While installing the equipment for a new two-node cluster， it is discovered that the disk subsystem Rack has one connection to the power mains. The explanation given was that the disks are to be mirrored between the two disk drawers in the rack and that a failure in either of the disk drawers would simply result in half of the disk being lost. What can be done to avoid a single point of failure？（）
- A Customize HACMP to overcome quorum issues.
- B Add a second PDU to separate the power connections.
- C Use RAID5 when implementing two IO racks in a HACMP environment.
- D Add another shared disk to each volume group to avoid quorum issues during fallover.