Users guide

7 Intel® RAID Software User Guide
“ABCsum”. When drive A fails, the controller uses the ABCsum to calculates what
remains on drives B+C. The remainder must be recreated onto new drive A.
Parity can be dedicated (all parity stripes are placed on the same drive) or
distributed
(parity stripes are spread across multiple drives). Calculating and
writing parity slows the write process but provides redundancy in a much smaller
space than mirroring. Parity checking is also used to detect errors in the data during
consistency checks and patrol reads.
RAID 5 uses distributed parity and RAID 6 uses dual distributed parity (two
dif
ferent sets of parity are calculated and written to different drives each time.)
RAID modes 1 and 5 can survive a single disk failure, although performance may
be degraded, especially during the rebuild. RAID modes 10 and 50 can survive
multiple disk failures across the spans, but only one failure per array. RAID mode 6
can survive up to two disk failures. RAID mode 60 can sustain up to two failures
per array.
Data protection is also provided by running calculations on
the drives to make sure data is
consistent and that drives are good. The controller uses consistency checks, background
initialization, and patrol reads. You should include these in regular maintenance
schedules.
• The consistency check operation verifies that data in the array matches the
redundancy data (parity or checksum). This is not provided in RAID 0 in which
there is no
fault tolerance.
• Background initialization is a consistency check that is forced five minutes after the
creation of a virtual disk. Background initialization also checks for media errors on
physical drives and ensures that striped data segments are the same on all physical
drives in an array.
• Patrol read checks for physical disk errors that could lead to drive failure. These
checks usually include an attempt at corrective action. Patrol read can be enabled or
disabled with automatic or manual activation. This process starts only when the
RAID controller is idle for a defined period of time and no other background tasks
are active, although a patrol read check can continue to run during heavy I/O
processes.
1.3.2 Enclosure Management
Enclosure management is the intelligent monitoring of the disk subsystem by software or
hardware usually within a disk enclosure. It increases the ability for the user to respond to
a drive or power supply failure by monitoring those sub systems.
1.3.3 Performance
Performance improvements come from multiple areas including disk striping and disk
spanning, accessing multiple disks simultaneously, and setting the percentage of
processing capability to use for a task.