Search in sources :

Example 66 with StreamExecutionEnvironment

use of org.apache.flink.streaming.api.environment.StreamExecutionEnvironment in project flink by apache.

the class LegacyCheckpointedStreamingProgram method main.

public static void main(String[] args) throws Exception {
    final String jarFile = args[0];
    final String host = args[1];
    final int port = Integer.parseInt(args[2]);
    StreamExecutionEnvironment env = StreamExecutionEnvironment.createRemoteEnvironment(host, port, jarFile);
    env.getConfig().disableSysoutLogging();
    env.enableCheckpointing(CHECKPOINT_INTERVALL);
    env.setRestartStrategy(RestartStrategies.fixedDelayRestart(1, 10000));
    env.disableOperatorChaining();
    DataStream<String> text = env.addSource(new SimpleStringGenerator());
    text.map(new StatefulMapper()).addSink(new NoOpSink());
    env.setParallelism(1);
    env.execute("Checkpointed Streaming Program");
}
Also used : StreamExecutionEnvironment(org.apache.flink.streaming.api.environment.StreamExecutionEnvironment)

Example 67 with StreamExecutionEnvironment

use of org.apache.flink.streaming.api.environment.StreamExecutionEnvironment in project flink by apache.

the class StreamingCustomInputSplitProgram method main.

public static void main(String[] args) throws Exception {
    final String jarFile = args[0];
    final String host = args[1];
    final int port = Integer.parseInt(args[2]);
    final int parallelism = Integer.parseInt(args[3]);
    Configuration config = new Configuration();
    config.setString(ConfigConstants.AKKA_ASK_TIMEOUT, "5 s");
    StreamExecutionEnvironment env = StreamExecutionEnvironment.createRemoteEnvironment(host, port, config, jarFile);
    env.getConfig().disableSysoutLogging();
    env.setParallelism(parallelism);
    DataStream<Integer> data = env.createInput(new CustomInputFormat());
    data.map(new MapFunction<Integer, Tuple2<Integer, Double>>() {

        @Override
        public Tuple2<Integer, Double> map(Integer value) throws Exception {
            return new Tuple2<Integer, Double>(value, value * 0.5);
        }
    }).addSink(new NoOpSink());
    env.execute();
}
Also used : Configuration(org.apache.flink.configuration.Configuration) Tuple2(org.apache.flink.api.java.tuple.Tuple2) StreamExecutionEnvironment(org.apache.flink.streaming.api.environment.StreamExecutionEnvironment) MapFunction(org.apache.flink.api.common.functions.MapFunction)

Example 68 with StreamExecutionEnvironment

use of org.apache.flink.streaming.api.environment.StreamExecutionEnvironment in project flink by apache.

the class FastFailuresITCase method testThis.

@Test
public void testThis() {
    Configuration config = new Configuration();
    config.setInteger(ConfigConstants.LOCAL_NUMBER_TASK_MANAGER, 2);
    config.setInteger(ConfigConstants.TASK_MANAGER_NUM_TASK_SLOTS, 2);
    LocalFlinkMiniCluster cluster = new LocalFlinkMiniCluster(config, false);
    cluster.start();
    StreamExecutionEnvironment env = StreamExecutionEnvironment.createRemoteEnvironment("localhost", cluster.getLeaderRPCPort());
    env.getConfig().disableSysoutLogging();
    env.setParallelism(4);
    env.enableCheckpointing(1000);
    env.getConfig().setRestartStrategy(RestartStrategies.fixedDelayRestart(210, 0));
    DataStream<Tuple2<Integer, Integer>> input = env.addSource(new RichSourceFunction<Tuple2<Integer, Integer>>() {

        @Override
        public void open(Configuration parameters) {
            if (FAILURES_SO_FAR.incrementAndGet() <= NUM_FAILURES) {
                throw new RuntimeException("fail");
            }
        }

        @Override
        public void run(SourceContext<Tuple2<Integer, Integer>> ctx) {
        }

        @Override
        public void cancel() {
        }
    });
    input.keyBy(0).map(new MapFunction<Tuple2<Integer, Integer>, Integer>() {

        @Override
        public Integer map(Tuple2<Integer, Integer> value) {
            return value.f0;
        }
    }).addSink(new SinkFunction<Integer>() {

        @Override
        public void invoke(Integer value) {
        }
    });
    try {
        env.execute();
    } catch (Exception e) {
        e.printStackTrace();
        fail(e.getMessage());
    }
}
Also used : Configuration(org.apache.flink.configuration.Configuration) MapFunction(org.apache.flink.api.common.functions.MapFunction) LocalFlinkMiniCluster(org.apache.flink.runtime.minicluster.LocalFlinkMiniCluster) AtomicInteger(java.util.concurrent.atomic.AtomicInteger) Tuple2(org.apache.flink.api.java.tuple.Tuple2) StreamExecutionEnvironment(org.apache.flink.streaming.api.environment.StreamExecutionEnvironment) Test(org.junit.Test)

Example 69 with StreamExecutionEnvironment

use of org.apache.flink.streaming.api.environment.StreamExecutionEnvironment in project flink by apache.

the class JobManagerHACheckpointRecoveryITCase method testCheckpointedStreamingSumProgram.

/**
	 * Simple checkpointed streaming sum.
	 *
	 * <p>The sources (Parallelism) count until sequenceEnd. The sink (1) sums up all counts and
	 * returns it to the main thread via a static variable. We wait until some checkpoints are
	 * completed and sanity check that the sources recover with an updated state to make sure that
	 * this test actually tests something.
	 */
@Test
@RetryOnFailure(times = 1)
public void testCheckpointedStreamingSumProgram() throws Exception {
    // Config
    final int checkpointingInterval = 200;
    final int sequenceEnd = 5000;
    final long expectedSum = Parallelism * sequenceEnd * (sequenceEnd + 1) / 2;
    final StreamExecutionEnvironment env = StreamExecutionEnvironment.createLocalEnvironment();
    env.setParallelism(Parallelism);
    env.enableCheckpointing(checkpointingInterval);
    env.addSource(new CheckpointedSequenceSource(sequenceEnd)).addSink(new CountingSink()).setParallelism(1);
    JobGraph jobGraph = env.getStreamGraph().getJobGraph();
    Configuration config = ZooKeeperTestUtils.createZooKeeperHAConfig(ZooKeeper.getConnectString(), FileStateBackendBasePath.getAbsoluteFile().toURI().toString());
    config.setInteger(ConfigConstants.TASK_MANAGER_NUM_TASK_SLOTS, Parallelism);
    ActorSystem testSystem = null;
    final JobManagerProcess[] jobManagerProcess = new JobManagerProcess[2];
    LeaderRetrievalService leaderRetrievalService = null;
    ActorSystem taskManagerSystem = null;
    try {
        final Deadline deadline = TestTimeOut.fromNow();
        // Test actor system
        testSystem = AkkaUtils.createActorSystem(new Configuration(), new Some<>(new Tuple2<String, Object>("localhost", 0)));
        // The job managers
        jobManagerProcess[0] = new JobManagerProcess(0, config);
        jobManagerProcess[1] = new JobManagerProcess(1, config);
        jobManagerProcess[0].startProcess();
        jobManagerProcess[1].startProcess();
        // Leader listener
        TestingListener leaderListener = new TestingListener();
        leaderRetrievalService = ZooKeeperUtils.createLeaderRetrievalService(config);
        leaderRetrievalService.start(leaderListener);
        // The task manager
        taskManagerSystem = AkkaUtils.createActorSystem(config, Option.apply(new Tuple2<String, Object>("localhost", 0)));
        TaskManager.startTaskManagerComponentsAndActor(config, ResourceID.generate(), taskManagerSystem, "localhost", Option.<String>empty(), Option.<LeaderRetrievalService>empty(), false, TaskManager.class);
        {
            // Initial submission
            leaderListener.waitForNewLeader(deadline.timeLeft().toMillis());
            String leaderAddress = leaderListener.getAddress();
            UUID leaderId = leaderListener.getLeaderSessionID();
            // Get the leader ref
            ActorRef leaderRef = AkkaUtils.getActorRef(leaderAddress, testSystem, deadline.timeLeft());
            ActorGateway leader = new AkkaActorGateway(leaderRef, leaderId);
            // Submit the job in detached mode
            leader.tell(new SubmitJob(jobGraph, ListeningBehaviour.DETACHED));
            JobManagerActorTestUtils.waitForJobStatus(jobGraph.getJobID(), JobStatus.RUNNING, leader, deadline.timeLeft());
        }
        // Who's the boss?
        JobManagerProcess leadingJobManagerProcess;
        if (jobManagerProcess[0].getJobManagerAkkaURL(deadline.timeLeft()).equals(leaderListener.getAddress())) {
            leadingJobManagerProcess = jobManagerProcess[0];
        } else {
            leadingJobManagerProcess = jobManagerProcess[1];
        }
        CompletedCheckpointsLatch.await();
        // Kill the leading job manager process
        leadingJobManagerProcess.destroy();
        {
            // Recovery by the standby JobManager
            leaderListener.waitForNewLeader(deadline.timeLeft().toMillis());
            String leaderAddress = leaderListener.getAddress();
            UUID leaderId = leaderListener.getLeaderSessionID();
            ActorRef leaderRef = AkkaUtils.getActorRef(leaderAddress, testSystem, deadline.timeLeft());
            ActorGateway leader = new AkkaActorGateway(leaderRef, leaderId);
            JobManagerActorTestUtils.waitForJobStatus(jobGraph.getJobID(), JobStatus.RUNNING, leader, deadline.timeLeft());
        }
        // Wait to finish
        FinalCountLatch.await();
        assertEquals(expectedSum, (long) FinalCount.get());
        for (int i = 0; i < Parallelism; i++) {
            assertNotEquals(0, RecoveredStates.get(i));
        }
    } catch (Throwable t) {
        // Reset all static state for test retries
        CompletedCheckpointsLatch = new CountDownLatch(2);
        RecoveredStates = new AtomicLongArray(Parallelism);
        FinalCountLatch = new CountDownLatch(1);
        FinalCount = new AtomicReference<>();
        LastElement = -1;
        // Print early (in some situations the process logs get too big
        // for Travis and the root problem is not shown)
        t.printStackTrace();
        // In case of an error, print the job manager process logs.
        if (jobManagerProcess[0] != null) {
            jobManagerProcess[0].printProcessLog();
        }
        if (jobManagerProcess[1] != null) {
            jobManagerProcess[1].printProcessLog();
        }
        throw t;
    } finally {
        if (jobManagerProcess[0] != null) {
            jobManagerProcess[0].destroy();
        }
        if (jobManagerProcess[1] != null) {
            jobManagerProcess[1].destroy();
        }
        if (leaderRetrievalService != null) {
            leaderRetrievalService.stop();
        }
        if (taskManagerSystem != null) {
            taskManagerSystem.shutdown();
        }
        if (testSystem != null) {
            testSystem.shutdown();
        }
    }
}
Also used : ActorSystem(akka.actor.ActorSystem) AkkaActorGateway(org.apache.flink.runtime.instance.AkkaActorGateway) Configuration(org.apache.flink.configuration.Configuration) ActorRef(akka.actor.ActorRef) TestingListener(org.apache.flink.runtime.leaderelection.TestingListener) ActorGateway(org.apache.flink.runtime.instance.ActorGateway) AkkaActorGateway(org.apache.flink.runtime.instance.AkkaActorGateway) UUID(java.util.UUID) SubmitJob(org.apache.flink.runtime.messages.JobManagerMessages.SubmitJob) Deadline(scala.concurrent.duration.Deadline) AtomicReference(java.util.concurrent.atomic.AtomicReference) CountDownLatch(java.util.concurrent.CountDownLatch) JobGraph(org.apache.flink.runtime.jobgraph.JobGraph) Some(scala.Some) LeaderRetrievalService(org.apache.flink.runtime.leaderretrieval.LeaderRetrievalService) JobManagerProcess(org.apache.flink.runtime.testutils.JobManagerProcess) AtomicLongArray(java.util.concurrent.atomic.AtomicLongArray) StreamExecutionEnvironment(org.apache.flink.streaming.api.environment.StreamExecutionEnvironment) Test(org.junit.Test) RetryOnFailure(org.apache.flink.testutils.junit.RetryOnFailure)

Example 70 with StreamExecutionEnvironment

use of org.apache.flink.streaming.api.environment.StreamExecutionEnvironment in project flink by apache.

the class TimestampITCase method testErrorOnEventTimeOverProcessingTime.

@Test
public void testErrorOnEventTimeOverProcessingTime() {
    StreamExecutionEnvironment env = StreamExecutionEnvironment.createRemoteEnvironment("localhost", cluster.getLeaderRPCPort());
    env.setParallelism(2);
    env.getConfig().disableSysoutLogging();
    env.setStreamTimeCharacteristic(TimeCharacteristic.ProcessingTime);
    DataStream<Tuple2<String, Integer>> source1 = env.fromElements(new Tuple2<>("a", 1), new Tuple2<>("b", 2));
    source1.keyBy(0).window(TumblingEventTimeWindows.of(Time.seconds(5))).reduce(new ReduceFunction<Tuple2<String, Integer>>() {

        @Override
        public Tuple2<String, Integer> reduce(Tuple2<String, Integer> value1, Tuple2<String, Integer> value2) {
            return value1;
        }
    }).print();
    try {
        env.execute();
        fail("this should fail with an exception");
    } catch (Exception e) {
    // expected
    }
}
Also used : Tuple2(org.apache.flink.api.java.tuple.Tuple2) ReduceFunction(org.apache.flink.api.common.functions.ReduceFunction) StreamExecutionEnvironment(org.apache.flink.streaming.api.environment.StreamExecutionEnvironment) Test(org.junit.Test)

Aggregations

StreamExecutionEnvironment (org.apache.flink.streaming.api.environment.StreamExecutionEnvironment)383 Test (org.junit.Test)286 Tuple2 (org.apache.flink.api.java.tuple.Tuple2)192 OneInputTransformation (org.apache.flink.streaming.api.transformations.OneInputTransformation)81 TimeWindow (org.apache.flink.streaming.api.windowing.windows.TimeWindow)75 Tuple3 (org.apache.flink.api.java.tuple.Tuple3)48 EventTimeTrigger (org.apache.flink.streaming.api.windowing.triggers.EventTimeTrigger)42 JobGraph (org.apache.flink.runtime.jobgraph.JobGraph)34 Properties (java.util.Properties)32 ListStateDescriptor (org.apache.flink.api.common.state.ListStateDescriptor)31 TumblingEventTimeWindows (org.apache.flink.streaming.api.windowing.assigners.TumblingEventTimeWindows)30 TypeHint (org.apache.flink.api.common.typeinfo.TypeHint)27 SuccessException (org.apache.flink.test.util.SuccessException)27 IOException (java.io.IOException)24 Configuration (org.apache.flink.configuration.Configuration)24 SlidingEventTimeWindows (org.apache.flink.streaming.api.windowing.assigners.SlidingEventTimeWindows)24 KeySelector (org.apache.flink.api.java.functions.KeySelector)22 ProcessingTimeTrigger (org.apache.flink.streaming.api.windowing.triggers.ProcessingTimeTrigger)21 ReducingStateDescriptor (org.apache.flink.api.common.state.ReducingStateDescriptor)20 MapFunction (org.apache.flink.api.common.functions.MapFunction)19