Cluster is unresponsive after sometime on VirtualBox hypervisor

[LalatenduMohanty]

While trying the instance on Virtualbox + macOS the network interfaces of the VMs does not work anymore after 1200secs of use. OIt looks like a timer triggers and the network driver gets reset, which causes the IP address to disappear and renders the instance unusable.

After we do ifdown/ifup for the interface, the instance works again and this never re-occurs. A screenshot and dmesg log can be found here: https://gist.github.com/gbraad/e492ea1e3444aa80a1d8cd7d48956982

We have tried different hardware emulation for the NIC, but all have the same problem (e1000, virtio, pcnet).

Workaround

• If your cluster lost connection then open the virtualbox manager and select the VM then login using username core and password coreos.
• Then become root user using sudo su and perform ifdown eth0 then followed by ifup eth0

[gbraad]

https://bugzilla.redhat.com/show_bug.cgi?id=1664548