Interview with Fujitsu
Any views or opinions represented or expressed in this interview belong solely to the interviewee and do not neccessarily represent those of the PGConf.EU 2024 organization, PostgreSQL Europe, or the wider PostgreSQL community, unless explicitly stated.
- In which areas do you expect PostgreSQL to grow most and how does your company contribute to and benefit from that growth?
-
In today's dynamic market landscape and the increasing amount of cloud-hosted data, the fundamental requirements for a DBMS are centered around high availability, seamless scalability, and minimal downtime. Recognizing this critical need, we anticipate Postgres's continued growth in the future, driven by advancements in logical replication to facilitate obstacle-free active-active replication for distributed systems.
Our company has dedicated significant efforts to enriching this essential layer and continues to pursue further improvements. Through these endeavors, we enhance user experiences and empower organizations to effectively address the ever-evolving demands of modern data management with unmatched efficiency and reliability.
- Which of your company's contributions to the PostgreSQL Project (code/community/conference) are you most proud of?
-
As a dedicated contributor to the Postgres community, our company is deeply engaged across its ecosystem. Over past year, we've undertaken many significant projects within Logical Replication layer, introducing four key features:
- Enabled synchronization of 'logical replication slots' between primary and replica nodes for seamless failover process.
- Made possible the smooth subscription resumption after node upgrade.
- nabled publisher node upgrade while safeguarding logical replication slots and subscriber data.
- Enhanced subscriber node performance through parallelizing the apply work for large transactions.
Beyond coding, we actively participate in community discussions, providing valuable feedback. We also oversee release management, Commitfest coordination, and global conference organization, ensuring Postgres's smooth operation & development worldwide. Moreover, our team members are esteemed speakers at prestigious Postgres conferences globally, including PGConf.EU, PGConf.NYC, PGCon (Canada).
n summary, we are proud of our significant contributions to Postgres and our representation on the global stage, reaffirming our dedication to the PostgreSQL community's advancement.
- Which PostgreSQL extension do you benefit from most, and why?
-
In our project debugging, testing and review processes, we heavily rely on two key tools: pg_walinspect and pageinspect.
pg_walinspect serves as a valuable asset in various project tasks as it offers a range of built-in functions that enable us to inspect the contents of Write-Ahead Logs (WAL), providing invaluable insights and facilitating efficient problem-solving and performance optimization efforts.
Pageinspect module assists us in conducting thorough critical code change reviews and debugging processes. It offers essential functions for inspecting the contents of database pages at a low level.
- What feature is missing in PostgreSQL and how would it help if added?
-
PostgreSQL has limits on handling conflicts in active-active replication setups. While Logical replication has unlocked numerous opportunities for distributed, highly available systems with finer replica control, the lack of conflict resolution mechanisms poses a significant challenge.
- What is the most annoying PostgreSQL problem, and do you have plans or ideas to fix it?
-
The absence of DDL replication in PostgreSQL is a significant limitation within logical replication. This gap results in manual intervention being required for DDL changes, leading to a cumbersome and error-prone replication process. Integrating DDL replication functionality would streamline operations by automatically replicating schema changes to subscriber nodes, enhancing efficiency and reducing the risk of human error in system maintenance and management.
To address this problem, our team has worked on designing and implementing the support for logical replication of DDLs. We have successfully developed a PoC patch for the feature, laying the groundwork for its integration into PostgreSQL.