We do not have the ability to create a secondary network on AWS infra that we control that spreads across multiple EC2 instances. However, we can create a secondary network within a single EC2 instance, and host all the KubeVirt VMs associated with a NodePool on that single Ec2 instance. This allows us to simulate testing bridge cni integration for custom network interfaces on AWS infrastructure.

We need a hypershift/kubevirt e2e test that exercises creating a KubeVirt platform HCP with nodepools that us a custom network. This test should also include verifying that cloud-provider-kubevirt does not mirror LBs in this scenario.

We have to play with dev-scripts EXTRA_NETWORKS https://github.com/openshift-metal3/dev-scripts/blob/master/config_example.sh#L346 at baremetalds CI

There is also an EXTRA_NETWORKS openshift ci step

/ci-operator/step-registry/baremetalds/devscripts/conf/extranetwork/

How to access hypervisor hypervisor

Packet ssh key is at     name: e2e-kubevirt-mce-baremetalds-conformance-cluster-profile .data.packet-ssh-key

Packet ip is at e2e-kubevirt-mce-baremetalds-conformance .data.server-ip

user root