Following https://issues.redhat.com/browse/CNV-28040
On CNV, when virtual machine, with secondary interfaces connected with bridge CNI, is live migrated we observe disruption at the VM inbound traffic.

The root cause for it is the migration target bridge interface advertise before the migration is completed.

When the migration destination pod is created an IPv6 NS (Neighbor Solicitation)
and NA (Neighbor Advertisement) are sent automatically by the kernel.
The switches at the endpoints (e.g.: migration destination node) tables
get updated and the traffic is forwarded to the migration destination before
the migration is completed [1].

The solution is to have the bridge CNI create the pod interface in "link-down" state [2], the IPv6 NS/NA packets are avoided, CNV in turn, set the pod interface to "link-up" [3].

CNV depends on bridge CNI with [2] bits, which is deployed by cluster-network-operator.

[1] https://bugzilla.redhat.com/show_bug.cgi?id=2186372#c6
[2] https://github.com/kubevirt/kubevirt/pull/11069
[3] https://github.com/containernetworking/plugins/pull/997  

4.16.0

100%

    1.
    2.
    3.
    

CNO deploys CNI bridge w/o an option to set the bridge interface down.

CNO to deploy bridge CNI with [1] changes, from release-4.16 branch. 

[1] https://github.com/containernetworking/plugins/pull/997

More https://issues.redhat.com/browse/CNV-28040