Dave Stokes Talks Open – Stoking the Flames of Open Source – Episode 2
summary
A conversational roundup that gives concrete takeaways: Postgres ships 170–190 features per yearly release and deliberately favors simple, maintainable designs over micro‑optimizations; a legacy macro trick ran for years until inline functions landed; JSONB compression caused a months‑long delay because changing on‑disk formats is permanent; monitoring with Percona PMM (MySQL MongoDB Postgres ProxySQL) surfaces bad queries; 32‑bit transaction ID can wrap in days if unmanaged; reordering columns once saves ~20% disk and reduces replication and memory pressure.
AI-written summary · about
chapters (8)
- 0:00Intro
- 0:55How did you get here
- 4:53The burning building
- 6:58The macro
- 8:35The unknown unknowns
- 13:23Longevity
- 14:54Delays
- 17:27Interview with Charlie
related talks
16mDave Stokes Talks Open - Episode 4
18mDave Stokes Talks Open - Episode 6
15mDave Stokes Talks Open - Episode 14
10mDave Stokes Talks Open - Episode 7
07mDave Stokes Talks Open - Episode 12
10mDave Stokes Talks Open - Episode 15
10mDave Stokes Talks Open - Episode 3
06mDave Stokes Talks Open - Episode 11