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
Hello.Reader13 分钟前
Flink Balanced Tasks Scheduling:并行度不一致时,怎么把 TaskManager “压得更均匀”
大数据·flink
爱吃羊的老虎1 小时前
【大模型应用】入门了解AI Agent
大数据·人工智能
JZC_xiaozhong1 小时前
什么是跨系统流程自动化?
大数据·运维·bpm·数据集成与应用集成·业务流程管理
海兰2 小时前
Elasticsearch 9.3.0 日志分类功能完整指南
大数据·elasticsearch·分类
是做服装的同学2 小时前
服装企业生产管理ERP系统的主要功能是什么?
大数据·经验分享·其他
Elastic 中国社区官方博客2 小时前
Elasticsearch:创建 tavily 网页搜索 workflow 及在 agent 中使用它
大数据·数据库·人工智能·elasticsearch·搜索引擎·ai·全文检索
Hello.Reader2 小时前
Flink 日志怎么用从“找得到”到“看得懂”,再到“可检索可关联”
大数据·flink
产品经理邹继强2 小时前
VTC财务与投资篇③:投资决策重构——用VTC穿透财务预测的迷雾
大数据·人工智能·重构
岁岁种桃花儿2 小时前
Flink从入门到上天系列第六篇:Flink运行时架构
大数据·flink
byte轻骑兵2 小时前
大数据场景时序数据库选型指南——Apache IoTDB实践与解析
大数据·数据库·apache·时序数据库·iotdb