RDF/Star Recovery Leg RCVY Mode for Faster Replica Resynchronization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data storage systems face challenges in efficiently managing replication and recovery processes across multiple data centers, particularly in asynchronous and synchronous replication configurations, leading to prolonged recovery times and potential disruptions during failures.
Innovation Solution
Implementing a data storage system with RDF/Star and RDF/Star-A configurations that utilize RDF/A RCVY mode and MSC RCVY mode, where group-level personalities are set to R1 on both sides of the recovery leg, capture cycles are performed, and transmit cycles are discarded, enabling accelerated recovery by maintaining continuous replication readiness and reducing the need for active data transfer during normal operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional replication recovery processes are used in RDF/Star configurations, then data consistency is maintained, but recovery time is prolonged causing significant disruption to host applications
Solution Approach 1:
The system performs preliminary actions by setting personalities of corresponding replicas to R1 on both sides of the recovery leg at activation time, and pre-configuring capture cycles and transmit cycles to be discarded. This preparation in advance enables accelerated recovery without compromising data consistency, as the replication infrastructure is pre-positioned for rapid failover.
Solution Approach 2:
The invention uses replica copying mechanisms where secondary replicas are maintained with R1 personalities, creating ready-to-use copies that can immediately take over primary functions. The capture cycles and transmit cycles are pre-established on both sides of the recovery leg, allowing instantaneous replication without traditional recovery delays.
2Reliability
If active data transfer is performed during normal operations in RDF/Star configurations, then data synchronization is maintained, but system complexity and resource consumption increase
Solution Approach 1:
The invention extracts the active data transfer function from normal operations by configuring transmit cycles to be discarded on both sides of the recovery leg. This separates the replication readiness function from active data transmission, simplifying normal operations while maintaining synchronization capability through capture cycles alone.
Solution Approach 2:
The recovery leg is configured with multi-functionality, serving both as a standby recovery path and an active replication path. By setting R1 personalities on both sides and configuring capture cycles, the same infrastructure handles both normal data synchronization and failure recovery scenarios, reducing overall system complexity.
Data Source
AI summary
A recovery leg in a Star or Star-A configured data replication system is maintained in a pseudo-active RCVY mode in which personalities of corresponding replicas of a mirror are set to R1 on both sides of the recovery leg, capture cycles are performed, and transmit cycles are discarded. During a recovery operation, tracks are marked for resynchronization in the proper direction and resynchronization is performed if the source side data is not the most recent. Group-level personalities of the replica on the target side of the recovery leg are then set to R2 (the target side multi-session consistency software instance converts the capture cycles to restore cycles, converts the transmit cycles to receive cycles, and swaps the cycle tags between the receive and restore cycles). The RCVY mode and multi-session consistency software instances accelerate transition from RCVY mode to normal operation on the recovery leg.