Garbage Collection Container Delete-Ready State
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In data deduplication systems, existing garbage collection methods often inadvertently delete relevant data due to arbitrary or Last-In First-Out (LIFO) sequencing, leading to loss of valuable information.
Innovation Solution
Introducing a 'delete-ready' state for containers, where unreferenced data is preserved and can be recovered, allowing for a First-In First-Out (FIFO) or near-FIFO sequencing of container deletion, ensuring that accidentally deleted data can be restored and disk space is reclaimed efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If arbitrary or LIFO sequencing is used for garbage collection, then disk space is reclaimed quickly, but valuable data is accidentally deleted
Solution Approach 1:
The system performs preliminary actions by maintaining metadata records and references to data segments before garbage collection occurs. This allows the system to identify and preserve data that should not be deleted, while still enabling rapid reclamation of space for data that is safe to remove.
Solution Approach 2:
The system uses feedback mechanisms through metadata tracking and reference counting to monitor data usage patterns. This feedback allows the garbage collection process to adaptively determine which data can be safely reclaimed and which must be preserved, resolving the contradiction between speed and data safety.
2Loss of information
If FIFO sequencing is used for container deletion, then data recovery is improved, but disk space reclamation efficiency decreases
Solution Approach 1:
The system applies different quality characteristics to different data segments within the same storage system. Critical data with active references receives protection and prioritized preservation, while non-critical data without references is marked for rapid reclamation. This local differentiation resolves the contradiction between data recovery capability and reclamation efficiency.
3Quantity of substance
If all unreferenced data is deleted immediately, then storage efficiency is maximized, but accidentally deleted data cannot be recovered
Solution Approach 1:
The system implements beforehand cushioning by maintaining metadata records and reference information even after data segments are deleted during garbage collection. This cushioning layer acts as a safety buffer that enables recovery of accidentally deleted data, while still allowing the system to maximize storage efficiency by removing unreferenced data.
Data Source
AI summary
An indication is received that a data object is to be deleted, wherein the data object comprises data stored in a segment within a container. It is determined no currently alive data object references any segment within the container. The container is placed in a delete-ready but not yet reclaimable state.