Selective File System Journaling for Data Renaming Operations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing file systems face data loss and corruption during renaming operations due to failures like spontaneous power loss or unexpected storage device disconnections, especially when non-committed transactions are involved, leading to potential duplication or loss of data objects.
Innovation Solution
The proposed solution involves selectively journaling renaming operations based on whether they can be performed atomically, creating a journal entry only if the operation is non-atomic, and updating it upon completion to mitigate potential data loss and corruption, allowing for replay or undo of non-committed transactions upon volume mounting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If journaling is performed for all renaming operations, then reliability is improved, but device complexity and performance deteriorate due to excessive write amplification
Solution Approach 1:
The patent applies local quality by differentiating journaling requirements based on the atomicity of renaming operations. Atomic operations (single-sector writes) are excluded from journaling, while non-atomic operations (multi-sector writes) are journaled. This selective approach reduces overall journaling overhead while maintaining reliability for operations that need it most.
Solution Approach 2:
The patent implements partial action by journaling only a subset of renaming operations—specifically those that are non-atomic and therefore at risk of corruption. This avoids the excessive action of journaling all renaming operations, reducing write amplification while still providing protection where needed.
2Reliability
If journaling is performed for non-atomic renaming operations, then data loss is reduced, but system performance deteriorates due to additional write operations
Solution Approach 1:
The patent applies local quality by targeting journaling specifically at non-atomic renaming operations that span multiple sectors. These operations are identified as having higher risk of data corruption, so they receive the protective journaling treatment while atomic operations proceed without journaling overhead, maintaining performance for low-risk operations.
3Loss of energy
If atomic renaming operations are excluded from journaling, then write amplification is reduced, but reliability may deteriorate if atomic operations fail
Solution Approach 1:
The patent applies local quality by recognizing that atomic renaming operations have inherently different risk profiles compared to non-atomic operations. Atomic operations, being confined to single-sector writes, are considered sufficiently reliable and are excluded from journaling to reduce write amplification, while non-atomic operations receive journaling protection.
Data Source
AI summary
This disclosure describes techniques performing data object renaming operations. This disclosure describes techniques for performing data object renaming operations in a manner that may reduce file system corruption and data loss. The techniques described herein may include selectively journaling particular types of renaming operations. Selectively journaling particular types of renaming operations may mitigate potential data loss and file system corruption.