Google Cloud Expands Its Contributions to PostgreSQL with Improvements in Replication and Upgrade Processes

Google Cloud presented details about its latest technical developments and upstream contributions to PostgreSQL, aimed at improving logical replication, upgrade processes, and the overall stability of the platform.

The information covers engineering work carried out between July and December 2025, with the main focus on expanding logical replication capabilities and preparing for active-active replication configurations.

Improvements in logical replication and active-active architectures

One of the key updates is the implementation of automatic conflict detection, which allows PostgreSQL to detect row-level conflicts during replication processes without manual intervention.

This addresses a long-standing issue in multi-node write environments, where conflicting changes between different nodes could block replication processes.

The evolution toward active-active replication architectures has also sparked discussions in the PostgreSQL community regarding consistency models and conflict management.

Franck Pachot commented that comparing bidirectional logical replication with Oracle RAC or distributed SQL solutions such as CockroachDB and YugabyteDB can lead to misunderstandings of database consistency models, as replication systems often use a “last write wins” approach, while distributed SQL platforms rely on ACID consistency.

On the other hand, Janardhan Korapala described the development as an important signal for positioning PostgreSQL as an enterprise standard, especially as hyperscaler companies begin integrating enterprise-grade replication capabilities directly into the upstream project.

Expanded replication functionality

Google Cloud has also contributed improvements that extend logical replication beyond standard table data and add replication support for sequences.

This reduces the need for manual synchronization during migration and version upgrade processes.

The company has also developed fixes for issues such as self-deadlocks in subscription management mechanisms, which can occur when replication commands attempt to access already locked resources on the same server.

Faster and more reliable upgrade processes

A significant part of the development is focused on improving the reliability and performance of upgrade processes through:

• optimizations in pg_upgrade
• better handling of large objects
• reduced upgrade time for large environments
• preservation of required WAL data during upgrades
• proper storage of schema constraints

These improvements are aimed at more stable and predictable migration processes in large-scale PostgreSQL deployment environments.

Additional improvements and bug fixes

In addition to replication and upgrade optimizations, Google Cloud has contributed a number of stability fixes to PostgreSQL, including:

• handling of invalid index pages in diagnostic tools
• fixes for loading extensions from nested paths
• improvements in WAL flush logic for better durability protection in edge cases

Work on future PostgreSQL features
The company stated that it continues to work on future PostgreSQL improvements, including:

• structured conflict logging for replication environments
• improvements in pg_dump parallel export capabilities
• optimizations for processing large-scale datasets

With these developments, Google Cloud continues to expand its participation in the PostgreSQL ecosystem and support the evolution of the platform as an enterprise-ready database for large-scale cloud and hybrid infrastructures.