从Kafka阅读到Flink Scala Shell

Question

我试图连接到我的本地机器上的Kafka（2.1），并在Flink（1.7.2）附带的scalashell中读取。

下面是我正在做的：

:require flink-connector-kafka_2.11-1.7.1.jar
:require flink-connector-kafka-base_2.11-1.7.1.jar

import org.apache.flink.streaming.connectors.kafka.FlinkKafkaConsumer
import org.apache.flink.streaming.util.serialization.SimpleStringSchema
import org.apache.flink.streaming.connectors.kafka.FlinkKafkaConsumerBase
import java.util.Properties

val properties = new Properties()
properties.setProperty("bootstrap.servers", "localhost:9092")
properties.setProperty("group.id", "test")
var stream = senv.addSource(new FlinkKafkaConsumer[String]("topic", new SimpleStringSchema(), properties)).print()

之后，最后一条语句我得到了以下错误：

scala> var stream = senv.addSource(new FlinkKafkaConsumer[String]("topic", new SimpleStringSchema(), properties)).print()
<console>:69: error: overloaded method value addSource with alternatives:
  [T](function: org.apache.flink.streaming.api.functions.source.SourceFunction.SourceContext[T] => Unit)(implicit evidence$10: org.apache.flink.api.common.typeinfo.TypeInformation[T])org.apache.flink.streaming.api.scala.DataStream[T] <and>
  [T](function: org.apache.flink.streaming.api.functions.source.SourceFunction[T])(implicit evidence$9: org.apache.flink.api.common.typeinfo.TypeInformation[T])org.apache.flink.streaming.api.scala.DataStream[T]
 cannot be applied to (org.apache.flink.streaming.connectors.kafka.FlinkKafkaConsumer[String])
   var stream = senv.addSource(new FlinkKafkaConsumer[String]("topic", new SimpleStringSchema(), properties)).print()

我已经创建了一个名为“topic”的主题，我能够通过另一个客户端正确地生成和读取来自它的消息。我正在使用java版本1.8.0\u 201，并遵循https://ci.apache.org/projects/flink/flink-docs-stable/dev/connectors/kafka.html。

有什么可能出错的吗？

Answer 1

很可能您应该在添加源之前导入Flink的Scala隐式：

import org.apache.flink.streaming.api.scala._

Answer 2

某些依赖项需要其他依赖项，这是隐含的。我们通常使用一些依赖关系管理器，如maven或sbt，当我们向项目中添加一些依赖关系时，依赖关系管理器将在后台提供其隐含的依赖关系。

另一方面，当您使用shell时，在没有依赖项管理器的地方，您有责任提供您的代码依赖项。使用Flink Kafka连接器显式需要Flink Connector Kafkajar，但您应该注意到Flink Connector Kafka也需要一些依赖项。您可以在页面底部找到它的依赖项，它位于编译依赖项部分。所以从这个序言开始，我将以下jar文件添加到目录FLINK_HOME/lib（Flink类路径）：

flink-connector-kafka-0.11_2.11-1.4.2.jar
flink-connector-kafka-0.10_2.11-1.4.2.jar    
flink-connector-kafka-0.9_2.11-1.4.2.jar   
flink-connector-kafka-base_2.11-1.4.2.jar  
flink-core-1.4.2.jar                                         
kafka_2.11-2.1.1.jar
kafka-clients-2.1.0.jar

我可以使用Flink shell中的以下代码成功使用Kafka消息：

scala> import org.apache.flink.streaming.connectors.kafka.FlinkKafkaConsumer011
import org.apache.flink.streaming.connectors.kafka.FlinkKafkaConsumer011

scala> import org.apache.flink.streaming.util.serialization.SimpleStringSchema
import org.apache.flink.streaming.util.serialization.SimpleStringSchema

scala> import java.util.Properties
import java.util.Properties

scala> val properties = new Properties()
properties: java.util.Properties = {}

scala> properties.setProperty("bootstrap.servers", "localhost:9092")
res0: Object = null

scala> properties.setProperty("group.id", "test")
res1: Object = null

scala> val stream = senv.addSource(new FlinkKafkaConsumer011[String]("topic", new SimpleStringSchema(), properties)).print()
warning: there was one deprecation warning; re-run with -deprecation for details
stream: org.apache.flink.streaming.api.datastream.DataStreamSink[String] = org.apache.flink.streaming.api.datastream.DataStreamSink@71de1091

scala> senv.execute("Kafka Consumer Test")
Submitting job with JobID: 23e3bb3466d914a2747ae5fed293a076. Waiting for job completion.
Connected to JobManager at Actor[akka.tcp://flink@localhost:40093/user/jobmanager#1760995711] with leader session id 00000000-0000-0000-0000-000000000000.
03/11/2019 21:42:39 Job execution switched to status RUNNING.
03/11/2019 21:42:39 Source: Custom Source -> Sink: Unnamed(1/1) switched to SCHEDULED 
03/11/2019 21:42:39 Source: Custom Source -> Sink: Unnamed(1/1) switched to SCHEDULED 
03/11/2019 21:42:39 Source: Custom Source -> Sink: Unnamed(1/1) switched to DEPLOYING 
03/11/2019 21:42:39 Source: Custom Source -> Sink: Unnamed(1/1) switched to DEPLOYING 
03/11/2019 21:42:39 Source: Custom Source -> Sink: Unnamed(1/1) switched to RUNNING 
03/11/2019 21:42:39 Source: Custom Source -> Sink: Unnamed(1/1) switched to RUNNING 
hello
hello

此外，将一些jar文件添加到Flink类路径的另一种方法是将jar作为Flink shell start命令的参数传递：

bin/start-scala-shell.sh local "--addclasspath <path/to/jar.jar>"

测试环境：

Flink 1.4.2
Kafka 2.1.0
Java  1.8 201
Scala 2.11