Users guide

33 Intel® RAID Software User Guide
• If the system shuts down, the initialization or rebuild process automatically
resumes on the next boot. Auto resume must be enabled prior to virtual drive
creation.
• Stripe size is user definable on a per drive basis and can be 8, 16, 32, 64, or 128 KB
in size. The default is 256 KB, which is optimal for many data access types.
• Hot spares can be set as global or dedicated. A global hot spare automatically
comes online to replace the first drive to fail on any array or disk group on the
controller. A dedicated hot spare is assigned to a specific array or disk group and
only comes online to rebuild this specific failed array or disk group. A hot spare
only comes online if it is the same size or larger than the failing drive (see drive
coercion as mentioned in the next point), and if a drive has been marked as failed. If
a drive is removed (and marked as failed) within a virtual drive, the hot spare
automatically comes online. However, there must be disk activity (I/O to the drive)
in order for a missing drive to be marked as failed.
• Drive coercion refers to the ability of the controller to recognize the size of the
physical drives connected and then force the larger drives to use only the amount of
space available on the smallest drive. Drive coercion allows an option to map out a
reserved space to compensate for slightly smaller drive sizes that may be added
later. The default is set to 1 GB. The coercion algorithm options are:
— None: No coercion of size.
— 128 MB: The software rounds the drive capacit
y down to the next 128 MB
boundary and then up to the nearest 10 MB until the coerced capacity is larger
than the actual drive size. It is then reduced by 10 MB.
— 1 GB: The software rounds the drive capacity down to the nearest 1 GB
bo
undary and then down by 1 MB. This corresponds to the terms most drive
manufacturers use.
3.5.2 Fault Tolerant Features
• Configuration on Disk (COD) and NVRAM (Non-volatile Random Access
Memory) storage of array and disk group configuration information. Array and disk
group configuration information is stored both on the hard drive (COD) and in
NVRAM. This helps protect against loss of the configuration due to adapter and/or
drive failure.
• Failed drives are automatically detected and a transparent rebuild of the failed array
automatically occurs using a hot-spare drive.
• Support for SAF-TE (SCSI Accessed Fault-Tolerant Enclosure) enabled enclosures
allows enhanced drive failure and rebuild reporting through enclosure LEDs
(Light-Emitting Diodes); support also includes hot swapping of hard drives.
• A battery backup for cache memory is available as an option. RAID controller
firmware automatically checks for the presence of the battery module, and if found,
allows the write back cache option. The adapter continuously tracks the battery
voltage and reports if the battery is low. If low, the battery is first given a fast
charge to replenish the charge and is then given a trickle charge to keep it at an
optimal power level. Adapters that support the battery module include a “dirty
cache” LED; when power is lost to the system and data remains in the cache