Storage Management System for Corrupt Data Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
RAID-based storage systems face challenges in maintaining data integrity and availability during rebuild procedures, where access to corrupted data can be restricted or compromised, affecting data recovery and storage performance.
Innovation Solution
A method is implemented to identify corrupt storage targets within a data array, establish a 'broken' mirror with a relevant backup copy to identify differences, and overwrite corrupt data portions with non-corrupt data from the backup, allowing for efficient data recovery and maintenance of data integrity during rebuilds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional rebuild procedures are used to restore corrupted data in RAID systems, then data integrity is restored, but data access is restricted or compromised during the rebuild process
Solution Approach 1:
The rebuild process is segmented into two distinct paths: a fast path for healthy data blocks that maintains normal read access, and a slow path for corrupted blocks that performs targeted recovery. This segmentation allows most data access to proceed uninterrupted while only affecting specific corrupted portions during rebuild.
Solution Approach 2:
The system applies different quality levels of data access handling: reads for healthy blocks proceed with high priority and immediate access, while reads for corrupted blocks are routed to backup copies with lower priority. This local differentiation maintains overall system accessibility while focusing recovery resources where needed.
2Ease of operation
If data is restored from backup copies during rebuild, then data availability is maintained, but rebuild time and system performance are affected
Solution Approach 1:
The system performs preliminary identification of corrupted blocks using checksums or parity information before initiating full rebuild operations. This preliminary detection allows the system to prepare recovery strategies in advance, reducing actual rebuild time by having backup copies ready and pre-identified as recovery sources.
Solution Approach 2:
The fast path mechanism allows reads for healthy data blocks to skip the rebuild process entirely and access data immediately from the RAID array. This skipping of unnecessary recovery operations for unaffected blocks significantly reduces overall rebuild time while maintaining data availability for the majority of the dataset.
3Reliability
If multiple backup copies are maintained for RAID recovery, then data recovery reliability is improved, but storage overhead and system complexity increase
Solution Approach 1:
The system maintains backup copies at a level that exceeds the minimum single-failure threshold, preparing for multiple simultaneous failures. This partial redundancy approach stores more backup data than strictly necessary for single-block recovery, but provides enhanced protection against cascading failures without requiring full duplication of all data.
Solution Approach 2:
The patent uses distributed copying of data blocks across multiple locations within the RAID array, where each block has associated backup copies stored on different physical devices. This copying strategy improves recovery reliability by ensuring backups are isolated from the original failure, while the selective use of copies only when corruption is detected minimizes unnecessary storage overhead.
Data Source
AI summary
A method, computer program product, and computing system for identifying a storage target that has become corrupt within a data array. The storage target includes one or more corrupt data portions. A relevant backup copy of the storage target is identified. A “broken” mirror is established between the storage target and the relevant backup copy that identifies one or more differences between the storage target and the relevant backup copy. The storage target is rebuilt by overwriting the corrupt data portions within the storage target with corresponding non-corrupt data obtained from the relevant backup copy.