OpenACID Blog

The Pitfalls of Raft Membership Change

50 minute read

Raft’s single-server membership change carries two problems: its middle state can turn one datacenter into a single point of failure, and the original algori...

Build a Distributed Timestamp Oracle with EzRaft

46 minute read

timestamp_oracle is a 200-line timestamp service on EzRaft. A background task reserves ranges through Raft; next_timestamp allocates from memory. Leader leas...

EzRaft: Build a Distributed KV Store in 100 Lines

56 minute read

EzRaft builds on OpenRaft and hides the storage, networking, and service layers. Define just four things—Request, Response, apply, and read—and you have a fa...

Do Not Read: Raf Is a Useless and Valueless Failed Experiment

77 minute read

raf is an experimental Raft variant that does not persist currentTerm as a separate field. Instead, each election reserves a log index, and that index become...

Histogram Done Right: 2KB Memory, 0.2% Error

41 minute read

A lightweight histogram that tracks latency distributions in 2KB of memory with sub-0.2% error. Uses a float-like encoding for O(1) bucket indexing and trape...

xp 的 AI 工作流

6 minute read

xp 的 AI 开发工作流

Raft Node Rejoin Bug

34 minute read

Analyzes a replication session isolation bug in Raft implementations. When a node rejoins the cluster within the same term, delayed AppendEntries responses c...

Raft 中的 IO 执行顺序：内存状态与持久化状态的陷阱

10 minute read

深入分析 Raft 实现中 IO 重排序导致数据丢失的问题。问题不在 Raft 的设计，而在于实现中内存状态与持久化状态的区分导致的陷阱

Raft Configuration Change with Single Log Entry

42 minute read

Is the single-log-entry approach to Raft configuration change simpler than the standard Joint Consensus?

Raft IO Execution Order (Revised)

38 minute read

I got it wrong in my previous article. The IO ordering bug in Raft isn’t about the protocol design—it’s about the subtle trap that emerges when implementatio...

The Hidden Danger in Raft: Why IO Ordering Matters

23 minute read

Writing logs before persisting term in Raft can silently destroy committed data. Here’s why production systems like TiKV and HashiCorp Raft carefully control...