Hi Folks, In the context of ovn, the way the ovn agents like ovn controller, ovn metadata etc, broadcast their aliveness status is by updating certain values in the local ovs db tables to show they are alive , which is then picked up by the ovn controller on that node and it updates the southbound db which passes on to the north daemon and it informs neutron. is this the correct understanding of how ovn agent check works ? i am facing an issue, where three neutron server container hosts are marked as 'gw as a chassis' to serve dhcp and neutron metadata requests for sriov based vms. these are grouped into a single ha chassis group with a priority no for each chassis. when I stopped one of the ovn metadata agents on an active chassis, the ovn is not changing the chassis entry of that sriov port in the port binding table of the sb. it is still showing the old chassis entry. I checked for any updates in the ovn controller logs of the new active and passive chassis' . but nothing in the logs. is this a bug or some internal communication is not happening among the chassis' ovn controllers, northd , nb and sb ? Thanks elinux