| 摘要 |
If necessary, problematic sectors are isolated when an attempt to write a group of data blocks to corresponding disk sectors returns a write error by attempting to write each data sector individually, registering each individual sector which returns a write error for separate, further examination. Since disk drive vendors do not employ consistent write error messages in device drivers, true write faults-those caused by failure of disk media such that a sector becomes unreliable or unrecoverable-are then identified by an operating system device manager for the disk drive by writing the data block to a reserved replacement sector and monitoring the status of that write. If the write to the replacement sector also fails with a write error, the error signaled likely indicates a true drive status and an appropriate error is returned to the file system which issued the write request. If the write to the replacement sector succeeds, however, the write error most likely reflects a failing sector. Relocation of the failing sector may then be completed by simply mapping the original data sector to the replacement sector, and marking the original data sector as bad.
|
