Slower Internet connections, network port transceiver indicating lamp high frequency flicker.
To view the CPU usage is high, higher than usual more than doubled.
To view the CPU usage found forwarding process CPU usage is high.
CPU Usage Stat. Cycle: 60 (Second)
CPU Usage : 89% Max: 100%
CPU Usage Stat. Time : 2012-12-12 09:01:15
CPU Usage Stat. Tick : 0xeec(CPU Tick High) 0x72079664(CPU Tick Low)
Actual Stat. Cycle : 0x0(CPU Tick High) 0x782026f5(CPU Tick Low)
TaskName CPU Runtime(CPU Tick High/CPU Tick Low)
DFWD 88% 0/720184a1
Show interface flow information
<USG2100>disp interface brief
*down: administratively down
InUti/OutUti: input utility/output utility
Interface PHY Protocol InUti OutUti inErrors outErrors
Cellular5/0/0 down up(s) 0% 0% 0 0
Ethernet0/0/0 up up 0% 0.01% 0 0
Ethernet1/0/0 down down 0% 0% 0 0
Ethernet1/0/1 up up 0.02% 78% 0 0
Ethernet1/0/6 up up 100% 100% 0 0
Ethernet1/0/7 up up 100% 100% 1 1
Show interface flow found part of the interface flow is very large, there are two interface flow similar, if more than one interface flow were similar, indicating the presence of loop may.To view the Ethernet 1/0/6 and 1/0/7 found that the interface transceiver in large quantity per second, which is the embodiment of the loop characteristics.
<USG2100>disp inter ethernet 1/0/6
Last 300 seconds input rate 42756224 bits/s, 83508 packets/s
Last 300 seconds output rate 42756224 bits/s, 83508 packets/s
Input: 48310854 packets, 3091894691 bytes
48310854 broadcasts, 0 multicasts
Output:48310856 packets, 3091894784 bytes
48310856 broadcasts, 0 multicasts
Further information IP statistics, found the IP statistical information is not abnormal, the loop is a two layer loop caused by.
<USG2100>disp ip statistics
Input: sum 401299 local 137708
bad protocol 0 bad format 0
bad checksum 0 bad options 0
TTL exceeded 147345
Output: forwarding 0 local 147477
dropped 0 no route 39535
Fragment: input 0 output 0
fragmented 0 couldn't fragment 0
Reassembling:sum 0 timeouts 0
Unplug the ethernet1/0/7 or 1/0/6 network, found the loop was removed, interface traffic returned to normal, the rate of use of CPU also returned to normal range.
Confirm the loop is in the Ethernet 1/0/6 and 1/0/7 produced between the lines, further investigation, and found that the two interface is cable short circuit, the removal of the reticle, the network back to normal.
Itself of this kind of phenomenon description of the network flow is very large lead, have two kinds of reasons:
1) normal business flow is very large;
2) network loop.
When the network speed is very slow, the first to troubleshoot a system CPU utilization is normal, if the CPU usage is high, then further view which process is CPU high usage rate, if forwarding process CPU high usage rate, further investigation and forward the traffic flow is normal or loop condition exists, if the switch, hybrid switch can be checked, switches are generally have a loop detection mechanism, can be very convenient location out of the loop where the.