I/O operations during the rebuild (synchronization) process of the failed
disk unit may not require additional disk I/O requests. This depends on where
the data is read from or written to on the disk unit that is in the synchronization
process. For example:
- A read operation from the disk area that already has been rebuilt requires
one read operation.
- A read operation from the disk area that has not been rebuilt is treated
as a read operation on a failed disk unit. See Read operations on a failed disk unit for more information.
- A write operation to the disk area that has already been
rebuilt requires normal read and write operations (two read operations and
two write operations).
- A write operation to the disk area that has not been rebuilt is treated
as a write operation to a failed disk unit. See Write operations on a failed disk unit for more information.
Note: - The rebuild process takes longer when read and write operations to a replaced
disk unit are also occurring. Every read request or every write request interrupts
the rebuild process to perform the necessary I/O operations.
- RAID 6 operations are derived from RAID 5, but at a further level of
complexity. Because the concept is similar to RAID 5, RAID 6 read operations
are not described here