Storage Failback Using Snapshot Sync Replication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current remote mirroring solutions for storage systems require a lengthy full copy process during failback, which is resource-intensive and time-consuming, especially over limited network connections, leading to potential data loss and service unavailability.
Innovation Solution
The method involves creating and synchronizing snapshots between primary and secondary devices, allowing for quick restoration and synchronization of changes, thereby eliminating the need for a full copy during failback, using Snapshot Sync Replication to identify a common base and transfer only the differences in data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a full copy process is used during failback, then data consistency is ensured, but failback time becomes excessively long and network resources are overwhelmed
Solution Approach 1:
Snapshots are created and synchronized in advance during normal operation, storing historical data states before potential failures occur. This preliminary action enables rapid restoration without requiring full data copying during failback, as the pre-synchronized snapshots provide a ready-to-use baseline for quick recovery
Solution Approach 2:
The data synchronization process is segmented into snapshot-based restoration and incremental change synchronization. Instead of copying all data at once, the system divides the process into restoring from pre-created snapshots and then applying only the necessary incremental changes, dramatically reducing failback time while maintaining data consistency
2Reliability
If a full copy process is used during failback, then complete data restoration is achieved, but network bandwidth is excessively consumed
Solution Approach 1:
The system extracts only the essential baseline data from pre-synchronized snapshots stored locally, rather than transferring complete data sets over the network. By taking out only the necessary snapshot information and applying incremental changes locally, network bandwidth consumption is dramatically reduced while achieving complete data restoration
Solution Approach 2:
Data snapshots are preliminarily prepared and stored during normal operation, creating a local copy of historical data states. This preliminary action eliminates the need for extensive network data transfer during failback, as the restoration process can proceed using pre-synchronized local snapshots combined with incremental change applications
3Reliability
If traditional failback methods are used, then data is restored, but service downtime is extended
Solution Approach 1:
Snapshots are preliminarily created and synchronized during normal system operation, preparing restoration data in advance. When failback is needed, the system can immediately restore from these pre-prepared snapshots without waiting for data collection or initial synchronization, significantly reducing service downtime while ensuring complete data restoration
Solution Approach 2:
The failback process is segmented into rapid snapshot restoration followed by incremental change synchronization. This segmentation allows the system to quickly restore service using pre-synchronized snapshots first, then gradually apply incremental changes in the background, minimizing service interruption while ensuring complete data restoration
Data Source
AI summary
Techniques for managing a storage system involve at a second device: in response to receiving a command to failback a first device, determining a snapshot synchronized between the second device and the first device before a failure of the first device, wherein the first device has been failed over to the second device; causing the first device to be restored based on the snapshot; determining a change in latest data of the second device relative to the snapshot; and synchronizing the change to the first device.