Postgres 18 just landed, bringing a host of goodies—from built-in UUIDv7, fancy “over explain” stats and composite index skip scans to the long-awaited asynchronous I/O support. In this deep-dive episode, Hussein unpacks how sync vs async calls work, revisits Postgres 17’s blocking I/O, then shows off new async modes: the trusty worker queue and the high-performance io_uring path (plus a fallback “sync” mode when you need it).
Along the way you’ll hear about the hurdles of nonblocking I/O inside Postgres, the roadmap for backend writers and direct I/O, and what still needs polishing. Whether you’re a DB nerd or just curious about squeezing every bit of speed from your storage, this show has the low-down on making your Postgres faster and happier.
Watch on YouTube
Top comments (0)