diff --git a/docs-2.0/1.introduction/3.nebula-graph-architecture/4.storage-service.md b/docs-2.0/1.introduction/3.nebula-graph-architecture/4.storage-service.md index 10d851a80d..301020187b 100644 --- a/docs-2.0/1.introduction/3.nebula-graph-architecture/4.storage-service.md +++ b/docs-2.0/1.introduction/3.nebula-graph-architecture/4.storage-service.md @@ -247,10 +247,17 @@ leader 切换对于负载均衡至关重要,当把某个分片从一台机器 为了避免脑裂,当一个 Raft group 的成员发生变化时,需要有一个中间状态,该状态下新旧 group 的多数派需要有重叠的部分,这样就防止了新的 group 或旧的 group 单方面做出决定。为了更加简化,Diego Ongaro 在自己的博士论文中提出每次只增减一个 peer 的方式,以保证新旧 group 的多数派总是有重叠。NebulaGraph 也采用了这个方式,只不过增加成员和移除成员的实现有所区别。具体实现方式请参见 Raft Part class 里 addPeer/removePeer 的实现。 + + ## 与 HDFS 的区别 Storage 服务基于 Raft 协议实现的分布式架构,与 HDFS 的分布式架构有一些区别。例如: diff --git a/docs-2.0/5.configurations-and-logs/1.configurations/4.storage-config.md b/docs-2.0/5.configurations-and-logs/1.configurations/4.storage-config.md index a8db8ae7ea..7e68a1e68b 100644 --- a/docs-2.0/5.configurations-and-logs/1.configurations/4.storage-config.md +++ b/docs-2.0/5.configurations-and-logs/1.configurations/4.storage-config.md @@ -174,6 +174,8 @@ rocksdb options 配置的格式为`{"":""}`,多个 {{ ent.ent_begin }} + ## Black box 配置