Netonix Forums

Hey there,

I have observed that our Netonix switches kinda flood the L2 Network with 3 "Gratuitous ARP" broadcast packets every 10 seconds.

Seems this behaviour has been hardcoded into /usr/bin/monitor:

Code: Select all

...@switch-..-199:~# cat /usr/bin/monitor
#!/bin/sh

date +%Y > /tmp/current_start_date

while true; do
  #Perform gratuitous arp broadcast to update neighbors
  arping -U $(ifconfig | grep -Eo 'inet (addr:)?([0-9]*\.){3}[0-9]*' | grep -Eo '([0-9]*\.){3}[0-9]*' | grep -v '127.0.0.1') -c 3
  #[...]
 fi

 sleep 10
done
...@switch-..-199:~#


Is there a technical reason why the switches kinda aggressively ARP themselves 3 times every 10 seconds, causing some background broadcast noise when running a few dozen Netonix devices on a network?

3 arp broadcasts every 10 seconds is a flood?

Just think how many BPDU packets are flying around with RSTP.

I think I remember why we did this but it was a long time ago, so when i get a chance I will confere with Stephen.

Pretty sure it is needed otherwise your wating on mac entries to expire when there is a toplogy change.

Even if someone was running a HUGE flat network with 100 switches, which is not a ver good or wise idea, your not talking any significant amount of traffic.

Yea, so i checked with Stephen, i asked him to put that in back in 2019.

It solves the problem i expressed above when swapping out equipment standard arp or mac table entries TTL is 5min.

This insignificant traffic does nothing negative but insures a more accurate mac table and faster acknoledgment of equipment changes.

It was very pronounced when we serialIze switches and swap out a new switch at same IP but now has a different MAC.

I had also seen it when swapping out a switch at a tower it took time for local layer 2 equipment seeing the replacement.

Again compare this traffic to STP LACP Loop Protection or many other protocols in layer 2 not even talking about layer 3 protocols like OSPF and alike.

Very insignificant amount of traffic.