KafkaSpout tuple replay 抛出空指针异常

Question

我正在使用storm 1.0.1和Kafka 0.10.0.0以及storm-kafka-client 1.0.3。

请在下面找到我的代码配置。

kafkaConsumerProps.put(KafkaSpoutConfig.Consumer.KEY_DESERIALIZER, "org.apache.kafka.common.serialization.ByteArrayDeserializer");
            kafkaConsumerProps.put(KafkaSpoutConfig.Consumer.VALUE_DESERIALIZER, "org.apache.kafka.common.serialization.ByteArrayDeserializer");


            KafkaSpoutStreams kafkaSpoutStreams = new KafkaSpoutStreamsNamedTopics.Builder(new Fields(fieldNames), topics)
                    .build();

            KafkaSpoutRetryService retryService = new KafkaSpoutRetryExponentialBackoff(TimeInterval.microSeconds(500),
                                TimeInterval.milliSeconds(2), Integer.MAX_VALUE, TimeInterval.seconds(10));


            KafkaSpoutTuplesBuilder tuplesBuilder = new KafkaSpoutTuplesBuilderNamedTopics.Builder(new TestTupleBuilder(topics))
                        .build();

            KafkaSpoutConfig kafkaSpoutConfig = new KafkaSpoutConfig.Builder<String, String>(kafkaConsumerProps, kafkaSpoutStreams, tuplesBuilder, retryService)
                                                                .setOffsetCommitPeriodMs(10_000)
                                                                .setFirstPollOffsetStrategy(LATEST)
                                                                .setMaxRetries(5)
                                                                .setMaxUncommittedOffsets(250)
                                                                .build();

当我使元组失败时，它不会被重播。 Spout 抛出错误。 请告诉我为什么它会抛出空指针异常。

53501 [Thread-359-test-spout-executor[295 295]] ERROR o.a.s.util - Async loop died!
java.lang.NullPointerException
    at org.apache.storm.kafka.spout.KafkaSpout.doSeekRetriableTopicPartitions(KafkaSpout.java:260) ~[storm-kafka-client-1.0.3.jar:1.0.3]
    at org.apache.storm.kafka.spout.KafkaSpout.pollKafkaBroker(KafkaSpout.java:248) ~[storm-kafka-client-1.0.3.jar:1.0.3]
    at org.apache.storm.kafka.spout.KafkaSpout.nextTuple(KafkaSpout.java:203) ~[storm-kafka-client-1.0.3.jar:1.0.3]
    at org.apache.storm.daemon.executor$fn__7885$fn__7900$fn__7931.invoke(executor.clj:645) ~[storm-core-1.0.1.jar:1.0.1]
    at org.apache.storm.util$async_loop$fn__625.invoke(util.clj:484) [storm-core-1.0.1.jar:1.0.1]
    at clojure.lang.AFn.run(AFn.java:22) [clojure-1.8.0.jar:?]
    at java.lang.Thread.run(Thread.java:745) [?:1.8.0_102]
53501 [Thread-359-test-spout-executor[295 295]] ERROR o.a.s.d.executor - 
java.lang.NullPointerException
    at org.apache.storm.kafka.spout.KafkaSpout.doSeekRetriableTopicPartitions(KafkaSpout.java:260) ~[storm-kafka-client-1.0.3.jar:1.0.3]
    at org.apache.storm.kafka.spout.KafkaSpout.pollKafkaBroker(KafkaSpout.java:248) ~[storm-kafka-client-1.0.3.jar:1.0.3]
    at org.apache.storm.kafka.spout.KafkaSpout.nextTuple(KafkaSpout.java:203) ~[storm-kafka-client-1.0.3.jar:1.0.3]
    at org.apache.storm.daemon.executor$fn__7885$fn__7900$fn__7931.invoke(executor.clj:645) ~[storm-core-1.0.1.jar:1.0.1]
    at org.apache.storm.util$async_loop$fn__625.invoke(util.clj:484) [storm-core-1.0.1.jar:1.0.1]
    at clojure.lang.AFn.run(AFn.java:22) [clojure-1.8.0.jar:?]
    at java.lang.Thread.run(Thread.java:745) [?:1.8.0_102]
53527 [Thread-359-test-spout-executor[295 295]] ERROR o.a.s.util - Halting process: ("Worker died")
java.lang.RuntimeException: ("Worker died")
    at org.apache.storm.util$exit_process_BANG_.doInvoke(util.clj:341) [storm-core-1.0.1.jar:1.0.1]
    at clojure.lang.RestFn.invoke(RestFn.java:423) [clojure-1.8.0.jar:?]
    at org.apache.storm.daemon.worker$fn__8554$fn__8555.invoke(worker.clj:761) [storm-core-1.0.1.jar:1.0.1]
    at org.apache.storm.daemon.executor$mk_executor_data$fn__7773$fn__7774.invoke(executor.clj:271) [storm-core-1.0.1.jar:1.0.1]
    at org.apache.storm.util$async_loop$fn__625.invoke(util.clj:494) [storm-core-1.0.1.jar:1.0.1]
    at clojure.lang.AFn.run(AFn.java:22) [clojure-1.8.0.jar:?]
    at java.lang.Thread.run(Thread.java:745) [?:1.8.0_102]

请在下面找到完整的 spout 配置 {key.deserializer=org.apache.kafka.common.serialization.ByteArrayDeserializer, value.deserializer=org.apache.kafka.common.serialization.ByteArrayDeserializer, group.id=test-group, ssl.keystore.location=C:/ test.jks, bootstrap.servers=localhost:1000, auto.commit.interval.ms=1000, security.protocol=SSL, enable.auto.commit=true, ssl.truststore.location=C:/test1.jks, ssl .keystore.password=pass123, ssl.key.password=pass123, ssl.truststore.password=pass123, session.timeout.ms=30000, auto.offset.reset=latest}

Answer 1

Storm 1.0.1 包含 beta 质量的storm-kafka-client。我们修复了一些问题，Storm 1.1 版本中提供了更稳定的版本，可用于 Kafka 0.10 及更高版本。 在您的拓扑中，您​​可以对 Storm-kafka-client 版本 1.1 和具有适当版本的 kafka-clients 依赖项进行依赖。你不需要升级storm集群本身。

Answer 2

我让 enable.auto.commit=true 将值设为 false 为我解决了这个问题。