Remotely Mounted File System Using Snapshot Stubs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for snapshot management in storage systems require large amounts of storage and processing resources, especially when recovering a portion of the data, as they involve reconstructing the entire snapshot tree from full and incremental snapshots, which is inefficient in terms of time and resources.
Innovation Solution
The implementation of a distributed storage system that uses a snapshot tree data structure, where snapshots are archived as serialized data, allowing for partial restoration of the snapshot tree using stub nodes and metadata, reducing the need to reconstruct the entire tree and minimizing storage and processing requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If full snapshot and incremental snapshots are stored to enable recovery, then data recovery capability is improved, but storage space requirement increases
Solution Approach 1:
The snapshot tree is segmented into individual nodes representing different data portions at different time points. Each node can be independently stored and retrieved, allowing selective recovery of only the needed portions rather than requiring complete snapshot reconstruction. This segmentation enables efficient storage by storing only necessary snapshot components.
Solution Approach 2:
Incremental snapshots are nested within the snapshot tree structure, where each incremental snapshot contains only the changes since the previous snapshot. The nested structure allows the system to store incremental data efficiently within the hierarchical tree, reducing redundant storage while maintaining full recovery capability.
2Reliability
If complete snapshot tree reconstruction is performed for data recovery, then data integrity is improved, but recovery time increases
Solution Approach 1:
The patent extracts and stores metadata about the snapshot tree structure separately from the actual data nodes. This metadata contains information about node relationships, paths, and hierarchical structure. During recovery, this extracted metadata enables direct navigation to needed nodes without reconstructing the entire tree, significantly reducing recovery time while maintaining data integrity.
Solution Approach 2:
The snapshot tree structure and metadata are prepared and organized in advance during normal operation. Node relationships, paths, and hierarchical information are pre-computed and stored. When recovery is needed, this preliminary organization allows immediate access to required nodes without performing complex reconstruction operations at recovery time.
3Reliability
If complete snapshot tree reconstruction is performed, then full data recovery is improved, but processing resources consumption increases
Solution Approach 1:
The system performs only the necessary partial action required for recovery by using the stored metadata to directly locate and retrieve specific nodes. Instead of performing excessive action by reconstructing the entire snapshot tree, the patent enables selective retrieval of only the nodes needed for the requested data, significantly reducing processing resource consumption while achieving complete recovery of the needed portions.
Data Source
AI summary
A request to access one or more files associated with a particular pint in time that are archived at a cloud storage is received. A file system metadata tree data structure based on metadata associated with the one or more requested files that is obtained from the cloud storage is partially restored at a storage system. The partially restored file system metadata tree data structure includes one or more stub nodes associated with one or more non-requested files. Access to the one or more requested files is provided using the partially restored file system metadata tree data structure. Metadata associated with the one or more non-requested files is prefetch from the cloud storage. A portion of the file system metadata tree data structure corresponding to the one or more non-requested files is restored in part by replacing the one or more stub nodes associated with the one or more non-requested files with metadata corresponding to the one or more non-requested files.