Spark Streaming Kafka：使用spark-submit运行时，ByteArrayDeserializer的ClassNotFoundException

Question

I'm new to Scala / Spark Streaming, and to StackOverflow so please excuse my formatting. I have made a Scala app that reads log files from a Kafka Stream. It runs fine within the IDE, but I'll be damned if I can get it to run using spark-submit . It always fails with:

ClassNotFoundException: org.apache.kafka.common.serialization.ByteArrayDeserializer

The line referenced in the Exception is the load command in this snippet:

val records = spark
  .readStream
  .format("kafka") // <-- use KafkaSource
  .option("subscribe", kafkaTopic)
  .option("kafka.bootstrap.servers", kafkaBroker) // 192.168.4.86:9092
  .load()
  .selectExpr("CAST(value AS STRING) AS temp")
  .withColumn("record", deSerUDF($"temp"))

• IDE: IntelliJ
• Spark: 2.2.1
• Scala: 2.11.8
• Kafka: kafka_2.11-0.10.0.0

Relevant parts of pom.xml :

<properties>
    <maven.compiler.source>1.8</maven.compiler.source>
    <maven.compiler.target>1.8</maven.compiler.target>
    <encoding>UTF-8</encoding>
    <scala.version>2.11.8</scala.version>
    <scala.compat.version>2.11</scala.compat.version>
    <spark.version>2.2.1</spark.version>
</properties>

<dependencies>
    <dependency>
        <groupId>org.scala-lang</groupId>
        <artifactId>scala-library</artifactId>
        <version>${scala.version}</version>
    </dependency>
    <dependency>
        <groupId>org.apache.spark</groupId>
        <artifactId>spark-core_2.11</artifactId>
        <version>${spark.version}</version>
    </dependency>
    <dependency>
        <groupId>org.apache.spark</groupId>
        <artifactId>spark-streaming_2.11</artifactId>
        <version>${spark.version}</version>
        <scope>provided</scope>
    </dependency>
    <dependency>
        <groupId>org.apache.spark</groupId>
        <artifactId>spark-sql_2.11</artifactId>
        <version>${spark.version}</version>
    </dependency>
    <dependency>
        <groupId>com.github.scala-incubator.io</groupId>
        <artifactId>scala-io-file_2.11</artifactId>
        <version>0.4.3-1</version>
    </dependency>
    <dependency>
        <groupId>org.apache.spark</groupId>
        <artifactId>spark-streaming-kafka-0-10_2.11</artifactId>
        <version>${spark.version}</version>
    </dependency>
    <dependency>
        <groupId>org.apache.spark</groupId>
        <artifactId>spark-sql-kafka-0-10_2.11</artifactId>
        <version>${spark.version}</version>
    </dependency>
    <dependency>
        <groupId>org.apache.kafka</groupId>
        <artifactId>kafka-clients</artifactId>
        <version>0.10.0.0</version>
        <!-- version>2.0.0</version -->
    </dependency>

Note: I don't think it is related, but I have to use zip -d BroLogSpark.jar "META-INF/*.SF" and zip -d BroLogSpark.jar "META-INF/*.DSA" to get past meaning about the manifest signatures.

My jar file does not include any of org.apache.kafka . I have seen several posts that strongly suggest I have a mismatch in versions, and I have tried countless permutations of changes to pom.xml and spark-submit . After each change, I confirm that it still runs within the IDE, then proceed to try using spark-submit on the same system, same user. Below is my most recent attempt, where my BroLogSpark.jar is in the current directory and "192.168.4.86:9092 profile" are input arguments.

spark-submit --packages org.apache.spark:spark-streaming-kafka-0-10_2.11:2.2.1,org.apache.kafka:kafka-clients:0.10.0.0 BroLogSpark.jar 192.168.4.86:9092 BroFile

Answer 1

Add below dependency too

<dependency>
    <groupId>org.apache.kafka</groupId>
    <artifactId>kafka_2.11</artifactId>
    <version>0.10.0.0</version>
</dependency>