RAID Reconstruction Using Distributed Hot Spare Writing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing RAID technologies experience low reconstruction speed when a physical disk fails, as data from multiple disks must be written into a single hot spare disk, leading to bottlenecks and reduced reliability.
Innovation Solution
A method and apparatus that utilize a pre-generated layout template to distribute data from faulty disks across multiple physical disks, performing exclusive-OR operations to write data into different hot spare disks, thereby reducing the burden on a single disk and improving reconstruction speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data from multiple physical disks is written into a single hot spare disk during RAID reconstruction, then the reconstruction process can be completed, but the written speed of the hot spare disk becomes a bottleneck resulting in low RAID reconstruction speed
Solution Approach 1:
The patent segments the single hot spare disk into multiple hot spare disks, distributing the data writing task across multiple targets. Instead of consolidating all reconstructed data into one disk, the system divides the data strips and assigns them to different hot spare disks, thereby parallelizing the writing process and eliminating the single-point bottleneck that limited reconstruction speed.
Solution Approach 2:
The patent transitions from a single-disk reconstruction target to a multi-disk target architecture. By adding the dimension of multiple hot spare disks, the system expands the writing capacity from a single bottleneck to a distributed parallel system, enabling simultaneous writing operations across multiple disks and significantly improving overall reconstruction throughput.
2Productivity
If a single hot spare disk is used for all reconstruction operations, then the system structure remains simple, but the written speed becomes a bottleneck
Solution Approach 1:
The patent applies segmentation by dividing the hot spare disk resources into multiple independent hot spare disks. Each hot spare disk can independently receive and write data strips, creating parallel writing channels. This segmentation transforms the single-disk architecture into a multi-disk system that achieves higher aggregate writing speed while maintaining manageable complexity through modular organization.
Solution Approach 2:
The patent makes the hot spare disk system multi-functional by enabling multiple disks to perform the same reconstruction writing function simultaneously. Instead of one disk handling all reconstruction tasks sequentially, multiple hot spare disks can work in parallel, each handling different data strips, thereby universalizing the reconstruction capability across the disk subsystem.
Data Source
AI summary
Embodiments of the present invention provide a method and apparatus for reconstructing a RAID, and a system, when a first physical disk is faulty, finding DUs of other physical disks, where the DUs are located on a same LDS as a DU of the first physical disk, in each logic module of a pre-generated layout template, performing exclusive-OR on data read from the DUs of other physical disks, and writing the data into a hot spare DU of a second physical disk; and in the layout template, replacing a DU identifier of the first physical disk with a DU identifier of the second physical disk, where DUs corresponding to DU identifiers in each LDS in the layout template belong to different physical disks. The present invention can reduce time needed for writing the data, shorten time needed for RAID reconstruction and restoration, and improve a reconstruction speed.


