GNU/Linux ◆ xterm-256color ◆ bash 663 views

This is a quick example of a dual-node KubeVirtCI cluster running a migration over a secondary network.

The clip first shows the list of environment variables used to properly configure the KubeVirtCI cluster. Those were defined before running make cluster-up.

It then shows and creates the a NAD on the secondary network (accessed through eth1 on every node). IPs are managed by whereabouts.

The KubeVirt custom resource is then edited to enable live migrations and define which NAD to use for migrations. That edit triggers a restart of the virt-handler pods.

Looking at one of the new virt-handler pods, we can see it is now connected to the migration bridge with an IP from whereabouts.

We then start a migratable VMI and wait for it to start, on node01.

Finally, we create a migration object against that VMI, and we can see it moved to node02. The migration status shows a target node address in the whereabouts IP range, showing it indeed used our dedicated migration network!