Feed on
Posts
Comments

What you’re seeing is the kernel limiting ICMP responses to 200/second.
If there are more than 200 ICMP requests per second, and you have
net.inet.icmp.icmplim set to 200 via sysctl (the default value), this
occurs.

This could be a ICMP flood attack. It could also be legimate traffic.
For your network, what would you consider to be a normal number of ICMP
requests per second?

231 packets/second is actually pretty slow if you’re on a high speed
local network, so in that situation it’s unlikely to be a deliberate
ping flood. I’ve had network monitoring tools that were badly configured
do something that looked much like this.

It might very well. If ‘named’ dies, and net.inet.udp.blackhole=0,
then the kernel will be generating ICMP error responses for UDP
packets sent to port 53.

SOLUTION

# sysctl net.inet.icmp.icmplim=500

Try that.

Also, if you want to make it permanent, add it to /etc/sysctl.conf:

net.inet.icmp.icmplim=500

(On a line all by itself)

Leave a Reply

Switch to our mobile site