Self-Healing RAID Geometry Conversion for Fault Tolerance

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In storage systems, when a device fails in a logical storage unit, especially in closed box arrays where spare drives are difficult to replace, existing technologies struggle to convert the critical unit to a fault-tolerant state without data loss, especially when no spare devices are available.

Innovation Solution

The method involves detecting device failures, draining data from the critical logical storage unit to fault-tolerant units, modifying the RAID geometry based on available discs, and rebalancing user data across the storage array to convert the critical unit into a fault-tolerant one.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If data is drained from the critical logical storage unit to fault-tolerant units, then the critical unit can be converted to fault-tolerant state, but system productivity decreases due to data migration overhead

Engineering Contradiction:

Improvefault toleranceVSAvoiddata migration overhead

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The system performs data draining and storage unit conversion in advance before the failure becomes critical, allowing the system to proactively reconfigure storage geometry and redistribute data before operational impact occurs. This preliminary action enables the critical unit to be converted to fault-tolerant state without interrupting ongoing operations.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system dynamically adjusts RAID geometry based on the number of available discs after failure, transforming the storage configuration from a static to a dynamic structure. The conversion process adapts the redundancy level and data distribution strategy in real-time according to the actual hardware availability, optimizing the balance between fault tolerance and performance.

Inventive Principle:

Principle #15Dynamics

2Adaptability or versatility

If RAID geometry is modified based on available discs, then the storage system adapts to hardware failure, but device complexity increases due to geometry conversion

Engineering Contradiction:

Improvehardware failure adaptationVSAvoidgeometry conversion

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system changes key parameters of the RAID geometry including redundancy level, stripe size, and data distribution pattern based on the number of available discs. By adjusting these parameters dynamically, the system adapts to hardware failures while maintaining operational integrity. The parameter changes enable flexible reconfiguration without requiring complete system redesign.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The conversion mechanism serves multiple functions: it detects hardware failures, recalculates optimal RAID geometry, migrates data, and reconfigures storage units. This multi-functional approach consolidates several complex operations into a unified process, reducing the net increase in device complexity while enhancing adaptability.

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Reliability

If user data is rebalanced across fault-tolerant storage units, then data integrity is maintained, but loss of time occurs during the rebalancing process

Engineering Contradiction:

Improvedata integrityVSAvoidrebalancing duration

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The data rebalancing process is performed continuously in the background during normal storage operations rather than requiring a complete system pause. The system maintains continuous data protection while progressively redistributing data across fault-tolerant units, ensuring data integrity is maintained throughout the process without complete operational interruption.

Inventive Principle:

Principle #20Continuity of useful action

Solution Approach 2:

The system performs partial rebalancing in phases rather than complete data migration at once. By moving data in manageable chunks and maintaining operational capacity throughout, the system reduces the perceived time loss while ensuring complete data integrity. The excessive action of maintaining full operational capacity during rebalancing compensates for the extended duration.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS10901634B2Self-healing in a storage system where critical storage group is rebuilt with different raid geometry using remaining functioning drives (in the critical storage group) that have not failed, and not all data, drained from the critical storage group, is stopped in the critical storage group that is rebuilt

Publication Date: 2021.01.26 SEAGATE TECH LLC

AI summary

A storage system may include a plurality of logical storage units that each include a plurality of storage devices. One or more logical unit numbers may be stored across one or more of the plurality of logical storage units, and the logical unit numbers may be accessible by one or more host devices. A logical storage unit may include a plurality of storage devices. Upon detection of failure of a storage device of a logical storage unit, data of the logical storage unit is drained to one or more fault tolerant logical storage unit. The logical storage unit with the defective device is converted to a fault-tolerant logical storage unit using the available and non-defective devices. Data is rebalanced across the fault-tolerant logical storage units.