Clusterless Database Backup GUID Indexing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In relational database management systems with clusterless availability groups, existing backup systems face challenges in recognizing database replicas after a forced failover, leading to unnecessary full backups instead of incremental or differential backups, which increases backup time significantly.
Innovation Solution
Utilizing a globally unique identifier (GUID) assigned by the database system to index and track database backups, allowing backups to be linked across different nodes, enabling efficient incremental or differential backups even after node changes due to forced failovers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If backups are indexed against node names in clusterless availability groups, then backup systems can identify databases for backing up, but after forced failover the backup system cannot recognize that the database is the same one previously backed up, leading to unnecessary full backups
Solution Approach 1:
The patent introduces a backup-guided failover mechanism where the backup system acts as an intermediary between the failover process and backup indexing. During failover, the backup system receives notification and updates the backup chain information, ensuring that the backup index correctly associates the failed-over database with its previous backup history. This intermediary role of the backup system in coordinating failover events resolves the recognition problem.
Solution Approach 2:
The patent implements a feedback mechanism where the backup system receives feedback about failover events and adjusts its backup indexing accordingly. When a forced failover occurs, the backup system is notified and updates the backup chain information to reflect the new primary node. This feedback loop ensures that the backup system maintains accurate knowledge of which databases have been previously backed up, preventing unnecessary full backups.
2Reliability
If manual forced failovers are allowed in relational database management systems, then database availability can be maintained during node failures, but the backup system may conduct full backups instead of incremental backups, dramatically increasing backup time
Solution Approach 1:
The patent applies preliminary action by having the backup system prepare and maintain backup chain information before failover events occur. The backup system proactively tracks which databases have been backed up and their associated backup chains. When failover is detected, this pre-established knowledge allows the backup system to immediately recognize the situation and perform appropriate incremental backups rather than unnecessary full backups, thus maintaining both availability and efficiency.
Solution Approach 2:
The patent replaces the mechanical approach of relying solely on node name indexing with a more intelligent software-based backup chain tracking mechanism. Instead of using simple node name matching which fails during failover, the system uses a software-managed backup chain data structure that persists across failover events. This substitution of mechanical indexing with intelligent software tracking resolves the contradiction between availability and backup efficiency.
Data Source
AI summary
A determination is made that a backup of a database in an availability group provided by a relational database management system (RDBMS) should be performed. The availability group includes a node functioning as a primary node and hosting a primary replica of the database and one or more other nodes functioning as secondary nodes and hosting secondary replicas of the database. The availability group is a clusterless availability group in which the one or more other nodes functioning as secondary nodes are not available as automatic failover targets should the primary node become unavailable. A command is issued to a node in the availability group to obtain a globally unique identifier (GUID) of the availability group. The node is instructed to index a backup of the database against the GUID of the availability group.