpls network cloud and B is connected via internet B is defauly gateway for internet users as well on th elan....now B has 4 BGP peers and there are four gre VPN tunnels........and couple of access-list and crypro maps as well....its running advance k9 12.4 version of IOS....A and B is running IBGP on lan there are around 163 prefixes recieving currently by B router from A.....now guys a very stange thing is happening we have to advertise a network via bgp suppose on the wan.....i have configured a static route of pointing towards the firewall on the LAN .....everything worked perfectly......now as soon as i config the same thing on router B after 30 min the router B hangs....i took off the config just from router B and saw that CPU utalization was 97% now as soon as i take off config from router A just the two routes the CPU utalization comes back to 20%....now guys what is going on there is a lot of diff between 20% and 97% just adding two routes in router A and router B the cpu utalization cant get that high according to my understanding.....i have done the bug tool test and there is nothing wrong with the IOS.....even if i take route of the B still it hangs th eonly time the router behaves properly is when i take the route from A.....guys can someone tell me what is going wrong......just two routes and its dead.........please help m eout as i am getting nuts.....plz help

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Giuseppe Larosa Thu, 07/16/2009 - 11:26

Hello Khan,

what speed are the WAN links?

there is any chance that a lot of traffic is expected / arriving for the new subnet loading the devices.

Have you verified routes consistency in the few minutes you can still access the devices (have you verified there is no routing loop with a traceroute from a remote site for example)

you can check also the FW log and FW routing table.

Hope to help


chinkevi_2 Thu, 07/16/2009 - 15:02

there are good links in Cisco online for high cpu troubleshooting. Do "sh proc cpu" and you can see which process is taking up the cpu, and if it is caused by traffic volume.

You could capture the output while problem occur and start from there.


This Discussion