Online Volume Repair for Compressed Data Storage Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current storage solutions face challenges in repairing errors in compressed data and metadata while maintaining online access, leading to increased latency and downtime due to the inability to perform online repairs when errors occur in volumes storing both together.
Innovation Solution
A machine-readable storage medium with instructions that can identify and repair errors in compression metadata and data while the volume is online, allowing for read, partial overwrite, or full page overwrite requests by alerting users, engaging in self-healing, or performing repairs at another location.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If offline repair is performed to ensure data integrity, then reliability is improved, but loss of time increases due to system downtime
Solution Approach 1:
The system performs preliminary actions by creating and storing repair information (error location, error type, repair method) in a repair information store before actual repair is needed. This allows the repair process to be pre-prepared and executed quickly when errors are detected, minimizing downtime while ensuring data integrity through verified repair procedures.
Solution Approach 2:
The patent introduces an intermediary repair information store that acts as a mediator between error detection and repair execution. This intermediary layer stores pre-analyzed repair information, allowing the system to quickly retrieve and apply repairs without performing complex analysis during downtime, thus reducing system unavailability while maintaining reliability.
2Reliability
If comprehensive error checking is performed to ensure data integrity, then reliability is improved, but processing power consumption increases
Solution Approach 1:
The system applies partial action by performing error checking only on specific data blocks that are being accessed or modified, rather than checking the entire volume. The repair information store contains pre-analyzed error data for specific blocks, allowing targeted verification without comprehensive scanning, thus reducing processing power consumption while maintaining reliable error detection for relevant data.
Solution Approach 2:
The patent implements local quality by maintaining repair information specific to individual data blocks rather than performing uniform comprehensive checking across the entire storage volume. Each block has its error characteristics stored locally in the repair information store, allowing the system to verify only the necessary portions of data with appropriate checking depth, optimizing the balance between reliability and processing power usage.
3Loss of time
If online repair is performed to minimize downtime, then loss of time is reduced, but device complexity increases due to concurrent operations
Solution Approach 1:
The system segments the repair process into distinct phases: error detection, repair information generation and storage, and repair execution. By dividing the complex online repair task into manageable segments with clear interfaces, the system can handle concurrent read/write operations and repair operations separately, reducing the complexity of managing concurrent operations while enabling continuous availability.
Solution Approach 2:
The repair information store serves as an intermediary that decouples error detection from repair execution. This intermediary layer allows the system to detect errors and prepare repair information without immediately executing repairs, enabling careful coordination of concurrent operations. The intermediary structure simplifies the management of concurrent read/write and repair operations by providing a buffer and coordination point.
Data Source
AI summary
Examples include repair of a volume while the volume is online. In some examples, a request is received to access a compressed page, wherein the compressed page includes compressed data and compression metadata, an error is identified in the compression metadata or the compressed data, and based on the error and the request, repair of the error is handled while the volume is online.