Rasilient PixelStor 5000, Руководство пользователя

Страница: 23 / 161

About This Manual
PixelStor 5000 User Manual
Rev. 1, May 2013
iii
tolerance.
Active-Active Controllers & Transparent Fail-over & Fail-back - Fast throughput and data
availability are paramount for dedicated storage systems. This is why PixelStor 5000 is
designed with active-active controllers to improve both performance and data availability.
Each Disk Group has an assigned “primary” controller to take advantage of the fastest route to
the drives. So, optimal performance is achieved when traffic goes through both controllers to
their associated Disk Groups. For availability, both controllers can serve any Disk Groups
within the system. When there is a controller failure Disk Groups assigned to the failed
controller will automatically “migrate” to the surviving controller without user intervention. When
the failed controller is replaced, the systems will automatically fail-back to its original state. In
PixelStor 5000, the relationship between two controllers is tight. Each controller has the
precise status of the other, which results in quick and accurate actions when failure happens.
Mirrored Cache -
With Dual controllers, the “auto cache” option puts a mirrored copy of the
write cache on the redundant controller. The data is fully protected. If there is a failure in any
controller, the system flushes the mirrored copy and automatically switches to the
“write-through” setting. This is the best compromise between performance and protection.
Multi-Path IO and Cache Coherency - The PixelStor 5000 makes no assumption regarding
the behavior of the hosts since they can switch from one path to the other. This is normally
referred as “multi-path IO” – a technology to ensure the end-to-end high availability for path
failure. RASILIENT Systems patent-pending technology supports path failure with our cache
coherency features. In the extreme case, the host can send commands to either controller
randomly.
Dedicated SATA Channels - PixelStor 5000
’s multi-path design provides a dedicated drive
channel from each controller. The elimination of inter-drive dependencies enhances
performance. The independent controller access provides additional data availability.
Disk Roaming
- In the PixelStor 5000, the concept of Disk Groups is introduced, which
represents a set of drives that share similar characteristics. The Disk Group configuration,
including RAID, and volume, roams with the drives. Once inserted, the Disk Group can be
discovered and the service can start with a mouse click. Within the enclosure, the drives can
also be moved around. The roaming operation can be carried out while the system is on-line.
Users can take advantage of this flexibility to arrange their storage for different on-line and
off-line tasks. This feature provides a basic and inexpensive way to upgrade a system and
retain your data or move drives to another system for basic system recovery.
Capacity Expansion
– The PixelStor 5000 has the best capacity expansion capability on the