Zero Downtime Database Upgrade via Schema Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Zero downtime upgrades in database systems face challenges due to the inability to automatically map database types, leading to table conversions being done in read-only mode, which impacts production and limits the ability to handle complex conversions without downtime.
Innovation Solution
The solution involves selectively cloning tables and generating new target tables with updated software content, computing differences in structure and database types between source and target systems, and migrating data using these differences, allowing for independent or concurrent data migration during zero downtime upgrades.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If table conversions are done using ABAP tools during zero downtime upgrades, then data can be migrated from source to target tables, but the tables must be put to read-only mode which impacts production write accesses
Solution Approach 1:
The patent segments the database system into multiple schemas (source schema, target schema, and optionally intermediate schemas) to isolate conversion operations. By dividing the monolithic table conversion process into schema-level operations, write accesses can continue in the source schema while conversions proceed in the target schema, resolving the contradiction between complete data migration and production availability.
Solution Approach 2:
The patent introduces intermediate schemas or bridge tables as mediators between the source and target tables. These intermediaries allow data to flow through multiple stages during conversion, enabling write operations to continue in the source system while conversions progress through intermediate representations before finalizing in the target schema.
2Adaptability or versatility
If complex database type conversions are performed during zero downtime upgrades, then data structure updates can be achieved, but the conversion process cannot be completed without putting tables to read-only mode
Solution Approach 1:
The patent performs preliminary actions by creating target table structures and defining conversion mappings before actual data conversion begins. By pre-configuring the target schema structure, conversion rules, and intermediate representations, the system prepares everything needed for complex type conversions to proceed without interrupting source table accessibility, making the operation easier to manage.
3Reliability
If database triggers are used to replicate data from production to upgrade copy during zero downtime upgrades, then data synchronization can be maintained, but the upgrade cannot be revoked by restoring a backup
Solution Approach 1:
The patent creates complete copies of the production database schema and tables in a separate upgrade schema. By copying the entire schema structure, triggers, and data to an isolated upgrade environment, the system maintains accurate data synchronization for conversion purposes while preserving the ability to revoke the upgrade by simply dropping the copied schema, eliminating the complexity of backup restoration.
Data Source
AI summary
A zero downtime upgrade of a source system to a target system is initiated. The source system and the target system operate in a first database layer and a second database layer. Thereafter, differences in structure and/or database type are computed between a source table in the source system and a target table in the target system in each of the two different database layers. The data is then migrated, using the computing differences and in the second database layer, from the source table to the target table. The upgrade procedure is later terminated after completion of the migration of the data. Related apparatus, systems, techniques and articles are also described.