Consistent, Distributed Elixir
summary
Claims a working Raft implemented in Elixir that provides strong consistency for state machines, demonstrates a replicated distributed lock across a three-node cluster with leader election and quorum commits, and shows live failover and recovery. Uses a RocksDB-backed durable log for replication and snapshotting, exercises consensus semantics (majority quorum, leader/follower, elections), and is validated with property tests. The demo succeeds: writes replicate, partitions yield errors on non‑leaders, and healed nodes catch up.
AI-written summary · about
chapters (14)
- 0:00Start
- 2:13ELIXIR IS AWESOME
- 5:24PROBLEM: "WE NEED TO LIMIT ACCESS TO AN EXTERNAL RESOURCE"
- 10:16PROBLEM: "WE NEED TO RUN MULTIPLE NODES"
- 20:31PROBLEM: "WE NEED TO KEEP TRACK OF COUNTS"
- 24:15PHOENIX PRESENCE
- 25:04SOME PROBLEMS NEED CONSISTENCY DISTRIBUTED LOCKING DATABASES DISTRIBUTED SCHEDULING AND COORDINATION CONFIGURATION AND METADATA STORAGE TRANSACTIONS
- 26:37CONSENSUS
- 33:45HOW DOES THIS WORK?
- 37:53TESTING PROPERTY TESTS JEPSEN
- 38:32WHAT CAN WE DO NOW?
- 38:43KV STORE SERVICE DISCOVERY DISTRIBUTED LOCK MANAGER DATABASE TRANSACTIONS CONFIGURATION MANAGEMENT
- 39:13TONIO
- 40:06TODO MORE TESTING BETTER OPERATIONS DOCUMENTATION LMDB STORAGE ADAPTER
