One of our cisco 7206VXR (NPE-G1) processor with ATM interface for DSL clients as route-bridged type is currently experiencing a High CPU utilization with 99%. From the cpu processes, we noticed a high utilization of ARP Input. Do you have any recommendations to fix the problem?
CPU utilization for five seconds: 99%/78%; one minute: 99%; five minutes: 98%
PID Runtime(ms) Invoked uSecs
5Sec 1Min 5Min TTY Process
20 1137746364 460512309 2470 12.89% 12.78% 12.81% 0 ARP Input
28 4306380 1743508 2469 0.07% 0.04% 0.04% 0 Net Background
50 322862452 722734161 446 4.19% 2.69% 2.81% 0 IP Input
52 114938036 553322027 207 1.02% 0.65% 0.90% 0 IP SNMP
116 28560224 284155006 100 0.31% 0.21% 0.24% 0 PDU DISPATCHER
127 152328304 286558936 531 1.34% 0.95% 1.27% 0 SNMP ENGINE
147 3449676 117315338 29 0.23% 0.04% 0.05% 0 PPP manager
{ 4 comments… read them below or add one }
Check the link out below
http://www.cisco.com/warp/public/63/highcpu_processes.html#arp
It says
ARP Input
High CPU utilization in the Address Resolution Protocol (ARP) Input process occurs if the router has to originate an excessive number of ARP requests. ARP requests for the same IP address are rate-limited to one request every two seconds, so an excessive number of ARP requests would have to originate for different IP addresses. This can happen if an IP route has been configured pointing to a broadcast interface. A most obvious example is a default route such as:
ip route 0.0.0.0 0.0.0.0 Fastethernet0/0In this case, the router generates an ARP request for each IP address that is not reachable through more specific routes, which practically means that the router generates an ARP request for almost every address on the Internet. For more information about configuring next hop address for static routing, see Specifying a Next Hop IP Address for Static Routes.
Alternatively, an excessive amount of ARP requests can be caused by a malicious traffic stream which scans through locally attached subnets. An indication of such a stream would be the presence of a very high number of incomplete ARP entries in the ARP table. Since incoming IP packets that would trigger ARP requests would have to be processed, troubleshooting this problem would essentially be the same as troubleshooting high CPU utilization in the IP Input process.
In addition to the other post would like to suggest to check out whether ip proxy-arp is enabled on any of the interfaces.
if yes try to disable the same on them and check out …
it also proves to be a possible cause for the same..
As per my understanding, proxy arp is enabled on all routers, all interfaces, by default.
Can someone explain what would happen if it is turned off ?
It is still the default in IOS to enable proxy ARP on the interfaces. An increasing number of customers are choosing to disable proxy ARP believing that there is some amount of security exposure in it.
If you disable proxy ARP then the router will not respond to any ARP request for a destination that is not on the local subnet(s) of the router interface. This may or may not have much impact depending on who (and why) they are ARPing. There are a number of things that may lead to proxy ARP:
- if a host ARPs for every destination (no matter whether it is local or is remote) it will have connectivity if the router enables proxy ARP and will not have connectivity if the router does not enable proxy ARP. The host might do this if it is misconfigured. If you disable proxy ARP it will potentially create a problem for the end station. You may choose to do this to force the users to get a correct configuration on their workstations.
- if a router configures a static route which uses an interface over a LAN interface instead of next hop address (for example ip route 0.0.0.0 0.0.0.0 FastEthernet0) it will ARP for every unique destination address to which it is forwarding. If the next hop router enables proxy ARP then the static route will work (and create extra work on the router on which it is configured). If the next hop router does not enable proxy ARP then the static route will not work
You must log in to post a comment.