Dynamic Path Priority for Scalable Storage Failure Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing access control systems for multiple host computers fail to recover from failures in a scalable manner, as they are not designed to handle multiple hosts effectively during data mirroring processes.
Innovation Solution
The system employs a dual-host and dual-storage configuration with prioritized paths for data mirroring, where the first storage system is the source and the second is the target during normal conditions, and reverses roles during failures, allowing write requests to be rerouted through lower-priority paths to ensure recovery without overloading a single host.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single host computer is used for data mirroring operations, then the complexity of the system is reduced, but the scalability of failure recovery is limited
Solution Approach 1:
The system segments the failure recovery function by introducing multiple host computers (first host and second host) that can independently perform data mirroring operations. This segmentation allows the system to handle failures at different locations simultaneously, improving scalability without significantly increasing overall system complexity.
Solution Approach 2:
Each host computer is designed with multi-functionality, capable of performing both primary data writing and backup data mirroring operations. The first host can write to the first storage apparatus and mirror to the second storage apparatus, while the second host can similarly perform these operations, creating a universal system that adapts to various failure scenarios.
2Ease of operation
If path priority is fixed during normal operations, then system operation is simplified, but the ability to handle failures dynamically is reduced
Solution Approach 1:
The path priority is designed to be dynamic rather than fixed. During normal operations, the first path has higher priority for simplified operations. However, upon detecting a failure in the first storage apparatus, the system automatically switches priorities so that the second path becomes higher priority, enabling dynamic adaptation to failure conditions while maintaining operational simplicity.
Solution Approach 2:
The system incorporates feedback mechanisms where the host computers continuously monitor the status of storage apparatuses and paths. When a failure is detected, this feedback triggers an automatic reconfiguration of path priorities, ensuring the system responds appropriately to changing conditions while maintaining ease of operation through automated control.
3Reliability
If data mirroring continues during failure recovery, then data consistency is maintained, but the load on storage systems increases
Solution Approach 1:
The data mirroring operation is structured as periodic rather than continuous during failure recovery. The system alternates between performing mirroring operations and pausing to reduce load, while still maintaining data consistency through controlled periodic synchronization. This periodic approach prevents overwhelming the storage systems during critical recovery periods.
Solution Approach 2:
During failure recovery, the system performs partial data mirroring rather than complete continuous mirroring. Only critical data consistency requirements are fulfilled, allowing the system to maintain essential data integrity while reducing the overall load on storage systems during the recovery process.
Data Source
AI summary
A first SS (storage system) provides a first SA (storage area), a second SS provides a second SA, and first and second HSs (host systems) are coupled to the first and second SSs. First and second paths connect the first HS to the first and second SAs. Third and fourth paths connect the second HS to the first and second SAs. When the first path has a higher priority than the second path after recovery where a WR (write request) is permitted to be transmitted to the first SS, if the first SA which is a write destination for the received WR is a target in a data mirroring process, the first SS transmits a notification to stop the process. The second SS performs control such that a WR specifying the second SA as a write destination fails, and then the second HS transmits the WR using the third path.