Spock PostgreSQL Multi-Master (Multi-Active) Replication from pgEdge
Postgres Conference Silicon Valley (April 20 and 21, 2023) is now behind us. It was an exciting time with friends, fellow PostgreSQL community members, and competitors.
I gave a presentation at the event about a new Postgres extension developed by pgEdge named Spock (part of the Community Licensed pgEdge Platform). Spock is an asynchronous multi-master (multi-active) replication system for PostgreSQL based on logical decoding. Spock started out as a fork of pgLogical, but we've added several important features and enhancements.
Optimal resource usage, with read-write tables on standby nodes.
Rapid replacement of lost nodes without data loss.
Cross-version replication that supports fast blue-green upgrades.
Selective replication of specific tables (with support for partitioning).
Distributed content management that tracks content location. This is important for supporting personal identifying information (pii) regulations related to the country of data residence.
... and more.
Over 20 years ago I created the Slony-I replication system for PostgreSQL. Slony has long-since been replaced with built-in replication (streaming and logical), but users are still not happy with the alternatives. PostgreSQL logical replication is missing features and functionality, pgLogical is now discontinued, and proprietary solutions are leaving many users unsupported and dissatisfied.
At pgEdge we are changing that.
While Spock is not yet ready for production, we encourage you to evaluate the current progress and chime in with feedback through GitHub or other channels. If you couldn't join us in San Jose or missed my talk, check out the slides or the video recording below.