Speak at Bengaluru Systems meet-up

What really happens inside PostgreSQL from the moment the postmaster starts to the moment a crash is recovered and normal operations resume?
This talk takes you through a connected journey across several key components of PostgreSQL internals, starting with the process architecture: how the postmaster starts, spawns child processes, and manages system activity. From there, we dive into the mechanisms of physical replication, and how write-ahead logs (WAL) are generated, sent, and applied on replicas to maintain consistency.
Next, we explore crash recovery how PostgreSQL replays WAL to recover the database to a consistent state, including the roles of checkpoints, full-page writes. Finally, we dig into the buffer manager, detailing how shared buffers are managed, flushed, and synced with WAL.
This talk is designed for future contributors/hackers, developers, and DBAs who want to understand PostgreSQL’s inner workings across its full lifecycle — from startup to crash recovery. You’ll leave with a clear mental model of how these pieces connect to make PostgreSQL reliable, performant, and resilient.