Intel AFCSASRISER, Руководство пользователя, Страница: 36 / 192

24 Intel
®
RAID Software User’s Guide
Typically, when a drive fails or is expected to fail, the data is rebuilt on a hot spare. The failed
drive is replaced with a new disk. Then the data is copied from the hot spare to the new drive,
and the hot spare reverts from a rebuild drive to its original hot spare status. The copyback
operation runs as a background activity, and the virtual drive is still available online to the
host.
Copyback is also initiated when the first Self-Monitoring Analysis and Reporting Technology
(SMART) error occurs on a drive that is part of a virtual drive. The destination drive is a hot
spare that qualifies as a rebuild drive. The drive with the SMART error is marked as "failed"
only after the successful completion of the copyback. This avoids putting the drive group in
degraded status.
Note: During a copyback operation, if the drive group involved in the copyback is deleted because of a
virtual drive deletion, the destination drive reverts to an Unconfigured Good state or hot spare state.
Order of Precedence
In the following scenarios, rebuild takes precedence over the copyback operation:
1. If a copyback operation is already taking place to a hot spare drive, and any virtual drive
on the controller degrades, the copyback operation aborts, and a rebuild starts. The
rebuild changes the virtual drive to the optimal state.
2. The rebuild operation takes precedence over the copyback operation when the
conditions exist to start both operations. For example:
— Where the hot spare is not configured (or unavailable) in the system.
— There are two drives (both members of virtual drives), with one drive exceeding the
SMART error threshold, and the other failed.
— If you add a hot spare (assume a global hot spare) during a copyback operation, the
copyback is aborted, and the rebuild operation starts on the hot spare.
Configuration Planning
Factors to consider when planning a configuration are the number of physical disks the RAID
controller can support, the purpose of the array, and the availability of spare drives.
Each type of data stored in the disk subsystem has a different frequency of read and write
activity. If you know the data access requirements, you can more successfully determine a
strategy for optimizing the disk subsystem capacity, availability, and performance.
Servers that support video-on-demand typically read the data often, but write data
infrequently. Both the read and write operations tend to be long. Data stored on a general-
purpose file server involves relatively short read and write operations with relatively small
files.
Dimmer Switch Feature
Powering and cooling drives represents a major cost for data centers. The new MegaRAID
Dimmer™ Switch reduces the power consumption of the devices connected to a MegaRAID
controller. This helps to share resources more efficiently and lower costs.