Layered RAID Fast Rebuild via Reserved Space
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data storage architectures face challenges in efficiently rebuilding a failed data storage device within a RAID implementation, which is a time-intensive process that increases downtime.
Innovation Solution
The system employs a scale out storage architecture with protection domains, utilizing reserved data storage space within each domain to quickly rebuild a failed data storage device by redistributing data across multiple devices, implementing a modified RAID 6 configuration that allows for rapid recovery from single or multiple device failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional RAID rebuilding is used, then data redundancy is maintained, but rebuild time is excessive and downtime increases
Solution Approach 1:
The system pre-allocates reserved space on each data storage device before failure occurs. This reserved space is prepared in advance and can be immediately utilized when a device fails, eliminating the need to allocate space during the rebuild process and significantly reducing rebuild time
Solution Approach 2:
The RAID array is divided into multiple protection domains, with each domain containing a subset of data storage devices. When a device fails, only the affected protection domain needs to be rebuilt rather than the entire array, segmenting the rebuild operation into smaller, faster tasks
2Speed
If reserved space is allocated for fast rebuild, then rebuild speed improves, but storage capacity is reduced
Solution Approach 1:
Each data storage device allocates reserved space locally based on its own capacity and the failure scenarios it needs to support. The reserved space ratio is configured per-device rather than uniformly across the entire array, allowing optimization of rebuild speed while minimizing overall capacity loss
Solution Approach 2:
The system allocates reserved space that is sufficient to handle single or multiple device failures within a protection domain, but not excessive. The reserved space is calculated to be exactly what is needed for the rebuild operation, avoiding unnecessary capacity reduction while ensuring fast rebuild capability
Data Source
AI summary
A System, Computer program product, and computer-executable method of rebuilding a failed data storage device within a storage architecture including a plurality of protection domains including two or more data storage devices, wherein a RAID protection is generated over volumes provided from the plurality of protection domains, the System, Computer program product, and computer-executable method including receiving a notification of a failed data storage device of the data storage devices within a first protection domain of the plurality of protection domains and rebuilding the failed data storage device using reserved data storage within a first data storage device within the first protection domain.