57 lines
1.2 KiB
Markdown
57 lines
1.2 KiB
Markdown
|
||
|
||
# 简介
|
||
|
||
HDFS的NameNode、Yarn的ResourceManager都是依靠ZK实现主备倒换的。核心的类为:ZKFailoverController.java,
|
||
选举的核心类为ActiveStandbyElector.java
|
||
|
||
## 主备选举
|
||
|
||
主备选举的核心类是ActiveStandbyElector。在初始化的时候需要创建zk连接并且尝试在zk上面创建文件。在创建连接或者创建文件的时候都会有回调事件。
|
||
|
||
回调处理的函数主要包含:
|
||
|
||
|
||
### 创建node节点回调
|
||
|
||
入口函数如下:
|
||
```java
|
||
public synchronized void processResult(int rc, String path, Object ctx,
|
||
String name) {
|
||
// .....
|
||
}
|
||
```
|
||
|
||
处理流程图如下:
|
||
|
||
![zk_failver_001](https://pan.zeekling.cn//zeekling/hadoop/common/zk_failover_001.png)
|
||
|
||
|
||
|
||
### 监控回调
|
||
|
||
入口函数如下:
|
||
```java
|
||
public synchronized void processResult(int rc, String path, Object ctx,
|
||
Stat stat) {
|
||
// ...
|
||
}
|
||
```
|
||
|
||
处理流程如下:
|
||
|
||
![zk_failver_003](https://pan.zeekling.cn//zeekling/hadoop/common/zk_failover_003.png)
|
||
|
||
### 事件回调
|
||
|
||
入口函数如下:
|
||
|
||
```java
|
||
synchronized void processWatchEvent(ZooKeeper zk, WatchedEvent event) {
|
||
// ..
|
||
}
|
||
```
|
||
|
||
![zk_failver_002](https://pan.zeekling.cn//zeekling/hadoop/common/zk_failover_002.png)
|
||
|