To allow communication (and nfs sharing) between host and VM it is necessary to create a bridge, which include the main physical interface.
This is the default configuration in other hypervisors like Proxmox.
After a reboot the WebUI is available for a few seconds, then when kube-bridge starts at boot the WebUI cannot be reached and it's impossible to ping the server.
I temporarly solved this issue with a fresh install of TrueNAS SCALE without settings a pool for Applications, so that Kubernetes scripts won't activate.
Replicate the problem:
In network menu:
- Select primary interface eno1, remove all the settings (ip, dhcp), apply.
- Create a bridge interface br01, include eno1 in the bridge, set static ip or DHCP. Apply
- Test changes then save changes. Now guest VM can ping host.
- Select the default pool for application (if not already done)
- Reboot. The WebUI won't be available