ml.shifu.guagua.master
类 SyncMasterCoordinator<MASTER_RESULT extends Bytable,WORKER_RESULT extends Bytable>

java.lang.Object
  ml.shifu.guagua.BasicCoordinator<MASTER_RESULT,WORKER_RESULT>
      ml.shifu.guagua.master.AbstractMasterCoordinator<MASTER_RESULT,WORKER_RESULT>
          ml.shifu.guagua.master.SyncMasterCoordinator<MASTER_RESULT,WORKER_RESULT>

类型参数：

MASTER_RESULT - master result for computation in each iteration.

WORKER_RESULT - worker result for computation in each iteration.

所有已实现的接口：

MasterInterceptor<MASTER_RESULT,WORKER_RESULT>, org.apache.zookeeper.Watcher

public class SyncMasterCoordinator<MASTER_RESULT extends Bytable,WORKER_RESULT extends Bytable>
extends AbstractMasterCoordinator<MASTER_RESULT,WORKER_RESULT>

SyncMasterCoordinator is used to as a barrier for each iteration.

For each iteration, SyncMasterCoordinator will wait until all workers are done.

To start a new iteration, SyncMasterCoordinator will write a znode for each iteration like '/_guagua/job_201312041304_189025/master/{currentIteration}' with with MasterComputable MASTER_RESULT, WORKER_RESULT as its data. This is like a signal to notify workers.

Workers are waiting on current master znode, if got current master znode, it will start another iteration

嵌套类摘要

从类 ml.shifu.guagua.master.AbstractMasterCoordinator 继承的嵌套类/接口

AbstractMasterCoordinator.FailOverCommand, AbstractMasterCoordinator.MasterElectionCommand

从类 ml.shifu.guagua.BasicCoordinator 继承的嵌套类/接口

BasicCoordinator.BasicCoordinatorCommand, BasicCoordinator.CoordinatorCommand, BasicCoordinator.RetryCoordinatorCommand

从接口 org.apache.zookeeper.Watcher 继承的嵌套类/接口

org.apache.zookeeper.Watcher.Event

字段摘要

从类 ml.shifu.guagua.BasicCoordinator 继承的字段

WAIT_SLOT_MILLS

构造方法摘要

SyncMasterCoordinator()

方法摘要

void	preApplication(MasterContext<MASTER_RESULT,WORKER_RESULT> context) The hook point for each application or each mapreduce job which is before all iterations started.
void	preIteration(MasterContext<MASTER_RESULT,WORKER_RESULT> context) The hook point for each iteration which is before MasterComputable.compute(MasterContext).

从类 ml.shifu.guagua.master.AbstractMasterCoordinator 继承的方法

getMyBid, isAllWorkersHalt, postApplication, postIteration, setMyBid, setWorkerResults, updateMasterHaltStatus

从类 ml.shifu.guagua.BasicCoordinator 继承的方法

checkAndSetZooKeeper, closeZooKeeper, getAppNode, getBaseMasterElectionNode, getBytesFromZNode, getCurrentMasterNode, getCurrentMasterSplitNode, getCurrentWorkerNode, getCurrentWorkerSplitNode, getLastMasterNode, getLastWorkerNode, getMasterBaseNode, getMasterElectionNode, getMasterNode, getMasterSerializer, getRootNode, getSleepTime, getWorkerBaseNode, getWorkerBaseNode, getWorkerNode, getWorkerSerializer, getZkConnLatch, getZooKeeper, initialize, isFixedTime, process, setBytesToZNode, setFixedTime, setMasterSerializer, setSleepTime, setWorkerSerializer, setZooKeeper, startHeartbeat, stopHeartBeat

从类 java.lang.Object 继承的方法

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

构造方法详细信息

SyncMasterCoordinator

public SyncMasterCoordinator()

方法详细信息

preApplication

public void preApplication(MasterContext<MASTER_RESULT,WORKER_RESULT> context)

从接口 MasterInterceptor 复制的描述

The hook point for each application or each mapreduce job which is before all iterations started.

preIteration

public void preIteration(MasterContext<MASTER_RESULT,WORKER_RESULT> context)

从接口 MasterInterceptor 复制的描述

The hook point for each iteration which is before MasterComputable.compute(MasterContext).