What is Patroni, really?
summary
Patroni uses an external distributed key‑value store (etcd, Consul, ZooKeeper, Kubernetes control‑plane) to serialize leader elections via atomic compare‑and‑swap and TTLs; a node that fails to refresh the leader key with TTL and CAS demotes itself, avoiding split‑brain. It writes cluster config, member connection URLs, WAL positions and implements retain_slots, quorum failover, synchronous_standby_names and a DCS failsafe so clusters can keep writing when the control store blinks. Cascading standby clusters support cross‑region replicas.
AI-written summary · about
chapters (6)
- 0:00Introduction & agenda
- 1:14What is Patroni?
- 2:11Automatic failover done wrong
- 9:55Patroni: how it works
- 18:51Notable features
- 21:16Additional resources
related talks
50mPatroni Deployment Patterns
52mPatroni 3.0: What's New and Future Plans
49mPatroni: what the blog posts don't tell you...
41mEnsuring High Availability in Self-Managed PostgreSQL with Patroni
54mContainerized Clustered PostgreSQL
33mCitus & Patroni: The Key to Scalable and Fault-Tolerant PostgreSQL
41mProtecting your data with Patroni and pgBackRest
34mElephants on Automatic: HA Clustered PostgreSQL with Helm [I]