Metadata Storage Unit Hot-Swapping for I/O Continuity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In a storage system, when the metadata storage unit fails, input/output operations for other storage units cannot be performed normally due to the loss of mapping information, leading to a buildup of unprocessed I/O requests.
Innovation Solution
A method is implemented to allocate a second metadata storage unit from a storage resource pool, where metadata is read into memory before failure, recorded in configuration information, and then stored in the second metadata storage unit, replacing the failed first metadata storage unit.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single metadata storage unit is used to store metadata of multiple storage units, then the storage system structure is simple, but the system reliability deteriorates when the metadata storage unit fails
Solution Approach 1:
The patent divides the metadata storage function into multiple independent metadata storage units (first metadata storage unit and second metadata storage unit). Each metadata storage unit can independently store metadata, so when one fails, the other can take over, improving reliability while maintaining relatively simple system structure through functional segmentation.
Solution Approach 2:
The patent changes the parameter of metadata storage from single-unit to multi-unit configuration. By introducing a standby metadata storage unit and implementing hot-swappable architecture, the system transitions from a single point of failure to a redundant configuration, improving reliability without significantly increasing operational complexity.
2Reliability
If metadata is stored only on physical disk without memory caching, then data consistency is ensured, but I/O operation speed deteriorates
Solution Approach 1:
The patent implements preliminary action by pre-loading metadata from the metadata storage unit into the metadata cache in memory before actual I/O operations occur. This allows the system to have metadata readily available in fast memory, significantly improving I/O operation speed while maintaining data consistency through the validation mechanism.
Solution Approach 2:
The patent introduces a metadata cache as an intermediary component between the physical disk and the I/O processing system. This cache layer accelerates metadata access speed while the validation mechanism ensures data consistency, effectively mediating between the conflicting requirements of speed and reliability.
3Device complexity
If a metadata storage unit fails and is not replaced, then system simplicity is maintained, but I/O request processing deteriorates due to request pileup
Solution Approach 1:
The patent implements self-service through automatic failure detection and hot-swappable metadata storage unit replacement. When a metadata storage unit fails, the system automatically detects the failure and can switch to the standby unit without requiring complex manual intervention, maintaining system simplicity while ensuring continuous I/O request processing capability.
Solution Approach 2:
The patent provides beforehand cushioning by maintaining a standby metadata storage unit that is ready to take over immediately upon failure of the active unit. This preemptive redundancy ensures that I/O request processing continues without interruption or significant delay, preventing request pileup while keeping failure management relatively simple through automated switching.
4Quantity of substance
If metadata is not read into memory before failure, then memory resource usage is reduced, but data recovery capability deteriorates
Solution Approach 1:
The patent applies preliminary action by reading metadata into the metadata cache in memory before failures occur. This ensures that critical metadata is already in fast memory and can be quickly accessed or backed up when failures happen, improving data recovery capability while managing memory resource usage through selective caching of essential metadata.
Solution Approach 2:
The patent extracts critical metadata from the physical disk storage and places it into the memory-based metadata cache. This extraction of essential data to fast memory improves data recovery capability and I/O performance, while memory resource usage is managed by caching only the necessary metadata rather than all data.
Data Source
AI summary
A technique for managing storage units involves: if it is determined that a first metadata storage unit for storing metadata of a plurality of storage units fails, allocating a second metadata storage unit for replacing the first metadata storage unit from a storage resource pool, wherein the metadata is read into a memory before the failure occurs, and first metadata of the first metadata storage unit is recorded in configuration information for indicating a valid metadata storage unit; recording second metadata of the second metadata storage unit in the configuration information; storing the metadata in the memory into the second metadata storage unit; and replacing the first metadata storage unit with the second metadata storage unit by validating the second metadata and invalidating the first metadata in the configuration information. Accordingly, such a technique improves reliability of a metadata storage unit.