User manual

FIBRE-to-SAS/SATA RAID SUBSYSTEM
56
User Manual
RAID 5 – Striping with Distributed Parity
A RAID 5 RAID data is striped and
transferred to disks by independent
read/write operations. The data chunks
that are written are also larger than in a
RAID 0. The RAID 5 uses parity that is
striped across all disks in the RAID. You
need at least 3 disks to implement a
RAID 5. A RAID 5 can withstand a single
disk failure without loosing data or
loosing access to the data. The following
illustration shows a typical RAID 5 RAID.
Advantages
Disadvantages Ideal Use
Read data
transactions are very
fast.
Write data transactions
are somewhat slower (due
to the parity that has to
be calculated)
Disk failures have an
effect on throughput,
although this is still
acceptable.
A good all-round sys-
tem that combines
efficient storage with
excellent security
and decent
performance.
Ideal for file and
application servers.
RAID 6 – Striping with Dual Distributed Parity
The RAID subsystem’s RAID 6 function
provides the highest level of data
protection by tolerating multiple
simultaneous drive failures without
downtime or data loss. RAID 6 probability
of data loss is 1/1,000 (three orders of
magnitude) less that of an equivalent
RAID 5 array.
Where RAID 5 writes one parity block for each
stripe across the array, RAID 6 saves two
parity blocks, allowing the RAID controller to
continue operating with two inoperative drives,
without the loss of data or availability. As drive sizes have increased, the time required
to rebuild an inoperative drive’s data has stretched to several hours. During a rebuild,
data on the remaining drives within a RAID 5 array is unprotected. Should a second drive
fail, the rebuild will halt, and the entire RAID data may be lost. Because RAID 6 can
tolerate two simultaneous drive failures, your data remains fully protected during the
rebuild process. RAID set verification may be run ad-hoc or scheduled to test every
sector for proactive error detection. Furthermore, the system constantly monitors drive
health to proactively predict failure, and to automatically copy data to a spare drive
before the drive fails.