知识卡片

普通读书笔记卡 · 2313.e.5.f.10.a.4.a.1.c.3.a

维护共识集群保留多数

理解等级:高级

五投票节点集群通常需至少三节点可协作。升级要逐台恢复后再动下一台;扩缩容还需状态追赶及安全成员变更,不能只修改节点列表。多数丢失使系统无法推进,安全协议应停写,而不是自动产生两个合法领导者。

这张卡片看不懂?

已有反馈 · 0 / 10 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • Consistency and Availability · 查看原书依据

      Operating a distributed system based on Paxos or Raft requires meticulous handling of cluster management operations. For instance, when upgrading the software version of a five-node cluster, administrators must carefully orchestrate the process to maintain quorum.

    参考来源

    - 位置:《Building Distributed Systems Designing scalable architectures with microservices, event-driven patterns, and robust deployment…》第04章《Consistency and Availability》(来源项目book-sources;源文件:extracted/Building Distributed Systems Designing scalable architectures with microservices, event-driven patterns, and robust deployment…/_epub-src/OPS/ch04.xhtml) - 结论依据:原文的机制或对比支持以下核心判断:五投票节点集群通常需至少三节点可协作。 本卡将相邻说明中的适用限制保留为解释。 - 原始内容:Operating a distributed system based on Paxos or Raft requires meticulous handling of cluster management operations. For instance, when upgrading the software version of a five-node cluster, administrators must carefully orchestrate the process to maintain quorum.