File System Clone Management via Selective Block Copying
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
File system replication methods often result in deep copying of clone files, leading to increased storage costs due to non-sharing of data blocks between clone files and their snapshot-files, which is inefficient and costly.
Innovation Solution
A method for managing file systems that involves identifying diverged metadata portions in object tree structures, allowing for the deletion or truncation of clone files and snapshot objects, and transferring ownership of data blocks from parent snapshot objects to child objects, thereby optimizing storage usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If deep copying of clone files is performed during file system replication, then data integrity and independence are improved, but storage costs increase due to non-sharing of data blocks
Solution Approach 1:
The patent uses selective copying instead of deep copying. During replication, the system checks whether data blocks are shared between source and target file systems using reference counting. Only diverged portions (blocks that have been modified) are copied to the target, while shared blocks remain referenced rather than duplicated. This reduces storage consumption while maintaining data integrity through the use of copy-on-write mechanisms for diverged blocks.
Solution Approach 2:
The patent merges the replication process with the decloning process. When a clone file is deleted, the system performs both replication and decloning operations simultaneously, transferring ownership of data blocks from parent snapshot objects to child objects and eliminating unnecessary duplications in a single coordinated operation, thereby reducing overall storage costs.
2Ease of manufacture
If clone files are handled as regular files during replication, then replication simplicity is improved, but storage efficiency deteriorates due to creation of unnecessary deep copies
Solution Approach 1:
The patent applies different treatment to different portions of clone files based on their characteristics. Shared data blocks are handled with reference counting and soft links, while diverged blocks are handled with copy-on-write. This local differentiation allows the system to maintain simplicity for shared portions while optimizing storage efficiency for diverged portions, avoiding unnecessary deep copying.
Solution Approach 2:
The patent changes the replication parameter from deep copy to selective copy based on block divergence status. By tracking modification timestamps and using reference counting, the system dynamically determines whether to copy or reference each data block, transforming the replication process from a uniform deep copy operation to a selective operation that improves storage efficiency.
3Stability of the object's composition
If parent snapshot objects retain ownership of shared data blocks, then snapshot consistency is improved, but storage overhead increases due to redundant ownership claims
Solution Approach 1:
The patent implements self-service through automatic reference counting and ownership transfer mechanisms. When child objects are created from parent snapshot objects, the system automatically tracks shared data blocks using reference counts. When the parent snapshot is deleted or modified, the system automatically transfers ownership of diverged blocks to child objects without manual intervention, maintaining consistency while reducing storage overhead.
Solution Approach 2:
The patent replaces manual ownership management with an automated reference counting system. Instead of manually tracking which objects own which data blocks, the system uses reference counts to automatically determine ownership and trigger transfers when appropriate, reducing storage overhead while maintaining snapshot consistency through automated rather than mechanical manual processes.
Data Source
AI summary
The present disclosure relates to management of cloned objects and corresponding snapshot objects in file systems by a file server. Aspects may relate to one or more processes for managing file system objects of a file system in a data storage system comprising a file system server, wherein the file system server comprises: a computer system including a processor configured to execute an operating system and further software components executed on the operating system, and a programmable logic device interfacing with the computer system and including a file system management unit configured to manage a file system.