我们的卡夫卡主题之一存在问题，DefaultKafkaConsumerFactory＆amp; ConcurrentMessageListenerContainer组合描述了here与工厂使用的JsonDeserializer。不幸的是，有人热情并且在主题上发布了一些无效的消息。似乎spring-kafka默默地无法处理这些消息中的第一条消息。是否有可能让spring-kafka记录错误并继续？查看记录的错误消息，似乎Apache kafka-clients库应该处理这样的情况：当迭代一批消息时，它们中的一个或多个可能无法解析？
以下代码是说明此问题的示例测试用例：
import org.apache.kafka.clients.consumer.ConsumerConfig;
import org.apache.kafka.clients.consumer.ConsumerRecord;
import org.apache.kafka.clients.producer.ProducerConfig;
import org.apache.kafka.common.serialization.Serializer;
import org.apache.kafka.common.serialization.StringDeserializer;
import org.apache.kafka.common.serialization.StringSerializer;
import org.junit.ClassRule;
import org.junit.Test;
import org.springframework.kafka.core.DefaultKafkaConsumerFactory;
import org.springframework.kafka.core.DefaultKafkaProducerFactory;
import org.springframework.kafka.core.KafkaTemplate;
import org.springframework.kafka.listener.KafkaMessageListenerContainer;
import org.springframework.kafka.listener.MessageListener;
import org.springframework.kafka.listener.config.ContainerProperties;
import org.springframework.kafka.support.SendResult;
import org.springframework.kafka.support.serializer.JsonDeserializer;
import org.springframework.kafka.support.serializer.JsonSerializer;
import org.springframework.kafka.test.rule.KafkaEmbedded;
import org.springframework.kafka.test.utils.ContainerTestUtils;
import org.springframework.util.concurrent.ListenableFuture;
import java.util.HashMap;
import java.util.Map;
import java.util.Objects;
import java.util.concurrent.BlockingQueue;
import java.util.concurrent.ExecutionException;
import java.util.concurrent.LinkedBlockingQueue;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.TimeoutException;
import static org.junit.Assert.assertEquals;
import static org.junit.Assert.assertThat;
import static org.springframework.kafka.test.hamcrest.KafkaMatchers.hasKey;
import static org.springframework.kafka.test.hamcrest.KafkaMatchers.hasValue;
/**
* @author jfreedman
*/
public class TestSpringKafka {
private static final String TOPIC1 = "spring.kafka.1.t";
@ClassRule
public static KafkaEmbedded embeddedKafka = new KafkaEmbedded(1, true, 1, TOPIC1);
@Test
public void submitMessageThenGarbageThenAnotherMessage() throws Exception {
final BlockingQueue<ConsumerRecord<String, JsonObject>> records = createListener(TOPIC1);
final KafkaTemplate<String, JsonObject> objectTemplate = createPublisher("json", new JsonSerializer<JsonObject>());
sendAndVerifyMessage(records, objectTemplate, "foo", new JsonObject("foo"), 0L);
// push some garbage text to Kafka which cannot be marshalled, this should not interrupt processing
final KafkaTemplate<String, String> garbageTemplate = createPublisher("garbage", new StringSerializer());
final SendResult<String, String> garbageResult = garbageTemplate.send(TOPIC1, "bar","bar").get(5, TimeUnit.SECONDS);
assertEquals(1L, garbageResult.getRecordMetadata().offset());
sendAndVerifyMessage(records, objectTemplate, "baz", new JsonObject("baz"), 2L);
}
private <T> KafkaTemplate<String, T> createPublisher(final String label, final Serializer<T> serializer) {
final Map<String, Object> producerProps = new HashMap<>();
producerProps.put(ProducerConfig.BOOTSTRAP_SERVERS_CONFIG, embeddedKafka.getBrokersAsString());
producerProps.put(ProducerConfig.CLIENT_ID_CONFIG, "TestPublisher-" + label);
producerProps.put(ProducerConfig.ACKS_CONFIG, "all");
producerProps.put(ProducerConfig.RETRIES_CONFIG, 2);
producerProps.put(ProducerConfig.MAX_IN_FLIGHT_REQUESTS_PER_CONNECTION, 1);
producerProps.put(ProducerConfig.REQUEST_TIMEOUT_MS_CONFIG, 5000);
producerProps.put(ProducerConfig.MAX_BLOCK_MS_CONFIG, 5000);
producerProps.put(ProducerConfig.KEY_SERIALIZER_CLASS_CONFIG, StringSerializer.class);
producerProps.put(ProducerConfig.VALUE_SERIALIZER_CLASS_CONFIG, serializer.getClass());
final DefaultKafkaProducerFactory<String, T> pf = new DefaultKafkaProducerFactory<>(producerProps);
pf.setValueSerializer(serializer);
return new KafkaTemplate<>(pf);
}
private BlockingQueue<ConsumerRecord<String, JsonObject>> createListener(final String topic) throws Exception {
final Map<String, Object> consumerProps = new HashMap<>();
consumerProps.put(ProducerConfig.BOOTSTRAP_SERVERS_CONFIG, embeddedKafka.getBrokersAsString());
consumerProps.put(ConsumerConfig.GROUP_ID_CONFIG, "TestConsumer");
consumerProps.put(ConsumerConfig.ENABLE_AUTO_COMMIT_CONFIG, true);
consumerProps.put(ConsumerConfig.AUTO_COMMIT_INTERVAL_MS_CONFIG, "100");
consumerProps.put(ConsumerConfig.SESSION_TIMEOUT_MS_CONFIG, 15000);
consumerProps.put(ConsumerConfig.KEY_DESERIALIZER_CLASS_CONFIG, StringDeserializer.class);
consumerProps.put(ConsumerConfig.VALUE_DESERIALIZER_CLASS_CONFIG, JsonDeserializer.class);
final DefaultKafkaConsumerFactory<String, JsonObject> cf = new DefaultKafkaConsumerFactory<>(consumerProps);
cf.setValueDeserializer(new JsonDeserializer<>(JsonObject.class));
final KafkaMessageListenerContainer<String, JsonObject> container = new KafkaMessageListenerContainer<>(cf, new ContainerProperties(topic));
final BlockingQueue<ConsumerRecord<String, JsonObject>> records = new LinkedBlockingQueue<>();
container.setupMessageListener((MessageListener<String, JsonObject>) records::add);
container.setBeanName("TestListener");
container.start();
ContainerTestUtils.waitForAssignment(container, embeddedKafka.getPartitionsPerTopic());
return records;
}
private void sendAndVerifyMessage(final BlockingQueue<ConsumerRecord<String, JsonObject>> records,
final KafkaTemplate<String, JsonObject> template,
final String key, final JsonObject value,
final long expectedOffset) throws InterruptedException, ExecutionException, TimeoutException {
final ListenableFuture<SendResult<String, JsonObject>> future = template.send(TOPIC1, key, value);
final ConsumerRecord<String, JsonObject> record = records.poll(5, TimeUnit.SECONDS);
assertThat(record, hasKey(key));
assertThat(record, hasValue(value));
assertEquals(expectedOffset, future.get(5, TimeUnit.SECONDS).getRecordMetadata().offset());
}
public static final class JsonObject {
private String value;
public JsonObject() {}
JsonObject(final String value) {
this.value = value;
}
public String getValue() {
return value;
}
public void setValue(final String value) {
this.value = value;
}
@Override
public boolean equals(final Object o) {
if (this == o) { return true; }
if (o == null || getClass() != o.getClass()) { return false; }
final JsonObject that = (JsonObject) o;
return Objects.equals(value, that.value);
}
@Override
public int hashCode() {
return Objects.hash(value);
}
@Override
public String toString() {
return "JsonObject{" +
"value='" + value + '\'' +
'}';
}
}
}
答案 0 :(得分：2)
从spring-kafka-2.x.x开始，我们现在可以轻松地在接口KafkaListenerErrorHandler的配置文件中声明bean，实现方式为
@Bean
public ConsumerAwareListenerErrorHandler listen3ErrorHandler() {
return (m, e, c) -> {
this.listen3Exception = e;
MessageHeaders headers = m.getHeaders();
c.seek(new org.apache.kafka.common.TopicPartition(
headers.get(KafkaHeaders.RECEIVED_TOPIC, String.class),
headers.get(KafkaHeaders.RECEIVED_PARTITION_ID, Integer.class)),
headers.get(KafkaHeaders.OFFSET, Long.class));
return null;
};
}
更多资源可在https://docs.spring.io/spring-kafka/reference/htmlsingle/#annotation-error-handling上找到，还有另一个链接也有类似的问题：Spring Kafka error handling - v1.1.x和How to handle SerializationException after deserialization
答案 1 :(得分：1)
我有一个解决方案，但我不知道它是否是最好的解决方案，我按如下方式扩展JsonDeserializer，导致spring-kafka消耗null值并需要必要的下游更改处理这种情况。
class SafeJsonDeserializer[A >: Null](targetType: Class[A], objectMapper: ObjectMapper) extends JsonDeserializer[A](targetType, objectMapper) with Logging {
override def deserialize(topic: String, data: Array[Byte]): A = try {
super.deserialize(topic, data)
} catch {
case e: Exception =>
logger.error("Failed to deserialize data [%s] from topic [%s]".format(new String(data), topic), e)
null
}
}