Namespace Salvage in Logging Filesystems via B+ Tree Key Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current filesystem recovery methods are inadequate for restoring namespace metadata after cyber-attacks, particularly in deduplicating filesystems using B+ Trees, where corruption of critical elements renders large amounts of data inaccessible.
Innovation Solution
The implementation of physical key salvage and additive combining techniques to recover namespace metadata, involving sequential scanning of damaged B+ Trees for embedded file metadata keys and combining historical records to construct a new tree structure, allowing for the recovery of lost data even when snapshots are corrupted.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional filesystem recovery methods are used, then the filesystem structure is preserved, but namespace metadata cannot be recovered after corruption
Solution Approach 1:
The recovery process segments the B+ Tree into individual pages that can be independently scanned and processed. Each page is examined separately for salvageable keys, allowing the system to recover metadata without requiring the entire tree structure to be intact. This segmentation enables recovery even when parts of the filesystem are corrupted.
Solution Approach 2:
The patent extracts keys from the corrupted B+ Tree pages by scanning through the raw page data and identifying valid key structures. These extracted keys are then used to reconstruct the namespace metadata, separating the recovery process from the corrupted structural elements and focusing only on salvageable information.
2Loss of information
If snapshots are retained for recovery purposes, then data can be restored, but storage space is consumed and snapshots may themselves be corrupted
Solution Approach 1:
The system performs self-service recovery by scanning its own corrupted B+ Tree pages and extracting salvageable keys directly from the existing storage structure. This eliminates the need for external backup snapshots, allowing the filesystem to recover from its own corrupted data without requiring additional storage space for separate backup copies.
Solution Approach 2:
The recovery process changes the parameter of data accessibility by transforming corrupted B+ Tree pages into a usable namespace structure through key extraction. This parameter change allows data to be recovered in a different form than the original structure, bypassing the need for identical snapshot copies.
3Speed
If B+ Tree structure is maintained for efficient file access, then file operations are fast, but corruption of critical elements renders large amounts of data inaccessible
Solution Approach 1:
By segmenting the B+ Tree into individual scanable pages, the system isolates corruption to specific pages rather than the entire structure. This allows efficient scanning of uncorrupted pages while recovering keys, maintaining fast recovery operations without requiring the complete tree structure to be intact.
Solution Approach 2:
The patent converts the harmful effect of corruption into a benefit by scanning corrupted pages for salvageable keys. Rather than treating corruption as a complete failure, the system identifies and extracts valid key information from corrupted pages, transforming the corrupted data into recoverable namespace metadata.
Data Source
AI summary
Recovering a namespace held in a tree structure that has been corrupted includes scanning a linear representation of pages belonging to the tree structure to copy out keys from uncorrupted leaf pages into a new tree structure. The scan is repeated over snapshots that have been taken of the tree to find keys that are missing from the new tree structure. Upon finding a missing key in an uncorrupted leaf page of a snapshot, the missing key is copied out into the new tree structure unless the missing key has already been found and copied out from an uncorrupted leaf page of a more recent snapshot taken of the tree structure.