Resulting from a discussion on Apr 3 among various stakeholders, we would like to understand how Windows VMs configured in a WSFC behave when the active path goes down and the VM in configured with the default IO error policy (error=stop). In this case, the VM will be paused by libvirt and resumed every 30 seconds by kubevirt to retry the IO. We expect that multipath should failover the downed path within 20 seconds and the subsequent retry would use the new active path.

The goal of this test is to see what impact this long pause has on the Windows VM and the health of the WSFC environment and whether this could be good enough to give to customers in the short term while we develop a qemu/multipath fix that can switch the paths more quickly.