反序列化 Avro 序列化的 Kafka 流的问题

Question

I'm receiving an exception when trying to materialize a store. I am running Kafka 1.0, Confluent's Schema Registry 4.0, and Avro 1.8.2. I have generated the Pojo with Avro's maven plugin and deployed the schema to the Confluent server with the Confluent maven plugin. I am able to produce a message to the STREAM1 topic. Here's the code to setup the stream:

Properties properties = new Properties();
properties.put(StreamsConfig.APPLICATION_ID_CONFIG, "streams-pipe");
properties.put(StreamsConfig.CLIENT_ID_CONFIG, "cleant-id");
properties.put(StreamsConfig.BOOTSTRAP_SERVERS_CONFIG, "http://localhost:9092");
properties.put(AbstractKafkaAvroSerDeConfig.SCHEMA_REGISTRY_URL_CONFIG, "http://localhost:8081");
properties.put(StreamsConfig.DEFAULT_KEY_SERDE_CLASS_CONFIG, Serdes.String().getClass());
properties.put(StreamsConfig.DEFAULT_VALUE_SERDE_CLASS_CONFIG, SpecificAvroSerde.class);
StreamsBuilder builder = new StreamsBuilder();

Serde<T> pojoSerde = new SpecificAvroSerde<>();
final Map<String, String> serdeConfig = Collections.singletonMap(
        AbstractKafkaAvroSerDeConfig.SCHEMA_REGISTRY_URL_CONFIG, "http://localhost:8081");
pojoSerde.configure(serdeConfig, false);

Consumed<String, Pojo> consumed = Consumed.with(Serdes.String(), pojoSerde);
KStream<String, Pojo> source = builder.stream(TopicName.STREAM1.toString(), consumed);
KTable<String, Long> storePojoCount = source
        .groupBy((key, value) -> key)
        .count(Materialized.as(StoreName.STORE_WORD_COUNT.toString()));

Produced<String, Long> produced = Produced.with(Serdes.String(), Serdes.Long());
storePojoCount.toStream().to(TopicName.STREAM2.toString(), produced);
KafkaStreams streams = new KafkaStreams(builder.build(), properties);
Runtime.getRuntime().addShutdownHook(new Thread(streams::close));
streams.start();

which produced the following exception.

Exception in thread "cleant-id-StreamThread-2" org.apache.kafka.streams.errors.StreamsException: Deserialization exception handler is set to fail upon a deserialization error. If you would rather have the streaming pipeline continue after a deserialization error, please set the default.deserialization.exception.handler appropriately.
    at org.apache.kafka.streams.processor.internals.RecordDeserializer.deserialize(RecordDeserializer.java:74)
    at org.apache.kafka.streams.processor.internals.RecordQueue.addRawRecords(RecordQueue.java:91)
    at org.apache.kafka.streams.processor.internals.PartitionGroup.addRawRecords(PartitionGroup.java:117)
    at org.apache.kafka.streams.processor.internals.StreamTask.addRecords(StreamTask.java:546)
    at org.apache.kafka.streams.processor.internals.StreamThread.addRecordsToTasks(StreamThread.java:920)
    at org.apache.kafka.streams.processor.internals.StreamThread.runOnce(StreamThread.java:821)
    at org.apache.kafka.streams.processor.internals.StreamThread.runLoop(StreamThread.java:774)
    at org.apache.kafka.streams.processor.internals.StreamThread.run(StreamThread.java:744)
Caused by: org.apache.kafka.common.errors.SerializationException: Error deserializing Avro message for id -1
Caused by: org.apache.kafka.common.errors.SerializationException: Unknown magic byte!

How can I configure this SpecificAvroSerde to successfully deserialize the stream?

Answer 1

The problem is that the Materialized object doesn't have the appropriate deserializers - Avro is attempting to deserialize the KTable values as Avro is the default value deserializer. It is unable to do so as the KTable values are actually Longs.

Creating a the materialized object with the correct deserializer will solve the problem.

protected <K, V> Materialized<K, V, KeyValueStore<Bytes, byte[]>> persistentStore(StoreName storeName, Serde<K> keyType, Serde<V> valueType) {
    KeyValueBytesStoreSupplier storeSupplier = Stores.persistentKeyValueStore(storeName.toString());
    return Materialized.<K, V>as(storeSupplier).withKeySerde(keyType).withValueSerde(valueType);
}

Any store supplier can be used here - this is just the one which fits my needs.