Distributed Extent Metadata for RAID Reliability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The centralized storage of metadata in RAID systems leads to reduced reliability and inefficiency, as damage to the metadata storage extent causes failure of the entire RAID mapping, and prevents expansion or reduction of allocated storage space, resulting in suboptimal storage management.
Innovation Solution
The storage system is divided into multiple extents, each with its own metadata storage region, allowing for independent replacement of damaged extents without affecting the RAID group, thereby increasing mapping reliability and enabling dynamic expansion and reduction of storage space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If metadata is stored centrally in a single extent, then the RAID system structure is simple, but the reliability decreases because damage to the metadata extent causes failure of the entire RAID mapping
Solution Approach 1:
The patent divides the centralized metadata storage into distributed metadata regions across multiple extents. Each extent contains its own metadata region storing mapping information specific to that extent, eliminating the single point of failure in centralized metadata storage. This segmentation improves reliability while the distributed structure itself becomes the new architecture.
2Adaptability or versatility
If storage space size is fixed after allocation, then the RAID group structure remains stable, but the adaptability decreases because expansion or reduction cannot be performed
Solution Approach 1:
The patent enables dynamic adjustment of storage space by allowing extents to be independently added or removed from the RAID group. The distributed metadata structure allows the system to adapt to changing storage requirements while maintaining data integrity through selective metadata updates rather than requiring complete restructuring.
3Productivity
If all I/O operations access centralized metadata, then the metadata management is simple, but the productivity decreases because the centralized metadata storage becomes a bottleneck
Solution Approach 1:
The patent segments the centralized metadata access into multiple distributed metadata regions across different extents. This allows I/O operations to access metadata in parallel from multiple locations rather than sequentially from a single centralized location, eliminating the bottleneck and improving throughput.
4Speed
If metadata is stored in the memory, then the access speed is fast, but the loss of information increases when the disk storing metadata is damaged
Solution Approach 1:
The patent implements redundancy by storing metadata information across multiple extents rather than in a single location. This distributed copying of metadata ensures that if one extent is damaged, the metadata information can be recovered from other extents, providing fault tolerance while maintaining access performance.
Data Source
AI summary
The present disclosure provides a storage management method, a device and a computer-readable medium. The method comprises: receiving a request for creating a storage space, the request at least comprising a storage capacity and a RAID configuration of the storage space, the RAID configuration at least indicating a RAID type; allocating an extent based on the storage capacity; creating a RAID group for the extent based on the RAID type; and storing metadata of the RAID group in the extent, the metadata indicating a configuration of the RAID group and a configuration of a user data region in the extent.