Replication Resynchronization via Transaction Log Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data replication systems face challenges in resynchronizing replicate databases with minimal interruption to primary databases, especially in production environments with continuous data generation, and require significant manual intervention.
Innovation Solution
A replication control system that allows for resynchronization of replicate databases without affecting primary database operations, using a 'resync mode' that manages transaction logs and purges replication queues, enabling database repopulation from a trusted source with minimal manual intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional resynchronization methods are used to repopulate replicate databases, then data consistency is restored, but primary database operations are interrupted and significant manual intervention is required
Solution Approach 1:
The patent segments the resynchronization process into distinct phases: identifying transactions since the last dump, purging old replication queue data, and applying only the necessary transaction log segment. This allows the replicate database to be resynchronized without requiring a complete shutdown or manual intervention, as the system automatically determines which transactions need to be applied.
Solution Approach 2:
The system performs preliminary actions by maintaining a record of the last successful dump timestamp and using this information to pre-determine which transactions need to be replicated. The replication queue is prepared in advance with the knowledge of what data needs to be synchronized, eliminating the need for manual intervention during the actual resynchronization process.
2Reliability
If complete database dumps are used for resynchronization, then data consistency is ensured, but the time and resources required increase significantly
Solution Approach 1:
The patent extracts only the necessary portion of transaction logs - specifically, transactions that occurred since the last successful dump - rather than processing complete database dumps. This extraction approach significantly reduces the time and computational resources required for resynchronization while maintaining data consistency, as only the changed data needs to be replicated.
Solution Approach 2:
The system changes the parameter of data transfer from complete database dumps to incremental transaction segments. By using the dump timestamp as a parameter to determine the scope of replication, the system transforms a resource-intensive full synchronization process into an efficient incremental update process, reducing both time and resource requirements.
3Reliability
If replication queues are not purged during resynchronization, then transaction history is preserved, but system resources are consumed and consistency is compromised
Solution Approach 1:
The patent implements a controlled discarding of old replication queue data by purging transactions that are older than the last successful dump. This selective discarding maintains data consistency by ensuring only relevant transactions are applied, while freeing up system resources that would otherwise be consumed by obsolete data. The process recovers system performance by removing unnecessary data while preserving the integrity of the replication process.
Data Source
AI summary
System, method, computer program product embodiments and combinations and sub-combinations thereof for replication resynchronization are provided. An embodiment includes providing a mode of operation in a replication control system for repopulating a replicated version of a primary database that does not affect an ability to continue activity in the primary database. Further included is managing the repopulating of the replicated version with attention to at least a resynchronization marker.