Parallel File System Upgrade in Virtual Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional file system upgrades in data storage systems are time-consuming, labor-intensive, and expensive due to the need for input and output operations on each block, which scales with the number of blocks, making it difficult to efficiently upgrade, replace, and decommission disk drives while maintaining data integrity and reducing system footprint.
Innovation Solution
The implementation of a concurrent file system upgrade method that identifies boundaries of hard disk sets within a block management data structure, allowing for parallel retrieval, upgrade, and reconstruction of metadata across multiple hard disks, resulting in an upgraded file system with reduced time-to-upgrade proportional to the largest set of hard disks rather than the entire system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional sequential file system upgrade method is used, then data integrity is maintained, but upgrade time scales linearly with the number of blocks making it extremely time-consuming
Solution Approach 1:
The patent divides the file system into multiple independent sets of blocks, where each set can be upgraded separately and concurrently. By segmenting the upgrade process into independent units of work, the system can process multiple block sets in parallel, reducing total upgrade time from O(n) to O(log n) or constant time relative to total block count, while maintaining data integrity through controlled concurrency boundaries
Solution Approach 2:
The patent performs preliminary actions by first identifying all block sets that need upgrading and preparing the upgrade plan before execution. Metadata is retrieved and prepared in advance, and the system establishes the concurrent upgrade structure beforehand, allowing the actual upgrade to proceed efficiently with pre-computed information rather than sequential processing
2Manufacturing precision
If conventional IO-intensive upgrade process is used, then each block is properly upgraded, but the process becomes labor-intensive and expensive due to sizeable IO operations
Solution Approach 1:
The patent merges multiple upgrade operations into a single concurrent execution framework. By combining the retrieval, upgrading, and writing operations across multiple block sets into a coordinated concurrent process, the system reduces total I/O operations and eliminates redundant processing, achieving both completeness and efficiency simultaneously
Solution Approach 2:
The patent creates a mapping structure that copies the logical organization of blocks into sets that can be processed in parallel. By establishing a concurrent block set structure that mirrors the original file system organization, the system can retrieve and upgrade multiple blocks simultaneously while maintaining the integrity relationships through the mapping layer
3Stability of the object's composition
If file system upgrade is performed on entire system at once, then consistency is maintained, but system footprint cannot be reduced and downtime is extended
Solution Approach 1:
The patent segments the file system into independent block sets that can be upgraded in parallel while maintaining consistency within each segment. This segmentation allows the system to reduce footprint incrementally by migrating data to new storage devices in concurrent operations, rather than requiring complete system downtime or maintaining full system footprint throughout the upgrade process
Data Source
AI summary
A method for upgrading a file system in a data storage environment is provided. A storage system identifies a set of boundaries of one or more sets of hard disks based on a first block management data structure (BMDS), where the first BMDS includes information describing a storage volume representing the one or more sets of hard disks. For each identified boundary, the storage system identifies a data block range of a set of data blocks, where the data block range includes metadata of one set of hard disks that is stored on the set of data blocks. Further, the storage system concurrently generates a plurality of second BMDSs, each of the second BMDS corresponding to one of the sets of hard disks, including simultaneously retrieving metadata of the sets of hard disks, where the second BMDSs are constructed based on the retrieved metadata, representing an upgraded file system.