Writeable Clone for Read-Only File System Without Snapshot Interdependencies
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage management techniques face challenges in making a read-only file system appear writeable, particularly in scenarios where interdependencies among snapshots complicate management and the technique is not applicable when the entire file system is read-only, such as in storage mirroring systems.
Innovation Solution
A method is introduced to create a writeable clone of a read-only persistent point-in-time image (RPPI) that appears modifiable from a client's perspective without breaking data consistency, by storing the writeable clone outside the active file system and using technologies like FlexClone to share data blocks, allowing changes to be made without overwriting original data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a read-only RPPI is made writeable by creating a virtual disk file in the active file system, then write operations can be performed on the RPPI, but interdependencies among snapshots are created that complicate management
Solution Approach 1:
The patent extracts the writeable clone from the active file system and places it in a separate clone file system. This separation removes the interdependency between the RPPI and the active file system, allowing the RPPI to be made writeable without creating management complexities in the snapshot hierarchy.
Solution Approach 2:
The patent introduces a clone file system as an intermediary layer between the read-only RPPI and the client. This intermediary allows write operations to be performed on the RPPI through the clone without directly modifying the original RPPI or creating dependencies in the active file system.
2Reliability
If the entire file system is read-only (e.g., in storage mirroring), then data consistency is maintained, but the technique to make RPPI writeable cannot be applied
Solution Approach 1:
The patent segments the file system into two independent parts: the original read-only file system (which maintains data consistency) and a new clone file system (which provides write capability). This segmentation allows the solution to work with read-only file systems by creating a separate writeable environment rather than trying to modify the original.
Solution Approach 2:
The patent creates a clone file system that is a copy of the original read-only file system. This clone serves as a writeable replica that mirrors the structure and data of the original, allowing write operations without affecting the integrity of the source read-only file system.
3Reliability
If data blocks are duplicated to create an RPPI, then the RPPI is fully independent and writeable, but storage space consumption increases
Solution Approach 1:
The patent merges the clone file system with the original file system's data blocks through shared storage. Instead of duplicating all data blocks, the clone shares the underlying storage infrastructure with the original, reducing storage consumption while maintaining functional independence through the clone file system layer.
Data Source
AI summary
A writeable clone of a read-only dataset, e.g., an RPPI of a file system, a mirror file system, etc., is created to make the read-only dataset appear from a client's perspective to be modifiable/writeable, yet without actually modifying the read-only dataset itself. Any change attempted to be made on the read-only dataset is stored in the writeable clone. In one embodiment, the clone shares data blocks with the read-only dataset.