flink-start源码

jobSubmit

testCse

java 复制代码
final StreamExecutionEnvironment env = new StreamExecutionEnvironment(configuration);
//将算子添加进transformArrayList中
        env.fromCollection(Collections.singletonList(42))
            .addSink(new DiscardingSink<>());
        return env.execute();

StreamExecutionEnvironment

构建streamGraph(进行相关配置初始化,并未真正进行构建)并异步执行

java 复制代码
    public JobExecutionResult execute(String jobName) throws Exception {
        Preconditions.checkNotNull(jobName, "Streaming Job name should not be null.");
        final StreamGraph streamGraph = getStreamGraph();
        streamGraph.setJobName(jobName);
        return execute(streamGraph);
    }

public JobExecutionResult execute(StreamGraph streamGraph) throws Exception {
    //异步执行streamGraph
     final JobClient jobClient = executeAsync(streamGraph);

            try {
                final JobExecutionResult jobExecutionResult;

                if (configuration.getBoolean(DeploymentOptions.ATTACHED)) {
                    jobExecutionResult = jobClient.getJobExecutionResult().get();
                } else {
                    jobExecutionResult = new DetachedJobExecutionResult(jobClient.getJobID());
                }

                jobListeners.forEach(
                        jobListener -> jobListener.onJobExecuted(jobExecutionResult, null));

                return jobExecutionResult;
            } catch (Throwable t) {
                ...
        }
    
    
    public JobClient executeAsync(StreamGraph streamGraph) throws Exception {
            final PipelineExecutorFactory executorFactory =
                executorServiceLoader.getExecutorFactory(configuration);


        CompletableFuture<JobClient> jobClientFuture =
                executorFactory
            //获取执行器,具体实现类如下图
                        .getExecutor(configuration)
            //实例化执行器并执行
                        .execute(streamGraph, configuration, userClassloader);

        try {
            JobClient jobClient = jobClientFuture.get();
            //提交成功
            jobListeners.forEach(jobListener -> jobListener.onJobSubmitted(jobClient, null));
            return jobClient;
        } catch (ExecutionException executionException) {
            final Throwable strippedException =
                    ExceptionUtils.stripExecutionException(executionException);
            jobListeners.forEach(
                //提交失败
                    jobListener -> jobListener.onJobSubmitted(null, strippedException));

            throw new FlinkException(
                    String.format("Failed to execute job '%s'.", streamGraph.getJobName()),
                    strippedException);
        }
    }

executor实现:

LocalExecutorFactory

java 复制代码
public class LocalExecutorFactory implements PipelineExecutorFactory {


@Override
public PipelineExecutor getExecutor(final Configuration configuration) {
    //创建执行器
    return LocalExecutor.create(configuration);
}
 ...   
}

LocalExecutor

java 复制代码
public static LocalExecutor create(Configuration configuration) {
//创建cluster
    return new LocalExecutor(configuration, MiniCluster::new);
}

LocalExecutor

java 复制代码
    @Override
    public CompletableFuture<JobClient> execute(
            Pipeline pipeline, Configuration configuration, ClassLoader userCodeClassloader)
            throws Exception {
        checkNotNull(pipeline);
        checkNotNull(configuration);

        Configuration effectiveConfig = new Configuration();
        effectiveConfig.addAll(this.configuration);
        effectiveConfig.addAll(configuration);

        // we only support attached execution with the local executor.
        checkState(configuration.getBoolean(DeploymentOptions.ATTACHED));

        final JobGraph jobGraph = getJobGraph(pipeline, effectiveConfig);

        return PerJobMiniClusterFactory.createWithFactory(effectiveConfig, miniClusterFactory)
            //提交作业
                .submitJob(jobGraph, userCodeClassloader);
    }

 /** Starts a {@link MiniCluster} and submits a job. */
    public CompletableFuture<JobClient> submitJob(
            JobGraph jobGraph, ClassLoader userCodeClassloader) throws Exception {
        MiniClusterConfiguration miniClusterConfig =
                getMiniClusterConfig(jobGraph.getMaximumParallelism());
        MiniCluster miniCluster = miniClusterFactory.apply(miniClusterConfig);
        
        //启动cluster
        miniCluster.start();

        return miniCluster
                .submitJob(jobGraph)
                .thenApplyAsync(
                        FunctionUtils.uncheckedFunction(
                                submissionResult -> {
                                    org.apache.flink.client.ClientUtils
                                            .waitUntilJobInitializationFinished(
                                                    () ->
                                                            miniCluster
                                                                    .getJobStatus(
                                                                            submissionResult
                                                                                    .getJobID())
                                                                    .get(),
                                                    () ->
                                                            miniCluster
                                                                    .requestJobResult(
                                                                            submissionResult
                                                                                    .getJobID())
                                                                    .get(),
                                                    userCodeClassloader);
                                    return submissionResult;
                                }))
                .thenApply(
                        result ->
                                new MiniClusterJobClient(
                                        result.getJobID(),
                                        miniCluster,
                                        userCodeClassloader,
                                        MiniClusterJobClient.JobFinalizationBehavior
                                                .SHUTDOWN_CLUSTER))
                .whenComplete(
                        (ignored, throwable) -> {
                            if (throwable != null) {
                                // We failed to create the JobClient and must shutdown to ensure
                                // cleanup.
                                shutDownCluster(miniCluster);
                            }
                        })
                .thenApply(Function.identity());
    }

MiniCluster

java 复制代码
public CompletableFuture<JobSubmissionResult> submitJob(JobGraph jobGraph) {
    final CompletableFuture<DispatcherGateway> dispatcherGatewayFuture =
            getDispatcherGatewayFuture();
    final CompletableFuture<InetSocketAddress> blobServerAddressFuture =
            createBlobServerAddress(dispatcherGatewayFuture);
    final CompletableFuture<Void> jarUploadFuture =
            uploadAndSetJobFiles(blobServerAddressFuture, jobGraph);
    final CompletableFuture<Acknowledge> acknowledgeCompletableFuture =
            jarUploadFuture
                    .thenCombine(
                            dispatcherGatewayFuture,
                            (Void ack, DispatcherGateway dispatcherGateway) ->
                                    dispatcherGateway.submitJob(jobGraph, rpcTimeout))
                    .thenCompose(Function.identity());
    return acknowledgeCompletableFuture.thenApply(
            (Acknowledge ignored) -> new JobSubmissionResult(jobGraph.getJobID()));
}
相关推荐
AnalogElectronic6 分钟前
拉多买菜项目报告
大数据·人工智能
智能化咨询7 分钟前
(199页PPT)DG企业架构企业IT战略规划架构设计方案(附下载方式)
大数据·架构
亚林瓜子7 分钟前
AWS Catalog中数据搬到Catalog中
大数据·python·spark·云计算·aws·pyspark·glue
金融小师妹11 分钟前
AI政策函数重构视角:凯文·沃什听证前信号释放与联储独立性再定价
大数据·人工智能·深度学习·能源
一生了无挂14 分钟前
Python大数据可视化:基于大数据技术的共享单车数据分析与辅助管理系统_flask+hadoop+spider
大数据·python·信息可视化
Dxy123931021615 分钟前
Elasticsearch 8 如何进行二维矩阵向量搜索
大数据·elasticsearch·矩阵
hf20001226 分钟前
Apache Iceberg vs Apache Paimon :数据湖表格式深度对比与选型指南
大数据·spark·数据湖·湖仓一体·lakehouse
隐于花海,等待花开43 分钟前
1.CONCAT / CONCAT_WS 函数深度解析
大数据·hive
CORNERSTONE36544 分钟前
生产管理六要素(PQCDSM)
大数据·运维·人工智能·生产管理