Fault Tolerant 24/7 Operations with Aeron Cluster
Talkat Open Source in Finance Forum New York 2022 · on Thursday, 2022-12-08· at 16:15· HUB 1, New York, NY, USA·
summary
Aeron Cluster (built on Aeron and Aeron Archive) implements a Raft-based replicated log with snapshots and replay, showing region‑wide fault tolerance and practical DR options: cold snapshots in S3, warm standby pulling logs or fully hot DR that flips in seconds. Measured peak throughput reaches 1.8M msgs/s with a sub‑20µs P99 roundtrip. Operational lessons: snapshot frequency vs replay time, asynchronous snapshot dissemination, run 5–7 node clusters for safe rolling upgrades, and version everything for mixed‑version deployments.
AI-written summary · about
chapters (6)
- 0:00Start
- 11:38Disaster Recovery
- 14:29Recovery Time Objective
- 18:0624/7 Upgrades
- 21:52Rolling Upgrades
- 24:57Hard Lessons
related talks
31mUser Attribution at Scale: All-Flash, ½ Trillion Objects
32mReal time customer experience across industries
46mEdge Performance with In memory NoSQL
22mBuilding CockroachDB for performance at scale
44mAvailability, Latency and Cost: Withstanding Outages
27mAvailability and Performance Tradeoffs in Global Database Deployments
40mCloud Databases: Pros and Cons
05mHighlights From Aerospike APJ Summit 2023
