ARP Storm Taking Down Default Gateway
Posted on 2013-01-05
We have been experiencing a problem in our local network where the default gateway is being taken down due to what appears to be an ARP storm.
Originally, the default gateway was set to a Cisco 2851 router that routed traffic between several VLANs and had its default route set to our Cisco ASA router. Both the 2951 and the ASA were connected to a Cisco 2560 switch.
When the outage occurred, we lost all routing from the 2851 although we could still access it via Telnet. Clearing the ARP cache would instantly bring all functionality back. We saw a large amount of ARP requests coming in (thousands per minute) and the routing would go back down within about 15 minutes.
To test, we changed the default gateway (set by DHCP) to the ASA router. We experienced the same behaviour of ARP traffic and it would take down the internal interface of the ASA. Clearing ARP instantly brought all functionality back.
We also tried setting up a temporary internet gateway using a Cradlepoint router hooked to a Verizon aircard. It was connected through an intermediant HP switch that was connected to the 2560 switch. After an hour or so, the Cradlepoint was overwhelemed and also went down.
A little more information: We experienced this behaviour two days in a row. Communication inside the same subnets worked fine. Routing would go down around 9:30 AM each day and everything would settle down and become stable around 4:30 pm.
We think the problem is originating from a laptop and only starts happening when the employee arrives to work and then it stops when the employee leaves with their laptop.
Is there any other likely cause to this problem? If it is a laptop, what is the best way to handle this problem? We can wait until it starts happening again on Monday and disconnect switches and ports until we identify the culprit. However, I'd like to prevent any more downtime.
Thanks in advance.