Apacer Technology AP-SRFD255AXXXXR-G - Manual (Page 10)

Serial ATA Flash Drive
AP-SRFD255AXXXXR-G
10
© 2008 Apacer Technology Inc. Rev 0.2 (Preliminary)
The RAID Processor supports a definition of Safe (RAID-1) set. The HBA (Host Bus Adapter) does not need
to know anything about the physical arrangement of the data. To the HBA, the virtualized drive behind RAID
Processor is a normal drive that just happens to be safer (fault tolerant). All error handling and Data rebuild is
managed by the RAID Processor. The two physical partitions used for mirroring must reside on two different
physical disks, as there is no benefit of mirroring the data on the same drive. The advantage of mirroring is
essentially safeguarding the data in case of a single physical drive failure. For a read operation, the RAID
Processor decides which physical partition to read from and for write, the data is written to both drivers. Note:
There is no benefit of performance due to mirroring. Often the performance may suffer slightly, as the data will
be transferred to the least common denominator of the two physical drives. The two physical partitions used
for mirroring must reside on two different physical disks, as there is no benefit of mirroring the data on the
same drive.
2.8 RAID-1 Racing
The Storage Processor improves the speed of RAID 1 (SAFE) mode by introducing RAID 1 drive racing.
When reading a sector, both mirrored drives race and the first one to get the data supplies the information,
lowering the average seek time.
2.9 Sector Recovery and Disk Rebuild
For Safe (RAID-1) volumes, the firmware implements three additional features to make the data more reliable,
Sector Recovery, Disk Rebuild and Disk Verify.
2.9.1 Sector Recovery
For Safe (RAID-1) volumes, at any time, if the Storage Processor detects any read or write error, it
automatically (without host intervention):
Take the Physical Partition Off-line
Attempt to copy the respective data from the other mirrored Partition (if available).
If the re-write is successful, then the problem is solved before it got to be catastrophic.
Bring the Physical Partition On-Line.
In case the problem still persists after re-write (Read-Verify fails), then it is assumed that Physical Partition on
the Physical Disk is bad (and all spare sectors implemented by the HDD itself are depleted). In this event, this
Physical Partition (with Error) will stay off-line. Other Physical Partitions on this HDD are still good and stay
on-line. In addition, that HDD’s LED will show the Error State. In the rare case, if there is another sector