elastic job源码分析 - 注册中心连接状态监听器

注册中心连接状态监听器io.elasticjob.lite.internal.listener.RegistryCenterConnectionStateListener实现了org.apache.curator.framework.state.ConnectionStateListener接口。

Class diagram

ConnectionStateListener只有一个方法stateChanged，它会在客户端与zk服务器连接状态改变时调用。

/**
 * Called when there is a state change in the connection
 *
 * @param client the client
 * @param newState the new state
 */
public void stateChanged(CuratorFramework client, ConnectionState newState);

elastic job通过该监听器，可以管理实例与zk连接状态不同时的行为。包括：SUSPENDED（暂停挂起），LOST（确认丢失）和RECONNECTED（重新连接）。

SUSPENDED
There has been a loss of connection. Leaders, locks, etc. should suspend until the connection is re-established. If the connection times-out you willreceive a LOST notice.

LOST
The connection is confirmed to be lost. Close any locks, leaders, etc. and attempt to re-create them. NOTE: it is possible to get a RECONNECTED state after this but you should still consider any locks, etc. as dirty/unstable.

RECONNECTED
A suspended, lost, or read-only connection has been re-established.

暂停挂起和确认丢失

当elastic job实例接收到与zk处于暂停挂起和确认丢失状态时，当前实例会暂停调度任务。

if (ConnectionState.SUSPENDED == newState || ConnectionState.LOST == newState) {
    jobScheduleController.pauseJob();
}

重新连接

当实例接收与zk服务端重新连接时，实例除了继续发起任务调度外，还会重新初始化在zk的状态。

1. 持久化作业服务器上线信息；

创建zk节点jobName/servers/实例IP.

2. 持久化作业运行实例上线相关信息；

创建zk临时节点jobName/instances/ip@-@pid.

3. 清除分配分片项的运行状态。

删除zk节点jobName/sharding/本地分片项/running（如果存在）.

if (ConnectionState.RECONNECTED == newState) {
    serverService.persistOnline(serverService.isEnableServer(JobRegistry.getInstance().getJobInstance(jobName).getIp()));
    instanceService.persistOnline();
    executionService.clearRunningInfo(shardingService.getLocalShardingItems());
    jobScheduleController.resumeJob();
}