Highly Available Redis Cluster Is Simple
summary
Describes deploying a highly available Redis cluster using 3 Redis instances (1 master + 2 slaves) with Redis Sentinel for automatic failover, HAProxy as a single entry point, and a second HAProxy plus keepalived to provide a floating VIP. Presents concrete configuration files, commands and ports, and shows a node stop/failover test that achieves failover in tens of seconds but can lose the most recent writes due to asynchronous replication.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 1:28What is Redis
- 4:21Why clusters
- 6:08Replication setup
- 14:53Sentinel monitoring
- 27:10Failover testing
- 33:47Switching master
- 35:34HAProxy load balancing
- 49:16Keepalived failover
- 1:00:15Verifying VIP
- 1:14:49Summary & Q&A
related talks
43mHot Deploying Low-Latency Services for 24/7 Operation
19mMulti-Cluster Service Deployments with Operators and KubeCarrier
38mDeploying Database Clusters in the Cloud
39mHighly Available & Distributed Containers
22mDrupal HA: a story from the trenches
36mHighly Available PostgreSQL Database Cluster Creation & Monitoring in Minutes
40mDIY Fault-Tolerant Microcluster
40mDIY Fault‑Tolerant Microcluster