Active-Active Dual-Node Storage Fail-Back with Segmented Large Writes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Active-active dual-node storage systems experience storage accessibility hiccups due to the complexity and need for managing large write operations during recovery from forced reboots, crashes, or disasters.
Innovation Solution
A specialized recovery protocol is executed by a surviving or rebooted node, treating each large write request as multiple small write requests, draining dedicated sub-ubers, and resuming normal treatment after completion, thereby avoiding the need to manage large writes and their corresponding sub-uber information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If large write operations are managed during recovery in active-active dual-node storage systems, then data integrity is maintained, but storage accessibility hiccups increase and recovery time extends
Solution Approach 1:
The patent segments large write operations into smaller write operations during recovery. The recovery node divides a large write request into multiple smaller writes that can be processed independently and in parallel, reducing the overall recovery time while maintaining data integrity. This segmentation allows the system to handle large writes without creating storage accessibility hiccups.
Solution Approach 2:
The patent implements preliminary actions by pre-processing and pre-positioning data during normal operation. Before a failure occurs, data is pre-processed and staged in a manner that facilitates rapid recovery. When recovery is needed, the system can quickly resume operations without having to perform extensive data processing from scratch, thereby reducing recovery time while ensuring data integrity.
2Reliability
If large write operations are processed during recovery, then complete data restoration is achieved, but storage accessibility performance deteriorates
Solution Approach 1:
The patent segments large write operations into smaller write operations during recovery. By dividing a large write into multiple smaller writes, the system can process them in parallel or sequentially without blocking storage accessibility for extended periods. This maintains storage performance while ensuring complete data restoration.
Solution Approach 2:
The patent dynamically adjusts the write operation strategy during recovery based on system state and performance conditions. The recovery node can switch between processing large writes as-is, segmenting them, or prioritizing smaller writes based on real-time storage accessibility performance, thereby maintaining optimal performance while achieving complete restoration.
3Stability of the object's composition
If sub-uber information is managed during large writes in recovery, then storage consistency is maintained, but operational complexity increases
Solution Approach 1:
The patent extracts and isolates sub-uber information management from the main recovery process. By separating the management of sub-uber metadata from the actual data recovery operations, the system maintains storage consistency without adding significant operational complexity to the core recovery workflow. The sub-uber information is handled as a separate, manageable component.
Solution Approach 2:
The patent introduces an intermediary mechanism or data structure that mediates between the recovery operations and sub-uber information management. This intermediary layer handles the complexity of tracking and consistent management of sub-uber metadata during recovery, shielding the main recovery logic from complexity while ensuring storage consistency is maintained.
Data Source
AI summary
Techniques for limiting or reducing storage accessibility “hiccups” in active-active clustered systems that perform large writes. The techniques can include executing, by a surviving, failover, or rebooted node of an active-active clustered system, a specialized recovery protocol that includes treating each large write request from a host computer as a plurality of small write requests while execution of the specialized recovery protocol is in progress, draining all dedicated sub-ubers for a primary and secondary node of the active-active clustered system, and, having completed execution of the specialized recovery protocol, resuming normal treatment of large write requests from the host computer. In this way, the need and complexity of managing large writes and maintaining their corresponding sub-uber information during recovery from a forced reboot, crash, or disaster involving the primary or secondary node can be avoided, and storage accessibility hiccups due to performing the large writes can be limited or reduced.