Class DatanodeAdminBackoffMonitor

java.lang.Object
org.apache.hadoop.hdfs.server.blockmanagement.DatanodeAdminMonitorBase
org.apache.hadoop.hdfs.server.blockmanagement.DatanodeAdminBackoffMonitor
All Implemented Interfaces:
Runnable, org.apache.hadoop.conf.Configurable, DatanodeAdminMonitorInterface

public class DatanodeAdminBackoffMonitor extends DatanodeAdminMonitorBase implements DatanodeAdminMonitorInterface
This class implements the logic to track decommissioning and entering maintenance nodes, ensure all their blocks are adequately replicated before they are moved to the decommissioned or maintenance state. This monitor avoids flooding the replication queue with all pending blocks and instead feeds them to the queue as the prior set complete replication. HDFS-14854 contains details about the overall design of this class.