| 摘要 |
A disk array architecture which supports compound RAID configurations and which can automatically switch between various RAID configurations in case of drive failures to prevent data loss and retain data redundancy. In the preferred embodiment, the disk array system of the present invention begins operation with an array configuration that implements a maximum of data redundancy using all of the available disk drives. As drive failures occur, the disk array system automatically reconfigures one or more of the drives to other RAID configurations which utilize less data redundancy and thus require a lesser number of drives. This automatic reconfiguration occurs until the least amount of redundancy remains, or no redundancy remains. In this manner, the disk array system initially takes advantage of all of the available drives for maximum performance and redundancy while still retaining a high level of fault tolerance. Also, no human intervention is immediately required to replace a failed disk drive because the disk array system automatically reconfigures one or more drives to a different scheme to compensate for the failed drive while still retaining some level of redundancy. As subsequent drive failures occur after a first failure and prior to installation of respective replacement drives, no data will be lost up to a point because the disk array system simply reconfigures one or more of the drives to a different RAID scheme using less redundancy. |
