Description of problem:
The initial problem was with unstable floating ip with intermittent failures. Finding this in a big cloud was not easy as the stuck ovn-controller was on a different compute node than the node hosting the VM with the FIP. After it was found on the fabric where the packets are being sent it was observed the node has arp responder flows installed. ovn-controller was at 100% CPU utilization and was not writing to the logs and connection to the southbound database was in CLOSE_WAIT state.

Version-Release number of selected component (if applicable):
ovn2.13-20.12.0-192.el8fdp.x86_64

How reproducible:
Once in a very long time

Steps to Reproduce:
1. Unknown
2.
3.

Actual results:

Expected results:

Additional info:
OVN DBs, local OVS db, ovn-controller log file and the core dump will be linked because it won't fit the BZ due to core dump being 6GB large.