elasticJob源码剖析

 

job的执行过程

首先从demo开始看

 

//com.dangdang.ddframe.job.example.JavaMain
public static void main(final String[] args) throws IOException {
    // CHECKSTYLE:ON
        //zk创建 
        CoordinatorRegistryCenter regCenter = setUpRegistryCenter();
        //作业事件配置
        JobEventConfiguration jobEventConfig = 
        new JobEventRdbConfiguration(setUpEventTraceDataSource());
        setUpSimpleJob(regCenter, jobEventConfig);//-->方法在下面
    }

 

//com.dangdang.ddframe.job.example.JavaMain
//overwrite(true)  是否覆盖zk的配置,如果为false,则改动后重启也不会覆盖zk已有的配置
private static void setUpSimpleJob(final CoordinatorRegistryCenter regCenter, final JobEventConfiguration jobEventConfig) {
    //作业的核心配置
    JobCoreConfiguration coreConfig = JobCoreConfiguration.newBuilder("javaSimpleJob", "0/1 * * * * ?", 3).shardingItemParameters("0=Beijing,1=Shanghai,2=Guangzhou").build();
    SimpleJobConfiguration simpleJobConfig = new SimpleJobConfiguration(coreConfig, JavaSimpleJob.class.getCanonicalName());
    //创建一个job,并且初始化
    new JobScheduler(regCenter, LiteJobConfiguration.newBuilder(simpleJobConfig).overwrite(true).build(), jobEventConfig).init();
}

构造一个JobScheduler

//io.elasticjob.lite.api.JobScheduler
private JobScheduler(final CoordinatorRegistryCenter regCenter, final LiteJobConfiguration liteJobConfig, final JobEventBus jobEventBus, final ElasticJobListener... elasticJobListeners) {
    //将job任务统一交由注册器JobRegistry统一管理
    JobRegistry.getInstance().addJobInstance(liteJobConfig.getJobName(), new JobInstance());
    this.liteJobConfig = liteJobConfig;
    this.regCenter = regCenter;
    //弹性化分布式作业监听器集合
    List<ElasticJobListener> elasticJobListenerList = Arrays.asList(elasticJobListeners);
    setGuaranteeServiceForElasticJobListeners(regCenter, elasticJobListenerList);
    schedulerFacade = new SchedulerFacade(regCenter, liteJobConfig.getJobName(), elasticJobListenerList);
    jobFacade = new LiteJobFacade(regCenter, liteJobConfig.getJobName(), Arrays.asList(elasticJobListeners), jobEventBus);
}

在init方法初始化作业

 

//io.elasticjob.lite.api.JobScheduler
/**
 * 初始化作业.
 */
public void init() {
    //1.更新作业配置.会判断,是否存在节点,或者是否LiteJobConfiguration里面的overwrite是否为true,如果满足一个条件就更新配置
    //然后直接从注册中心而非本地缓存获取作业节点最新数据.
    LiteJobConfiguration liteJobConfigFromRegCenter = schedulerFacade.updateJobConfiguration(liteJobConfig);
    //2.设置当前分片总数.和作业名称
    JobRegistry.getInstance().setCurrentShardingTotalCount(liteJobConfigFromRegCenter.getJobName(), liteJobConfigFromRegCenter.getTypeConfig().getCoreConfig().getShardingTotalCount());
    //3.创建一个作业调度控制器.
    JobScheduleController jobScheduleController = new JobScheduleController(
            //3.1根据配置文件创建一个调度器
            createScheduler(),
            //3.2根据执行任务的class创建一个调度器详情实例
            createJobDetail(liteJobConfigFromRegCenter.getTypeConfig().getJobClass()),
            //3.3获取job名称
            liteJobConfigFromRegCenter.getJobName());
    //4.job实例设置job的名称,作业调度控制器,和用于协调分布式服务的注册中心.
    JobRegistry.getInstance().registerJob(liteJobConfigFromRegCenter.getJobName(), jobScheduleController, regCenter);
    //5.注册并且启动信息,启动各种监听器
    schedulerFacade.registerStartUpInfo(!liteJobConfigFromRegCenter.isDisabled());
    //6.根据cron表达式调度作业
    jobScheduleController.scheduleJob(liteJobConfigFromRegCenter.getTypeConfig().getCoreConfig().getCron());
}

1.1

/**
 * 更新作业配置.
 * SchedulerFacade.java
 * @param liteJobConfig 作业配置
 * @return 更新后的作业配置
 */
public LiteJobConfiguration updateJobConfiguration(final LiteJobConfiguration liteJobConfig) {
    //更新或者添加作业配置
    configService.persist(liteJobConfig);
    //从zk去获取更新后的作业配置
    return configService.load(false);
}

1.2.1

/**
 * 读取作业配置.
 * 
 * @param fromCache 是否从缓存中读取
 * @return 作业配置
 */
public LiteJobConfiguration load(final boolean fromCache) {
    String result;
    if (fromCache) {
        result = jobNodeStorage.getJobNodeData(ConfigurationNode.ROOT);
        if (null == result) {
            result = jobNodeStorage.getJobNodeDataDirectly(ConfigurationNode.ROOT);
        }
    } else {
        //直接从注册中心而非本地缓存获取作业节点数据.
        result = jobNodeStorage.getJobNodeDataDirectly(ConfigurationNode.ROOT);
    }
    return LiteJobConfigurationGsonFactory.fromJson(result);
}

2.1

/**
 * 设置当前分片总数.
 *
 * @param jobName 作业名称
 * @param currentShardingTotalCount 当前分片总数
 */
public void setCurrentShardingTotalCount(final String jobName, final int currentShardingTotalCount) {
    currentShardingTotalCountMap.put(jobName, currentShardingTotalCount);
}

3.1.1

//io.elasticjob.lite.api.JobScheduler
/**
 * 创建调度程序
 * @return
 */
private Scheduler createScheduler() {
    Scheduler result;
    try {
        StdSchedulerFactory factory = new StdSchedulerFactory();
        //获取基本Quartz的属性
        factory.initialize(getBaseQuartzProperties());
        //从工厂获取一个调度程序
        result = factory.getScheduler();
        //往调度程序添加一个作业触发监听器
        result.getListenerManager().addTriggerListener(schedulerFacade.newJobTriggerListener());
    } catch (final SchedulerException ex) {
        throw new JobSystemException(ex);
    }
    return result;
}

/**
 * 获取基本Quartz的属性
 * @return
 */
private Properties getBaseQuartzProperties() {
    Properties result = new Properties();
    result.put("org.quartz.threadPool.class", org.quartz.simpl.SimpleThreadPool.class.getName());
    result.put("org.quartz.threadPool.threadCount", "1");
    result.put("org.quartz.scheduler.instanceName", liteJobConfig.getJobName());
    result.put("org.quartz.jobStore.misfireThreshold", "1");
    result.put("org.quartz.plugin.shutdownhook.class", JobShutdownHookPlugin.class.getName());
    result.put("org.quartz.plugin.shutdownhook.cleanShutdown", Boolean.TRUE.toString());
    return result;
}

3.2.1

//io.elasticjob.lite.api.JobScheduler
/**
 * 创建一个job详情
 * @param jobClass
 * @return
 */
private JobDetail createJobDetail(final String jobClass) {
    //3.2.1.1通过建造者模式,创建一个调度作业
    JobDetail result = JobBuilder.newJob(LiteJob.class).withIdentity(liteJobConfig.getJobName()).build();
    result.getJobDataMap().put(JOB_FACADE_DATA_MAP_KEY, jobFacade);
    Optional<ElasticJob> elasticJobInstance = createElasticJobInstance();
    //flase
    if (elasticJobInstance.isPresent()) {
        result.getJobDataMap().put(ELASTIC_JOB_DATA_MAP_KEY, elasticJobInstance.get());
    } else if (!jobClass.equals(ScriptJob.class.getCanonicalName())) {
        try {
            //把job的实例放进jobDataMap
            result.getJobDataMap().put(ELASTIC_JOB_DATA_MAP_KEY, Class.forName(jobClass).newInstance());
        } catch (final ReflectiveOperationException ex) {
            throw new JobConfigurationException("Elastic-Job: Job class '%s' can not initialize.", jobClass);
        }
    }
    return result;
}

3.2.1.1具体创建一个调度作业的流程

//io.elasticjob.lite.internal.schedule.LiteJob
/**
 * Lite调度作业.
 *
 * @author zhangliang
 */
public final class LiteJob implements Job {
    
    @Setter
    private ElasticJob elasticJob;
    
    @Setter
    private JobFacade jobFacade;
    
    @Override
    public void execute(final JobExecutionContext context) throws JobExecutionException {
        //通过 JobExecutorFactory 获得到作业执行器并进行执行
        JobExecutorFactory.getJobExecutor(elasticJob, jobFacade).execute();
    }
}

根据不同类型获取不同的作业执行器

//io.elasticjob.lite.executor.JobExecutorFactory
/**
 * 作业执行器工厂.
 *
 * @author zhangliang
 */
@NoArgsConstructor(access = AccessLevel.PRIVATE)
public final class JobExecutorFactory {
    
    /**
     * 获取作业执行器.
     * 首先会根据elasticJob的类型去找到相应的执行器
     * @param elasticJob 分布式弹性作业
     * @param jobFacade 作业内部服务门面服务
     * @return 作业执行器
     */
    @SuppressWarnings("unchecked")
    public static AbstractElasticJobExecutor getJobExecutor(final ElasticJob elasticJob, final JobFacade jobFacade) {
        if (null == elasticJob) {
            return new ScriptJobExecutor(jobFacade);
        }
        if (elasticJob instanceof SimpleJob) {
            return new SimpleJobExecutor((SimpleJob) elasticJob, jobFacade);
        }
        if (elasticJob instanceof DataflowJob) {
            return new DataflowJobExecutor((DataflowJob) elasticJob, jobFacade);
        }
        throw new JobConfigurationException("Cannot support job type '%s'", elasticJob.getClass().getCanonicalName());
    }
}

执行作业

//io.elasticjob.lite.executor.AbstractElasticJobExecutor
/**
 * 执行作业.
 */
public final void execute() {
    // 检查 作业执行环境
    try {
        jobFacade.checkJobExecutionEnvironment();
    } catch (final JobExecutionEnvironmentException cause) {
        jobExceptionHandler.handleException(jobName, cause);
    }
    // 获取 当前作业服务器的分片上下文
    ShardingContexts shardingContexts = jobFacade.getShardingContexts();
    // 发布作业状态追踪事件(State.TASK_STAGING)
    if (shardingContexts.isAllowSendJobEvent()) {
        jobFacade.postJobStatusTraceEvent(shardingContexts.getTaskId(), State.TASK_STAGING, String.format("Job '%s' execute begin.", jobName));
    }
    // 跳过 存在运行中的被错过作业
    if (jobFacade.misfireIfRunning(shardingContexts.getShardingItemParameters().keySet())) {
        // 发布作业状态追踪事件(State.TASK_FINISHED)
        if (shardingContexts.isAllowSendJobEvent()) {
            jobFacade.postJobStatusTraceEvent(shardingContexts.getTaskId(), State.TASK_FINISHED, String.format(
                    "Previous job '%s' - shardingItems '%s' is still running, misfired job will start after previous job completed.", jobName, 
                    shardingContexts.getShardingItemParameters().keySet()));
        }
        return;
    }
    // 执行 作业执行前的方法
    try {
        jobFacade.beforeJobExecuted(shardingContexts);
        //CHECKSTYLE:OFF
    } catch (final Throwable cause) {
        //CHECKSTYLE:ON
        jobExceptionHandler.handleException(jobName, cause);
    }
    // 执行 普通触发的作业
    execute(shardingContexts, JobExecutionEvent.ExecutionSource.NORMAL_TRIGGER);
    // 执行 被跳过触发的作业
    while (jobFacade.isExecuteMisfired(shardingContexts.getShardingItemParameters().keySet())) {
        jobFacade.clearMisfire(shardingContexts.getShardingItemParameters().keySet());
        execute(shardingContexts, JobExecutionEvent.ExecutionSource.MISFIRE);
    }
    // 执行 作业失效转移
    jobFacade.failoverIfNecessary();
    // 执行 作业执行后的方法
    try {
        jobFacade.afterJobExecuted(shardingContexts);
        //CHECKSTYLE:OFF
    } catch (final Throwable cause) {
        //CHECKSTYLE:ON
        jobExceptionHandler.handleException(jobName, cause);
    }
}

//io.elasticjob.lite.executor.AbstractElasticJobExecutor
private void execute(final ShardingContexts shardingContexts, final JobExecutionEvent.ExecutionSource executionSource) {
    // 无可执行的分片，发布作业状态追踪事件
    if (shardingContexts.getShardingItemParameters().isEmpty()) {
        if (shardingContexts.isAllowSendJobEvent()) {
            jobFacade.postJobStatusTraceEvent(shardingContexts.getTaskId(), State.TASK_FINISHED, String.format("Sharding item for job '%s' is empty.", jobName));
        }
        return;
    }
    // 注册作业启动信息
    jobFacade.registerJobBegin(shardingContexts);
    // 发布作业状态追踪事件(State.TASK_RUNNING)
    String taskId = shardingContexts.getTaskId();
    if (shardingContexts.isAllowSendJobEvent()) {
        jobFacade.postJobStatusTraceEvent(taskId, State.TASK_RUNNING, "");
    }
    try {
        //执行作业
        process(shardingContexts, executionSource);
    } finally {
        // TODO 考虑增加作业失败的状态，并且考虑如何处理作业失败的整体回路
        // 注册作业完成信息
        jobFacade.registerJobCompleted(shardingContexts);
        // 根据是否有异常，发布作业状态追踪事件
        if (itemErrorMessages.isEmpty()) {
            if (shardingContexts.isAllowSendJobEvent()) {
                jobFacade.postJobStatusTraceEvent(taskId, State.TASK_FINISHED, "");
            }
        } else {
            if (shardingContexts.isAllowSendJobEvent()) {
                jobFacade.postJobStatusTraceEvent(taskId, State.TASK_ERROR, itemErrorMessages.toString());
            }
        }
    }
}

//io.elasticjob.lite.executor.AbstractElasticJobExecutor
/**
 *  执行作业
 * @param shardingContexts 分片上下文集合.
 * @param executionSource 执行来源
 */
private void process(final ShardingContexts shardingContexts, final JobExecutionEvent.ExecutionSource executionSource) {
    Collection<Integer> items = shardingContexts.getShardingItemParameters().keySet();
    // 1个分片，直接执行
    if (1 == items.size()) {
        int item = shardingContexts.getShardingItemParameters().keySet().iterator().next();
        JobExecutionEvent jobExecutionEvent = new JobExecutionEvent(shardingContexts.getTaskId(), jobName, executionSource, item);
        // 执行一个作业
        process(shardingContexts, item, jobExecutionEvent);
        return;
    }
    // 多分片，并行执行
    final CountDownLatch latch = new CountDownLatch(items.size());
    for (final int each : items) {
        final JobExecutionEvent jobExecutionEvent = new JobExecutionEvent(shardingContexts.getTaskId(), jobName, executionSource, each);
        if (executorService.isShutdown()) {
            return;
        }
        //作业执行线程池
        executorService.submit(new Runnable() {

            @Override
            public void run() {
                try {
                    // 执行一个作业
                    process(shardingContexts, each, jobExecutionEvent);
                } finally {
                    latch.countDown();
                }
            }
        });
    }
    // 等待多分片全部完成
    try {
        latch.await();
    } catch (final InterruptedException ex) {
        Thread.currentThread().interrupt();
    }
}

private void process(final ShardingContexts shardingContexts, final int item, final JobExecutionEvent startEvent) {
    // 发布执行事件(开始)
    if (shardingContexts.isAllowSendJobEvent()) {
        jobFacade.postJobExecutionEvent(startEvent);
    }
    log.trace("Job '{}' executing, item is: '{}'.", jobName, item);
    JobExecutionEvent completeEvent;
    try {
        // 执行单个作业
        process(new ShardingContext(shardingContexts, item));
        // 发布执行事件(成功)
        completeEvent = startEvent.executionSuccess();
        log.trace("Job '{}' executed, item is: '{}'.", jobName, item);
        if (shardingContexts.isAllowSendJobEvent()) {
            jobFacade.postJobExecutionEvent(completeEvent);
        }
        // CHECKSTYLE:OFF
    } catch (final Throwable cause) {
        // CHECKSTYLE:ON
        // 发布执行事件(失败)
        completeEvent = startEvent.executionFailure(cause);
        jobFacade.postJobExecutionEvent(completeEvent);
        // 设置该分片执行异常信息
        itemErrorMessages.put(item, ExceptionUtil.transform(cause));
        jobExceptionHandler.handleException(jobName, cause);
    }
}

5.1

//io.elasticjob.lite.internal.schedule.SchedulerFacade
/**
 * 注册作业启动信息.
 * 
 * @param enabled 作业是否启用
 */
public void registerStartUpInfo(final boolean enabled) {
    //开启所有监听器.
    listenerManager.startAllListeners();
    //选举主节点.
    leaderService.electLeader();
    //持久化作业服务器上线信息.
    serverService.persistOnline(enabled);
    //持久化作业运行实例上线相关信息.
    instanceService.persistOnline();
    //设置需要重新分片的标记.
    shardingService.setReshardingFlag();
    //初始化作业监听服务.
    monitorService.listen();
    //如果没运行就运行起来
    if (!reconcileService.isRunning()) {
        reconcileService.startAsync();
    }
}

6.1 最后执行调度作业

//io.elasticjob.lite.internal.schedule.SchedulerFacade
/**
 * 调度作业.
 * 
 * @param cron CRON表达式
 */
public void scheduleJob(final String cron) {
    try {
        if (!scheduler.checkExists(jobDetail.getKey())) {
            scheduler.scheduleJob(jobDetail, createTrigger(cron));
        }
        scheduler.start();
    } catch (final SchedulerException ex) {
        throw new JobSystemException(ex);
    }
}

重新分片的4种情况:

1.注册作业启动信息时

2.作业分片总数变化时

3.服务器变化时

    服务器变化有2种情况

        1.服务器被开启或禁用

        2.作业节点新增或移除

4.待补充

只有在主节点的时候才分片

执行失败后会把失败信息放进itemErrorMessages的map里面,然后下面就会判断执行是否有异常,如果有异常就重试

失效转移是通过监听器,监听zk的节点变化来执行的

通过FailoverListenerManager这个失效转移监听管理器来监听zk的节点

 

最后附上执行任务流程图