Examples with KeyedSerializationSchemaWrapper org.apache.flink.streaming.util.serialization.KeyedSerializationSchemaWrapper used on opensource projects

Example 1 with KeyedSerializationSchemaWrapper

use of org.apache.flink.streaming.util.serialization.KeyedSerializationSchemaWrapper in project flink by apache.

the class KafkaProducerTestBase method runCustomPartitioningTest.

/**
	 * 
	 * <pre>
	 *             +------> (sink) --+--> [KAFKA-1] --> (source) -> (map) --+
	 *            /                  |                                       \
	 *           /                   |                                        \
	 * (source) ----------> (sink) --+--> [KAFKA-2] --> (source) -> (map) -----+-> (sink)
	 *           \                   |                                        /
	 *            \                  |                                       /
	 *             +------> (sink) --+--> [KAFKA-3] --> (source) -> (map) --+
	 * </pre>
	 * 
	 * The mapper validates that the values come consistently from the correct Kafka partition.
	 * 
	 * The final sink validates that there are no duplicates and that all partitions are present.
	 */
public void runCustomPartitioningTest() {
    try {
        LOG.info("Starting KafkaProducerITCase.testCustomPartitioning()");
        final String topic = "customPartitioningTestTopic";
        final int parallelism = 3;
        createTestTopic(topic, parallelism, 1);
        TypeInformation<Tuple2<Long, String>> longStringInfo = TypeInfoParser.parse("Tuple2<Long, String>");
        StreamExecutionEnvironment env = StreamExecutionEnvironment.createRemoteEnvironment("localhost", flinkPort);
        env.setRestartStrategy(RestartStrategies.noRestart());
        env.getConfig().disableSysoutLogging();
        TypeInformationSerializationSchema<Tuple2<Long, String>> serSchema = new TypeInformationSerializationSchema<>(longStringInfo, env.getConfig());
        TypeInformationSerializationSchema<Tuple2<Long, String>> deserSchema = new TypeInformationSerializationSchema<>(longStringInfo, env.getConfig());
        // ------ producing topology ---------
        // source has DOP 1 to make sure it generates no duplicates
        DataStream<Tuple2<Long, String>> stream = env.addSource(new SourceFunction<Tuple2<Long, String>>() {

            private boolean running = true;

            @Override
            public void run(SourceContext<Tuple2<Long, String>> ctx) throws Exception {
                long cnt = 0;
                while (running) {
                    ctx.collect(new Tuple2<Long, String>(cnt, "kafka-" + cnt));
                    cnt++;
                }
            }

            @Override
            public void cancel() {
                running = false;
            }
        }).setParallelism(1);
        Properties props = new Properties();
        props.putAll(FlinkKafkaProducerBase.getPropertiesFromBrokerList(brokerConnectionStrings));
        props.putAll(secureProps);
        // sink partitions into 
        kafkaServer.produceIntoKafka(stream, topic, new KeyedSerializationSchemaWrapper<>(serSchema), props, new CustomPartitioner(parallelism)).setParallelism(parallelism);
        // ------ consuming topology ---------
        Properties consumerProps = new Properties();
        consumerProps.putAll(standardProps);
        consumerProps.putAll(secureProps);
        FlinkKafkaConsumerBase<Tuple2<Long, String>> source = kafkaServer.getConsumer(topic, deserSchema, consumerProps);
        env.addSource(source).setParallelism(parallelism).map(new RichMapFunction<Tuple2<Long, String>, Integer>() {

            private int ourPartition = -1;

            @Override
            public Integer map(Tuple2<Long, String> value) {
                int partition = value.f0.intValue() % parallelism;
                if (ourPartition != -1) {
                    assertEquals("inconsistent partitioning", ourPartition, partition);
                } else {
                    ourPartition = partition;
                }
                return partition;
            }
        }).setParallelism(parallelism).addSink(new SinkFunction<Integer>() {

            private int[] valuesPerPartition = new int[parallelism];

            @Override
            public void invoke(Integer value) throws Exception {
                valuesPerPartition[value]++;
                boolean missing = false;
                for (int i : valuesPerPartition) {
                    if (i < 100) {
                        missing = true;
                        break;
                    }
                }
                if (!missing) {
                    throw new SuccessException();
                }
            }
        }).setParallelism(1);
        tryExecute(env, "custom partitioning test");
        deleteTestTopic(topic);
        LOG.info("Finished KafkaProducerITCase.testCustomPartitioning()");
    } catch (Exception e) {
        e.printStackTrace();
        fail(e.getMessage());
    }
}

Example 2 with KeyedSerializationSchemaWrapper

use of org.apache.flink.streaming.util.serialization.KeyedSerializationSchemaWrapper in project flink by apache.

the class KafkaConsumerTestBase method writeSequence.

protected String writeSequence(String baseTopicName, final int numElements, final int parallelism, final int replicationFactor) throws Exception {
    LOG.info("\n===================================\n" + "== Writing sequence of " + numElements + " into " + baseTopicName + " with p=" + parallelism + "\n" + "===================================");
    final TypeInformation<Tuple2<Integer, Integer>> resultType = TypeInformation.of(new TypeHint<Tuple2<Integer, Integer>>() {
    });
    final KeyedSerializationSchema<Tuple2<Integer, Integer>> serSchema = new KeyedSerializationSchemaWrapper<>(new TypeInformationSerializationSchema<>(resultType, new ExecutionConfig()));
    final KeyedDeserializationSchema<Tuple2<Integer, Integer>> deserSchema = new KeyedDeserializationSchemaWrapper<>(new TypeInformationSerializationSchema<>(resultType, new ExecutionConfig()));
    final int maxNumAttempts = 10;
    for (int attempt = 1; attempt <= maxNumAttempts; attempt++) {
        final String topicName = baseTopicName + '-' + attempt;
        LOG.info("Writing attempt #1");
        // -------- Write the Sequence --------
        createTestTopic(topicName, parallelism, replicationFactor);
        StreamExecutionEnvironment writeEnv = StreamExecutionEnvironment.createRemoteEnvironment("localhost", flinkPort);
        writeEnv.getConfig().setRestartStrategy(RestartStrategies.noRestart());
        writeEnv.getConfig().disableSysoutLogging();
        DataStream<Tuple2<Integer, Integer>> stream = writeEnv.addSource(new RichParallelSourceFunction<Tuple2<Integer, Integer>>() {

            private boolean running = true;

            @Override
            public void run(SourceContext<Tuple2<Integer, Integer>> ctx) throws Exception {
                int cnt = 0;
                int partition = getRuntimeContext().getIndexOfThisSubtask();
                while (running && cnt < numElements) {
                    ctx.collect(new Tuple2<>(partition, cnt));
                    cnt++;
                }
            }

            @Override
            public void cancel() {
                running = false;
            }
        }).setParallelism(parallelism);
        // the producer must not produce duplicates
        Properties producerProperties = FlinkKafkaProducerBase.getPropertiesFromBrokerList(brokerConnectionStrings);
        producerProperties.setProperty("retries", "0");
        producerProperties.putAll(secureProps);
        kafkaServer.produceIntoKafka(stream, topicName, serSchema, producerProperties, new Tuple2Partitioner(parallelism)).setParallelism(parallelism);
        try {
            writeEnv.execute("Write sequence");
        } catch (Exception e) {
            LOG.error("Write attempt failed, trying again", e);
            deleteTestTopic(topicName);
            JobManagerCommunicationUtils.waitUntilNoJobIsRunning(flink.getLeaderGateway(timeout));
            continue;
        }
        LOG.info("Finished writing sequence");
        // -------- Validate the Sequence --------
        // we need to validate the sequence, because kafka's producers are not exactly once
        LOG.info("Validating sequence");
        JobManagerCommunicationUtils.waitUntilNoJobIsRunning(flink.getLeaderGateway(timeout));
        final StreamExecutionEnvironment readEnv = StreamExecutionEnvironment.createRemoteEnvironment("localhost", flinkPort);
        readEnv.getConfig().setRestartStrategy(RestartStrategies.noRestart());
        readEnv.getConfig().disableSysoutLogging();
        readEnv.setParallelism(parallelism);
        Properties readProps = (Properties) standardProps.clone();
        readProps.setProperty("group.id", "flink-tests-validator");
        readProps.putAll(secureProps);
        FlinkKafkaConsumerBase<Tuple2<Integer, Integer>> consumer = kafkaServer.getConsumer(topicName, deserSchema, readProps);
        readEnv.addSource(consumer).map(new RichMapFunction<Tuple2<Integer, Integer>, Tuple2<Integer, Integer>>() {

            private final int totalCount = parallelism * numElements;

            private int count = 0;

            @Override
            public Tuple2<Integer, Integer> map(Tuple2<Integer, Integer> value) throws Exception {
                if (++count == totalCount) {
                    throw new SuccessException();
                } else {
                    return value;
                }
            }
        }).setParallelism(1).addSink(new DiscardingSink<Tuple2<Integer, Integer>>()).setParallelism(1);
        final AtomicReference<Throwable> errorRef = new AtomicReference<>();
        Thread runner = new Thread() {

            @Override
            public void run() {
                try {
                    tryExecute(readEnv, "sequence validation");
                } catch (Throwable t) {
                    errorRef.set(t);
                }
            }
        };
        runner.start();
        final long deadline = System.nanoTime() + 10_000_000_000L;
        long delay;
        while (runner.isAlive() && (delay = deadline - System.nanoTime()) > 0) {
            runner.join(delay / 1_000_000L);
        }
        boolean success;
        if (runner.isAlive()) {
            // did not finish in time, maybe the producer dropped one or more records and
            // the validation did not reach the exit point
            success = false;
            JobManagerCommunicationUtils.cancelCurrentJob(flink.getLeaderGateway(timeout));
        } else {
            Throwable error = errorRef.get();
            if (error != null) {
                success = false;
                LOG.info("Attempt " + attempt + " failed with exception", error);
            } else {
                success = true;
            }
        }
        JobManagerCommunicationUtils.waitUntilNoJobIsRunning(flink.getLeaderGateway(timeout));
        if (success) {
            // everything is good!
            return topicName;
        } else {
            deleteTestTopic(topicName);
        // fall through the loop
        }
    }
    throw new Exception("Could not write a valid sequence to Kafka after " + maxNumAttempts + " attempts");
}

Example 3 with KeyedSerializationSchemaWrapper

use of org.apache.flink.streaming.util.serialization.KeyedSerializationSchemaWrapper in project flink by apache.

the class KafkaConsumerTestBase method runStartFromLatestOffsets.

/**
	 * This test ensures that when explicitly set to start from latest record, the consumer
	 * ignores the "auto.offset.reset" behaviour as well as any committed group offsets in Kafka.
	 */
public void runStartFromLatestOffsets() throws Exception {
    // 50 records written to each of 3 partitions before launching a latest-starting consuming job
    final int parallelism = 3;
    final int recordsInEachPartition = 50;
    // each partition will be written an extra 200 records
    final int extraRecordsInEachPartition = 200;
    // all already existing data in the topic, before the consuming topology has started, should be ignored
    final String topicName = writeSequence("testStartFromLatestOffsetsTopic", recordsInEachPartition, parallelism, 1);
    // the committed offsets should be ignored
    KafkaTestEnvironment.KafkaOffsetHandler kafkaOffsetHandler = kafkaServer.createOffsetHandler();
    kafkaOffsetHandler.setCommittedOffset(topicName, 0, 23);
    kafkaOffsetHandler.setCommittedOffset(topicName, 1, 31);
    kafkaOffsetHandler.setCommittedOffset(topicName, 2, 43);
    // job names for the topologies for writing and consuming the extra records
    final String consumeExtraRecordsJobName = "Consume Extra Records Job";
    final String writeExtraRecordsJobName = "Write Extra Records Job";
    // seriliazation / deserialization schemas for writing and consuming the extra records
    final TypeInformation<Tuple2<Integer, Integer>> resultType = TypeInformation.of(new TypeHint<Tuple2<Integer, Integer>>() {
    });
    final KeyedSerializationSchema<Tuple2<Integer, Integer>> serSchema = new KeyedSerializationSchemaWrapper<>(new TypeInformationSerializationSchema<>(resultType, new ExecutionConfig()));
    final KeyedDeserializationSchema<Tuple2<Integer, Integer>> deserSchema = new KeyedDeserializationSchemaWrapper<>(new TypeInformationSerializationSchema<>(resultType, new ExecutionConfig()));
    // setup and run the latest-consuming job
    final StreamExecutionEnvironment env = StreamExecutionEnvironment.createRemoteEnvironment("localhost", flinkPort);
    env.getConfig().disableSysoutLogging();
    env.setParallelism(parallelism);
    final Properties readProps = new Properties();
    readProps.putAll(standardProps);
    // this should be ignored
    readProps.setProperty("auto.offset.reset", "earliest");
    FlinkKafkaConsumerBase<Tuple2<Integer, Integer>> latestReadingConsumer = kafkaServer.getConsumer(topicName, deserSchema, readProps);
    latestReadingConsumer.setStartFromLatest();
    env.addSource(latestReadingConsumer).setParallelism(parallelism).flatMap(new FlatMapFunction<Tuple2<Integer, Integer>, Object>() {

        @Override
        public void flatMap(Tuple2<Integer, Integer> value, Collector<Object> out) throws Exception {
            if (value.f1 - recordsInEachPartition < 0) {
                throw new RuntimeException("test failed; consumed a record that was previously written: " + value);
            }
        }
    }).setParallelism(1).addSink(new DiscardingSink<>());
    final AtomicReference<Throwable> error = new AtomicReference<>();
    Thread consumeThread = new Thread(new Runnable() {

        @Override
        public void run() {
            try {
                env.execute(consumeExtraRecordsJobName);
            } catch (Throwable t) {
                if (!(t.getCause() instanceof JobCancellationException)) {
                    error.set(t);
                }
            }
        }
    });
    consumeThread.start();
    // wait until the consuming job has started, to be extra safe
    JobManagerCommunicationUtils.waitUntilJobIsRunning(flink.getLeaderGateway(timeout), consumeExtraRecordsJobName);
    // setup the extra records writing job
    final StreamExecutionEnvironment env2 = StreamExecutionEnvironment.createRemoteEnvironment("localhost", flinkPort);
    DataStream<Tuple2<Integer, Integer>> extraRecordsStream = env2.addSource(new RichParallelSourceFunction<Tuple2<Integer, Integer>>() {

        private boolean running = true;

        @Override
        public void run(SourceContext<Tuple2<Integer, Integer>> ctx) throws Exception {
            // the extra records should start from the last written value
            int count = recordsInEachPartition;
            int partition = getRuntimeContext().getIndexOfThisSubtask();
            while (running && count < recordsInEachPartition + extraRecordsInEachPartition) {
                ctx.collect(new Tuple2<>(partition, count));
                count++;
            }
        }

        @Override
        public void cancel() {
            running = false;
        }
    }).setParallelism(parallelism);
    kafkaServer.produceIntoKafka(extraRecordsStream, topicName, serSchema, readProps, null);
    try {
        env2.execute(writeExtraRecordsJobName);
    } catch (Exception e) {
        throw new RuntimeException("Writing extra records failed", e);
    }
    // cancel the consume job after all extra records are written
    JobManagerCommunicationUtils.cancelCurrentJob(flink.getLeaderGateway(timeout), consumeExtraRecordsJobName);
    consumeThread.join();
    kafkaOffsetHandler.close();
    deleteTestTopic(topicName);
    // check whether the consuming thread threw any test errors;
    // test will fail here if the consume job had incorrectly read any records other than the extra records
    final Throwable consumerError = error.get();
    if (consumerError != null) {
        throw new Exception("Exception in the consuming thread", consumerError);
    }
}