A. Virtual machine data in the failed server rack is rebuilt immediately using parity information. vSphere HA powers off and restarts all of the virtual machines to rebalance storage utilization across the cluster.

B. All virtual machine data remains accessible. Virtual machines running on nodes in the failed server rack are restarted on nodes in the other two server racks. Virtual machines in the other two server racks continue to run.

C. Virtual machines running on nodes in the failed server rack remain offline for 60 minutes. Virtual machines in the other two server racks continue to run and are marked "Degraded".

D. Since there are four nodes in each server rack, erasure coding is used to rebuild data that was lost due to the server rack power failure. vSphere HA monitors the rebuild process and powers on all affected virtual machines after the rebuild process is complete.