Don’t Hassle Me, I’m Stateful
summary
Deploys a full MongoDB replica set on OpenShift using Kubernetes StatefulSet (introduced in Kubernetes 1.5, heading GA in 1.9), with a single shard, config servers, and ordered pod ordinals 0…n. Each replica gets per-pod persistent volumes provisioned before the pod, sticky network names via a headless service DNS names, and preserved PVs on delete. Live demo kills a primary and shows re‑election, scaling adds matching PVCs, and rolling/partitioned updates preserve index‑based roles. Overall the pattern automates migration of legacy clustered apps to containers.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 0:18Agenda
- 0:45Ground Rules
- 1:13What are Staple Sets
- 1:31Features of Staple Sets
- 2:30Problems with Staple Sets
- 4:21Cassandra
- 4:46Confluence
- 5:33Kubernetes
- 9:24Live Demo
- 19:38Addresses
- 22:46Individual Instance
- 23:33Staple Set Pod
- 24:53Deleting Pods
- 26:08Stateful Objects
- 28:28Network Configuration
- 30:33Deployment History
related talks
32mMulti-Cluster Stateful Set Migration: A Solution To Upgrade Pain
46mConfiguring and Deploying Stateful Applications with PetSet
25mStatefulsets on Kubernetes: The Guardians of Your Data Galaxy
29mStateful Workloads in Kubernetes: A Deep Dive
37mStateful vs Stateless Container Patterns in Kubernetes, a Deeper Analysis
01h 19mTutorial: Back to Basics: Hands-On Deployment of Stateful Workloads on Kubernetes
05mUpgrading Stateful Sets in Constrained Environment
35mWhen StatefulSets are not enough