Database Technology

MariaDB is ending support for its MySQL build in September while shifting focus to a separate premium replication technology. This move has reignited concerns regarding the long-term open-source status of the Galera cluster replication engine. The company is actively developing new proprietary replication capabilities as part of this strategic pivot.

• MySQL build support terminates in September; plan migration or extended support contracts immediately.
• Galera's open-source status is under question as MariaDB prioritizes premium replication tech.
• New proprietary replication features are in development, potentially creating vendor lock-in risks.
• Review current Galera deployments for compliance with the new licensing and support model.

The hba_file GUC specifies the path to the pg_hba.conf authentication rules rather than containing the rules inline. Changes within the referenced file take effect immediately upon a configuration reload, avoiding downtime. However, if you move or replace the hba_file itself, a full server restart is required for the change to apply.

• hba_file is a pointer to pg_hba.conf, not an inline config block.
• Editing the target file allows instant auth changes via reload.
• Moving or replacing the file itself mandates a server restart.
• Verify file paths carefully to avoid restart requirements.

PostgreSQL 19 addresses the data lineage problem by providing native capabilities to track how data moves through transformation steps. This feature allows engineers to trace values back to their source tables, views, and ETL scripts without relying on external tools or fragmented documentation. The update aims to resolve the common operational pain point of verifying data origins during financial or analytical audits.

• PostgreSQL 19 adds native lineage tracking to trace data origins across views and transformations
• Reduces manual effort in debugging data discrepancies by automating traceability
• Eliminates dependency on external lineage tools or fragmented ETL documentation
• Critical for validating financial metrics and ensuring audit compliance

Radim Marek argues that PostgreSQL's Multi-Version Concurrency Control is often mischaracterized as a 40-year-old mistake, when it is actually a deliberate architectural decision. He demonstrates that common pain points like table bloat, write amplification, and vacuum overhead are inherent trade-offs of this design, reproducible on a live PostgreSQL 19 beta. The piece contextualizes criticisms from entities like Uber and Andy Pavlo’s group as symptoms of this choice rather than bugs.

• MVCC bloat and vacuum overhead are intentional trade-offs, not software defects.
• PostgreSQL 19 beta reproduces classic MVCC behaviors like dead tuple accumulation.
• Write amplification metrics cited by Uber and others stem from core design.
• Treating MVCC traits as bugs leads to incorrect operational expectations.
• Understanding the 'why' behind MVCC helps plan for vacuum and bloat management.

AWS now supports a serverless pipeline that irreversibly redacts over 30 types of personally identifiable information from Amazon RDS for PostgreSQL audit logs. The system removes sensitive data such as Social Security numbers, credit cards, and names before storing the clean logs in Amazon S3. These redacted logs remain queryable via Amazon Athena for ongoing analysis and compliance needs.

• Automated redaction handles 30+ PII types including SSNs and credit cards before storage
• Serverless pipeline writes irreversibly redacted logs to Amazon S3 for cost-effective archiving
• Redacted audit logs remain fully queryable through Amazon Athena for compliance auditing
• Simplifies data privacy compliance by removing sensitive fields at the ingestion layer