Building the Multi-Cluster Data Layer
summary
Deploys YugabyteDB across three GKE clusters using Helm and Istio multi-primary with remote secrets, per-pod ClusterIP services and Kiali for visibility, running six tservers and three masters with replication factor 3. Shows that masters and tservers reach each other across networks after exposing per-pod services, then simulates a full region outage and the app continues to read and write from the remaining clusters. Also demonstrates leader locality via yb-admin to pin shard leaders and reduce cross-region latency.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 2:15Database Design
- 2:47Benefits of Deployment
- 4:29Yugabyte Model
- 5:51Installing on Kubernetes
- 8:17Highlevel requirements
- 9:14Simplified view
- 9:51Steps
- 10:25Service Discovery
- 11:45Connectivity
- 13:08Cross Cluster Access
- 14:10Replication Factor 3 Deployment
- 15:42Cluster IP Services
- 20:13Fault Tolerance Demo
- 24:55Region Level Outage Demo
- 28:46Use Cases
related talks
32mZero-Touch Fault-Tolerance for Cloud-Native Geo-Distributed Databa...
01h 41mMastering Multi-Region Deployments With YugabyteDB
30mYugabyteDB Elasticity and Resilience in Containers
31mEnd-to-End Stack: Harnessing YugabyteDB & Ecosystem Solutions, Build Apps on Global Scale
26mTip and Tricks to Deploying YugabyteDB in a Multi-Cloud World
46mA DB on K8s, Cloud-Native, Open-Source, and PostgreSQL-Compatible: YugabyteDB
26mA Winning Multi Cloud Data Strategy
30mBuilding Scalable Spring Applications with YugabyteDB
