Expeditious data recovery in Not Only SQL (NoSQL) databases using a log-based approach

Not Only SQL (NoSQL) databases lack standardized backup and recovery procedures, making fault tolerance and rapid restoration more challenging than in traditional relational systems. Achieving efficient recovery in distributed NoSQL environments requires mechanisms that can manage operational failures while preserving availability, consistency, and performance. This study investigates existing recovery strategies for NoSQL databases and introduces a novel log-based expeditious recovery mechanism tailored for distributed architectures. Our primary contribution is a transaction evaluator and cleaner system that performs real-time inspection of operational logs, identifies failed or inconsistent transactions, and prevents their propagation from the primary node to secondary replica nodes. Unlike conventional log-shipping and replication approaches, the proposed method integrates selective log forwarding with automated transaction sanitization, thereby reducing replica divergence, data loss, and overall recovery time. We implemented the system using Node.js and MongoDB Atlas, incorporating modules for log management, collection operations, and a lightweight journaling process for secondary nodes. Experimental evaluation shows that our approach improves recovery speed and reliability compared to standard replication-based techniques. These findings demonstrate a practical and robust solution for accelerating recovery in NoSQL databases, contributing meaningful advancement to failure-recovery research in distributed data systems.