Improving Raft Survivability in Real-World Conditions
summary
Describes practical hardening of the Raft consensus algorithm for real-world production conditions. Covers specific implementation issues observed in deployed systems, failure modes that reduce availability, and pragmatic changes and testing approaches to make Raft more resilient in the face of network partitions, process crashes, and operational variability.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 5:55Original Raft problems
- 10:54Tarantool replication features
- 15:33Raft problem solutions
- 23:39Leader election speedup
- 26:07Preventing split-brain
- 33:06Q&A
- 43:16Summary and call to action
related talks
47mUnderstanding Raft Consensus
41mThe Raft Consensus Algorithm
01h 32mHow to Reach Consensus with Paper, Pencil and the Raft Algorithm
47mDistributed Systems: When Democracy and Computer Science Come Together
01h 27mHow to Reach Consensus Using Paper, Pencil and the Raft Algorithm
55mImplementing Raft in RabbitMQ
59mSolving Raft's practical problems in Tarantool. What, how and why
14mDistributed Consensus in 15 Minutes!