org.apache.flink.runtime.leaderelection

接口 LeaderElectionDriver

所有超级接口:

AutoCloseable

所有已知实现类:

ZooKeeperLeaderElectionDriver

public interface LeaderElectionDriver
extends AutoCloseable

A LeaderElectionDriver is responsible for performing the leader election and storing the leader information. All the leader internal state is guarded by lock in LeaderElectionService. Different driver implementations do not need to care about the lock. And it should use LeaderElectionEventHandler if it want to respond to the leader change events.

Important: The LeaderElectionDriver could not guarantee that there is no LeaderElectionEventHandler callbacks happen after AutoCloseable.close().

方法概要

限定符和类型	方法和说明
boolean	hasLeadership() Check whether the driver still have the leadership in the distributed coordination system.
void	writeLeaderInformation(LeaderInformation leaderInformation) Write the current leader information to external persistent storage(e.g.

从接口继承的方法 java.lang.AutoCloseable

close

方法详细资料

writeLeaderInformation

void writeLeaderInformation(LeaderInformation leaderInformation)

Write the current leader information to external persistent storage(e.g. Zookeeper, Kubernetes ConfigMap). This is a blocking IO operation. The write operation takes effect only when the driver still has the leadership.

参数:

leaderInformation - current leader information. It could be LeaderInformation.empty(), which means the caller want to clear the leader information on external storage. Please remember that the clear operation should only happen before a new leader is elected and has written his leader information on the storage. Otherwise, we may have a risk to wrongly update the storage with empty leader information.

hasLeadership

boolean hasLeadership()

Check whether the driver still have the leadership in the distributed coordination system.

返回:

Return whether the driver has leadership.