Smart Array backplane RAID controllers using parallel
                      SCSI expect to be connected to enclosures that have
                      individual built in enclosure processors. This
                      processor or Environmental Monitoring Unit (EMU)
                      monitors fans, controls LEDs and so on. The EMU also
                      provides information about hot-plug events available
                      in the enclosure. The Smart Array controller is thus
                      always aware of the insertion or removal of a SCSI
                      disk drive when used with such type of enclosures,
                      like MSA30.

                      The 64xx series Smart Array card can also be cabled
                      to the internal parallel SCSI drives on applicable I64
                      platforms. These drives are housed in an enclosure
                      that is totally passive with respect to hot-plug event
                      notification. The Smart Array controller cannot
                      detect that a physical disk has been hot-inserted.
                      If a RAID physical disk fails and must be replaced,
                      the controller will not start the recovery process
                      on the affected RAID volumes.

                      After replacing the bad disk(s) with the new disk(s),
                      if the system is rebooted before issuing the SCAN ALL
                      command, then all the affected RAID volumes configured
                      on these disks will move to the "VOLUME ready for
                      recovery operation" state. The REBUILD operation will
                      not start automatically.

                      The "START RECOVER" command will initiate the REBUILD
                      operation for all the UNITs configured in the
                      disk group sequentially.

                      Note that "hiding" the insertion of a new disk from the
                      controller can result in the REBUILD operation not
                      starting automatically, even if the controller can detect
                      hot-plug events, such as when connected to a MSA30.
                      Removing a disk while the host system is running, and
                      replacing the disk while the host system is powered off,
                      will effectively hide the insertion. The START RECOVER
                      command will initiate the REBUILD operation in such
                      scenarios also.

                      This command is supported on both OpenVMS Alpha and
                      OpenVMS I64 platforms.

                      Note: Refer to SCAN ALL command for additional
                            information on rebuild of volumes.

                      Format:
                              START RECOVER