PostgreSQL on PetSet
summary
Deploys PostgreSQL on Kubernetes using PetSet/StatefulSet, Patroni for automatic primary election and etcd as the distributed config store, with a live demo on a portable micro‑cluster showing failover and replica join working. Shows concrete fixes for stable pod identities and predictable DNS names, explains replication slot buildup and need for a watchdog when IPs change, and demonstrates a minimal topology: headless service + stateful set + optional persistent volumes. Notes remaining issues with DNS, connection pools and PV reclaim policies.
AI-written summary · about
chapters (11)
- 0:00Start
- 2:01Patroni
- 3:05Single Master Replication
- 4:27Shard Identity
- 5:24Replication Identity
- 7:21Replication Slots
- 10:50Storage Issues
- 11:34Demon Set Abuse
- 24:39Kubernetes Annotations
- 27:58Leader Election
- 31:14Complex Replication Topologies
related talks
54mContainerized Clustered PostgreSQL
36mKube-native Postgres [I]
26mKubernetes from the Database Out
48mPostgreSQL in Kubernetes: Operation - Day 2
37mData On Kubernetes, Deploying And Running PostgreSQL And Patterns For Databases In a Kubernetes Cluster.
17mDebate: PostgreSQL on K8S
30mUSING KUBERNETES/DOCKER/HELM TO DEPLOY ON-DEMAND POSTGRESQL STREAMING REPLICAS
34mRun PostgreSQL:The Kubernetes way