[c-nsp] ETH spikes

[Howard Jones]

Dracul wrote:
> Hi All,
>
> Anybody experienced sudden big spikes with their eth interfaces in their
> routers? WAN bandwidth is just 2MB but suddenly the MRTG graphs
> registered a 100MB bandwidth on the eth0 going to the Proxy server. although
> on the proxy server eth1 (going to customer network) everything is normal.
>
> network fluke or something?
>   
Did you restart the SNMP service on the machine or clear the counters? 
Sometimes you get wierd things like this when MRTG/Cacti/Whatever looks 
every five minutes and sees that the uptime is still going up, but the 
octets counter has gone back to zero, and the only 'logical' explanation 
is that you passed 100M of traffic in the last 5 minutes to wrap the 
counters.

Howie