Hadoop2.6.0中YARN底层状态机实现分析

标签：

前言

Hadoop2.x.x版本的底层实现中作了很多优化：用状态机对各种对象生命周期和状态转移进行管理；采用事件机制避免线程同步与阻塞；采用Protocol Buffers优化RPC性能；采用Apache Avro优化日志等。本文主要针对YARN中状态机的实现进行分析，在这个过程中，会捎带一些事件的内容。

事件

YARN中的很多组件之间进行通信，主要借助于事件。为了可读性、可维护性及可扩展性，YARN中的事件由事件名称和事件类型组成。比如JobImpl处理的事件名称为JobEvent，而事件类型为JobEventType。有关Hadoop2.6.0的事件分类与实现可以参考《Hadoop2.6.0的事件分类与实现》一文。

状态

YARN中的每个组件都有其自身所处的一系列状态，比如JobImpl内部的一系列状态都定义在JobStateInternal中，如代码清单1所示。

代码清单1

public enum JobStateInternal {
  NEW,
  SETUP,
  INITED,
  RUNNING,
  COMMITTING,
  SUCCEEDED,
  FAIL_WAIT,
  FAIL_ABORT,
  FAILED,
  KILL_WAIT,
  KILL_ABORT,
  KILLED,
  ERROR,
  REBOOT
}

转换（过渡）

我们已经了解了事件与状态的基本实现与概念，那么事件与状态有什么关系？从哲学角度讲，状态是一个事物的静止属性，而事件则是一个事物与外界沟通的桥梁，只有静止却没有变化，那么它只是一滩死水。事物只有在接收信息后动起来，才算与外界有了互动。一个事物动起来就会潜移默化的发生改变，它内部就会发生转换。一个对象当前处于状态state0，当对象接收到事件Event后，将引发转换动作transition，最终当前对象的状态过渡到state1，这个过程可以用图1来表示。

图1 状态迁移示例

YARN中与过渡相关的类图如图2所示。

图2 YARN中与过渡相关的类图

YARN中的各个组件的变化都离不开状态的过度与变化，于是对这种行为进行了抽象，这种转换分为两类：单弧过渡与多弧过渡。（这种翻译不知道是否准确，我认为从一个状态到另一个状态的转换发生时，就像是在两个状态之间划了一道弧线一样）

单弧过渡

YARN中单弧过渡的实现代码如代码清单2，它的作用是当有限状态机（FSM）中的状态转换为已经注册到状态机的某种状态时，伴随的行为。

代码清单2

@Public
@Evolving
public interface SingleArcTransition<OPERAND, EVENT> {
  /**
   * Transition hook.
   * 
   * @param operand the entity attached to the FSM, whose internal 
   *                state may change.
   * @param event causal event
   */
  public void transition(OPERAND operand, EVENT event);

}

由于SingleArcTransition的具体实现类只负责接收到事件后的具体操作或行为，并没有包含状态相关的信息，所以在状态机执行状态过渡时，并不是直接调用SingleArcTransition具体实现类的transition方法，而是由接口Transition定义（见代码清单3）真正的转态过渡（包括行为和状态改变）。

代码清单3

  private interface Transition<OPERAND, STATE extends Enum<STATE>,
          EVENTTYPE extends Enum<EVENTTYPE>, EVENT> {
    STATE doTransition(OPERAND operand, STATE oldState,
                       EVENT event, EVENTTYPE eventType);
  }

SingleInternalArc作为Transition接口的实现类，在代理SingleArcTransition的同时，负责状态变换，见代码清单4。

代码清单4

  private class SingleInternalArc
                    implements Transition<OPERAND, STATE, EVENTTYPE, EVENT> {

    private STATE postState;
    private SingleArcTransition<OPERAND, EVENT> hook; // transition hook

    SingleInternalArc(STATE postState,
        SingleArcTransition<OPERAND, EVENT> hook) {
      this.postState = postState;
      this.hook = hook;
    }

    @Override
    public STATE doTransition(OPERAND operand, STATE oldState,
                              EVENT event, EVENTTYPE eventType) {
      if (hook != null) {
        hook.transition(operand, event);
      }
      return postState;
    }
  }

多弧过渡

YARN中多弧过渡的实现代码如代码清单5，它的作用是当有限状态机（FSM）中的状态转换为已经注册到状态机的多个有效状态中的一个时，伴随的行为与操作。

代码清单5

@Public
@Evolving
public interface MultipleArcTransition
        <OPERAND, EVENT, STATE extends Enum<STATE>> {

  /**
   * Transition hook.
   * @return the postState. Post state must be one of the 
   *                      valid post states registered in StateMachine.
   * @param operand the entity attached to the FSM, whose internal 
   *                state may change.
   * @param event causal event
   */
  public STATE transition(OPERAND operand, EVENT event);

}

代码清单6

  private class MultipleInternalArc
              implements Transition<OPERAND, STATE, EVENTTYPE, EVENT>{

    // Fields
    private Set<STATE> validPostStates;
    private MultipleArcTransition<OPERAND, EVENT, STATE> hook;  // transition hook

    MultipleInternalArc(Set<STATE> postStates,
                   MultipleArcTransition<OPERAND, EVENT, STATE> hook) {
      this.validPostStates = postStates;
      this.hook = hook;
    }

    @Override
    public STATE doTransition(OPERAND operand, STATE oldState,
                              EVENT event, EVENTTYPE eventType)
        throws InvalidStateTransitonException {
      STATE postState = hook.transition(operand, event);

      if (!validPostStates.contains(postState)) {
        throw new InvalidStateTransitonException(oldState, eventType);
      }
      return postState;
    }
  }

为了将所有状态机中的状态过渡与状态建立起映射关系，YARN中提供了ApplicableTransition接口用于将SingleInternalArc和MultipleInternalArc添加到状态机的拓扑表中，提高在检索状态对应的过渡实现时的性能，ApplicableTransition的实现类为ApplicableSingleOrMultipleTransition类，其apply方法用于代理SingleInternalArc和MultipleInternalArc，将它们添加到状态拓扑表中。ApplicableTransition接口的定义见代码清单7，ApplicableSingleOrMultipleTransition的实现见代码清单8。

代码清单7

  private interface ApplicableTransition
             <OPERAND, STATE extends Enum<STATE>,
              EVENTTYPE extends Enum<EVENTTYPE>, EVENT> {
    void apply(StateMachineFactory<OPERAND, STATE, EVENTTYPE, EVENT> subject);
  }

代码清单8

  static private class ApplicableSingleOrMultipleTransition
             <OPERAND, STATE extends Enum<STATE>,
              EVENTTYPE extends Enum<EVENTTYPE>, EVENT>
          implements ApplicableTransition<OPERAND, STATE, EVENTTYPE, EVENT> {
    final STATE preState;
    final EVENTTYPE eventType;
    final Transition<OPERAND, STATE, EVENTTYPE, EVENT> transition;

    ApplicableSingleOrMultipleTransition
        (STATE preState, EVENTTYPE eventType,
         Transition<OPERAND, STATE, EVENTTYPE, EVENT> transition) {
      this.preState = preState;
      this.eventType = eventType;
      this.transition = transition;
    }

    @Override
    public void apply
             (StateMachineFactory<OPERAND, STATE, EVENTTYPE, EVENT> subject) {
      Map<EVENTTYPE, Transition<OPERAND, STATE, EVENTTYPE, EVENT>> transitionMap
        = subject.stateMachineTable.get(preState);
      if (transitionMap == null) {
        // I use HashMap here because I would expect most EVENTTYPE‘s to not
        //  apply out of a particular state, so FSM sizes would be 
        //  quadratic if I use EnumMap‘s here as I do at the top level.
        transitionMap = new HashMap<EVENTTYPE,
          Transition<OPERAND, STATE, EVENTTYPE, EVENT>>();
        subject.stateMachineTable.put(preState, transitionMap);
      }
      transitionMap.put(eventType, transition);
    }
  }

状态机

YARN中状态机的实现类是StateMachineFactory，它主要包含4个属性信息：

• transitionsListNode：过渡列表节点。根据其名字不太容易理解，我这里说得简单点，就是将状态机的一个个过渡的ApplicableTransition实现串联为一个列表，每个节点包含一个ApplicableTransition实现及指向下一个节点的引用，其实现见代码清单9所示。

代码清单9

  private class TransitionsListNode {
    final ApplicableTransition<OPERAND, STATE, EVENTTYPE, EVENT> transition;
    final TransitionsListNode next;

    TransitionsListNode
        (ApplicableTransition<OPERAND, STATE, EVENTTYPE, EVENT> transition,
        TransitionsListNode next) {
      this.transition = transition;
      this.next = next;
    }
  }

transitionsListNode形成的过渡列表节点可以用图3表示。

图3 transitionsListNode过渡链表结构

• stateMachineTable：状态拓扑表，为了提高检索状态对应的过渡map而冗余的数据结构，此结构在optimized为真时，通过对transitionsListNode链表进行处理产生。stateMachineTable的结构可以用图4来表示。

图4 状态拓扑表数据结构

• defaultInitialState：对象创建时，内部有限状态机的默认初始状态。比如：JobImpl的内部状态机默认初始状态是JobStateInternal.NEW。
• optimized：布尔类型，用于标记当前状态机是否需要优化性能，即构建状态拓扑表stateMachineTable。

共有构造器

StateMachineFactory的公有构造器只有一个，其实现见代码清单10。

代码清单10

  public StateMachineFactory(STATE defaultInitialState) {
    this.transitionsListNode = null;
    this.defaultInitialState = defaultInitialState;
    this.optimized = false;
    this.stateMachineTable = null;
  }

私有构造器

StateMachineFactory的私有构造器有两个，其中代码清单11中的构造器在addTransition方法中使用。从其实现看出，此构造器的主要作用是构建transitionsListNode链表。

代码清单11

  private StateMachineFactory
      (StateMachineFactory<OPERAND, STATE, EVENTTYPE, EVENT> that,
       ApplicableTransition<OPERAND, STATE, EVENTTYPE, EVENT> t) {
    this.defaultInitialState = that.defaultInitialState;
    this.transitionsListNode 
        = new TransitionsListNode(t, that.transitionsListNode);
    this.optimized = false;
    this.stateMachineTable = null;
  }

代码清单12

  private StateMachineFactory
      (StateMachineFactory<OPERAND, STATE, EVENTTYPE, EVENT> that,
       boolean optimized) {
    this.defaultInitialState = that.defaultInitialState;
    this.transitionsListNode = that.transitionsListNode;
    this.optimized = optimized;
    if (optimized) {
      makeStateMachineTable();
    } else {
      stateMachineTable = null;
    }
  }

代码清单13

  private void makeStateMachineTable() {
    Stack<ApplicableTransition<OPERAND, STATE, EVENTTYPE, EVENT>> stack =
      new Stack<ApplicableTransition<OPERAND, STATE, EVENTTYPE, EVENT>>();

    Map<STATE, Map<EVENTTYPE, Transition<OPERAND, STATE, EVENTTYPE, EVENT>>>
      prototype = new HashMap<STATE, Map<EVENTTYPE, Transition<OPERAND, STATE, EVENTTYPE, EVENT>>>();

    prototype.put(defaultInitialState, null);

    // I use EnumMap here because it‘ll be faster and denser.  I would
    //  expect most of the states to have at least one transition.
    stateMachineTable
       = new EnumMap<STATE, Map<EVENTTYPE,
                           Transition<OPERAND, STATE, EVENTTYPE, EVENT>>>(prototype);

    for (TransitionsListNode cursor = transitionsListNode;
         cursor != null;
         cursor = cursor.next) {
      stack.push(cursor.transition);
    }

    while (!stack.isEmpty()) {
      stack.pop().apply(this);
    }
  }

1. 创建堆栈stack，用于将transitionsListNode链表中各个节点持有的ApplicableSingleOrMultipleTransition压入栈中；
2. 创建状态拓扑表stateMachineTable，并在此拓扑表中插入一个额外的默认初始状态defaultInitialState与null的映射；
3. 迭代访问transitionsListNode链表，并将各个节点持有的ApplicableSingleOrMultipleTransition压入栈中；
4. 依次弹出栈顶的ApplicableSingleOrMultipleTransition，并应用其apply方法（已在前面小节介绍），持续不断的构建状态拓扑表stateMachineTable。

状态机构建

为了简化叙述，本节以JobImpl中状态机的构建为例。由于JobImpl的状态机预设的（调用addTransition方法）加入的ApplicableSingleOrMultipleTransition非常多，我们节选其中的2个作为典型进行分析。最后还会分析installTopology方法的实现。JobImpl中状态机的定义见代码清单14。

代码清单14

  protected static final
    StateMachineFactory<JobImpl, JobStateInternal, JobEventType, JobEvent> 
       stateMachineFactory
     = new StateMachineFactory<JobImpl, JobStateInternal, JobEventType, JobEvent>
              (JobStateInternal.NEW)

          // Transitions from NEW state
          .addTransition(JobStateInternal.NEW, JobStateInternal.NEW,
              JobEventType.JOB_DIAGNOSTIC_UPDATE,
              DIAGNOSTIC_UPDATE_TRANSITION)
          .addTransition(JobStateInternal.NEW, JobStateInternal.NEW,
              JobEventType.JOB_COUNTER_UPDATE, COUNTER_UPDATE_TRANSITION)
          .addTransition
              (JobStateInternal.NEW,
              EnumSet.of(JobStateInternal.INITED, JobStateInternal.NEW),
              JobEventType.JOB_INIT,
              new InitTransition())
          // 省略其它addTransition调用
          // create the topology tables
          .installTopology();

1. 调用StateMachineFactory构造器创建一个初始的状态机；
2. 调用addTransition(STATE preState, STATE postState, EVENTTYPE eventType, SingleArcTransition<OPERAND, EVENT> hook)方法添加单弧过渡。从其实现（见代码清单15）可以知道addTransition方法将SingleArcTransition封装为SingleInternalArc，然后将SingleInternalArc封装为ApplicableSingleOrMultipleTransition，最后调用之前说的第一个私有构造器构建transitionsListNode链表；
3. 调用addTransition(STATE preState, Set<STATE> postStates, EVENTTYPE eventType, MultipleArcTransition<OPERAND, EVENT, STATE> hook)方法添加多弧过渡。从其实现（见代码清单16）可以知道addTransition方法将MultipleArcTransition封装为MultipleInternalArc，然后将MultipleInternalArc封装为ApplicableSingleOrMultipleTransition，最后调用之前说的第一个私有构造器构建transitionsListNode链表；
4. 最后调用installTopology方法，其实现见代码清单17。installTopology正是在使用之前说的第二个私有构造器构建状态拓扑表stateMachineTable；

  public StateMachineFactory
             <OPERAND, STATE, EVENTTYPE, EVENT>
          addTransition(STATE preState, STATE postState,
                        EVENTTYPE eventType,
                        SingleArcTransition<OPERAND, EVENT> hook){
    return new StateMachineFactory<OPERAND, STATE, EVENTTYPE, EVENT>
        (this, new ApplicableSingleOrMultipleTransition<OPERAND, STATE, EVENTTYPE, EVENT>
           (preState, eventType, new SingleInternalArc(postState, hook)));
  }

  public StateMachineFactory
             <OPERAND, STATE, EVENTTYPE, EVENT>
          addTransition(STATE preState, Set<STATE> postStates,
                        EVENTTYPE eventType,
                        MultipleArcTransition<OPERAND, EVENT, STATE> hook){
    return new StateMachineFactory<OPERAND, STATE, EVENTTYPE, EVENT>
        (this,
         new ApplicableSingleOrMultipleTransition<OPERAND, STATE, EVENTTYPE, EVENT>
           (preState, eventType, new MultipleInternalArc(postStates, hook)));
  }

  public StateMachineFactory
             <OPERAND, STATE, EVENTTYPE, EVENT>
          installTopology() {
    return new StateMachineFactory<OPERAND, STATE, EVENTTYPE, EVENT>(this, true);
  }

  private static final DiagnosticsUpdateTransition
      DIAGNOSTIC_UPDATE_TRANSITION = new DiagnosticsUpdateTransition();

  private static class DiagnosticsUpdateTransition implements
      SingleArcTransition<JobImpl, JobEvent> {
    @Override
    public void transition(JobImpl job, JobEvent event) {
      job.addDiagnostic(((JobDiagnosticsUpdateEvent) event)
          .getDiagnosticUpdate());
    }
  }

  public static class InitTransition 
      implements MultipleArcTransition<JobImpl, JobEvent, JobStateInternal> {

    @Override
    public JobStateInternal transition(JobImpl job, JobEvent event) {
        // 省略具体逻辑
      }
    }
  }

  private STATE doTransition
           (OPERAND operand, STATE oldState, EVENTTYPE eventType, EVENT event)
      throws InvalidStateTransitonException {
    // We can assume that stateMachineTable is non-null because we call
    //  maybeMakeStateMachineTable() when we build an InnerStateMachine ,
    //  and this code only gets called from inside a working InnerStateMachine .
    Map<EVENTTYPE, Transition<OPERAND, STATE, EVENTTYPE, EVENT>> transitionMap
      = stateMachineTable.get(oldState);
    if (transitionMap != null) {
      Transition<OPERAND, STATE, EVENTTYPE, EVENT> transition
          = transitionMap.get(eventType);
      if (transition != null) {
        return transition.doTransition(operand, oldState, event, eventType);
      }
    }
    throw new InvalidStateTransitonException(oldState, eventType);
  }

后记：个人总结整理的《深入理解Spark：核心思想与源码分析》一书现在已经正式出版上市，目前京东、当当、天猫等网站均有销售，欢迎感兴趣的同学购买。

京东：http://item.jd.com/11846120.html 

当当：http://product.dangdang.com/23838168.html 

Hadoop2.6.0中YARN底层状态机实现分析

标签：

原文地址：http://blog.csdn.net/beliefer/article/details/51190842