Parallel Lock Remastering During Database Reconfiguration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Database reconfiguration causes downtime due to the need to freeze access and scan every lock, which is time-consuming and affects database availability, especially in high-availability scenarios like online banking services.
Innovation Solution
Implement a method for continuous database locking during reconfiguration using event-based and request-based lock state remastering processes, allowing resources to be made available before the remastering process is fully completed, with non-blocking request-based support for resources previously or to be mastered by added or removed nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If database reconfiguration is implemented by freezing access and scanning every lock, then lock state remastering can be completed, but database downtime increases and availability decreases
Solution Approach 1:
The patent segments the lock scanning process into two distinct approaches: event-based lock state remastering that scans all locks systematically, and request-based lock state remastering that processes only specific locks on-demand. This segmentation allows the system to balance between thorough remastering and minimizing downtime by allowing selective access during reconfiguration.
Solution Approach 2:
The patent implements preliminary action by allowing request-based lock state remastering to process specific locks before the complete event-based remastering is finished. This enables resources to be made available earlier than traditional approaches would allow, reducing overall reconfiguration time while maintaining data integrity.
2Manufacturing precision
If all locks are scanned during reconfiguration, then complete lock state remastering is achieved, but processing time increases significantly
Solution Approach 1:
The patent applies partial action by allowing the request-based remastering process to handle specific lock requests without waiting for complete event-based remastering of all locks. This partial processing approach maintains sufficient lock state accuracy for accessed resources while significantly improving reconfiguration speed by avoiding unnecessary scanning of all locks.
Solution Approach 2:
The patent introduces dynamics by allowing the remastering approach to adapt based on request patterns. The system can dynamically switch between event-based comprehensive scanning and request-based selective processing, optimizing the balance between accuracy and speed based on actual workload requirements during reconfiguration.
Data Source
AI summary
A computer program product, system, and computer implemented method for continuous database locking during database reconfiguration is provided herein. The present approach provides two different processing approaches to remaster locks that can execute in parallel. The first is an event-based lock state remastering process that executes a processing flow to ensure that all locks that need to be remastered are in fact remastered. As second is a request-based lock state remastering process that executes processing solely for the requested resource in order to quickly make the requested resource accessible. Additionally, each process is responsive to the other in the request-based lock state remastering process can continue from where the event-based lock state remastering process left the corresponding resource and the event-based lock state remastering process avoids further processing for locks that are, or have been, processed using the request-based lock state remastering process.