SparkStreaming读写kerberos Hive（Yarn Cluster）

àì夳堔傛蜴生んèń 2022-01-31 02:51 161阅读 0赞

yarn Cluster模式也可适用于从集群外部提交（接口机）

##### 提交命令 #####

spark-submit的最后一行为args参数，经多次尝试发现yarn cluster模式通过该方式传入kerberos文件可行

1.  krb5.conf
    
    需要在yarn集群所有节点相同位置放置
2.  client.keytab
    
    需要在命令提交节点，yarn集群所有节点相同位置放置，container连接HDFS\[kerberos\]集群使用
3.  kafka\_client\_jaas.conf
    
    需要在yarn集群所有节点相同位置放置，container连接kafka\[kerberos\]集群使用

spark-submit --master yarn \
    --deploy-mode cluster \
    --class com.liubin.spark.kerberos.HiveSinkDemoKerberosYarnCluster \
    --keytab /tmp/client.keytab \
    --principal test@LIUBIN.COM \
    --conf "spark.executor.extraJavaOptions=-Djava.security.auth.login.config=/tmp/kafka_client_jaas.conf" \
    spark-example-1.0.0.jar \
    /tmp/krb5.conf abcd@LIUBIN.COM /tmp/client.keytab /tmp/kafka_client_jaas.conf

##### resources文件 #####

需要从kerberos集群中将相关文件拷贝到项目的resource目录下（需打入jar包中）

core-site.xml
    hdfs-site.xml
    yarn-site.xml
    hive-site.xml

##### util类 #####

package com.liubin.spark.kerberos
    
    import java.util.concurrent.Future
    import org.apache.kafka.clients.producer.{KafkaProducer, ProducerRecord, RecordMetadata}
    
    import scala.collection.JavaConversions._
    
    class KafkaSink[K, V](createProducer: () => KafkaProducer[K, V]) extends Serializable {
    
      lazy val producer = createProducer()
    
      def send(topic: String, key: K, value: V): Future[RecordMetadata] =
        producer.send(new ProducerRecord[K, V](topic, key, value))
    
      def send(topic: String, value: V): Future[RecordMetadata] =
        producer.send(new ProducerRecord[K, V](topic, value))
    }
    
    object KafkaSink {
    
      def apply[K, V](config: Map[String, Object]): KafkaSink[K, V] = {
        val createProducerFunc = () => {
          val producer = new KafkaProducer[K, V](config)
          sys.addShutdownHook {
            producer.close()
          }
          producer
        }
        new KafkaSink(createProducerFunc)
      }
    
      def apply[K, V](config: java.util.Properties): KafkaSink[K, V] = apply(config.toMap)
    }

##### 主代码 #####

package com.liubin.spark.kerberos
    
    import java.util.Properties
    
    import org.apache.kafka.common.serialization.{StringDeserializer, StringSerializer}
    import org.apache.spark.broadcast.Broadcast
    import org.apache.spark.sql.SparkSession
    import org.apache.spark.streaming.{Seconds, StreamingContext}
    import org.apache.spark.{SparkConf, SparkContext}
    
    /**
      * author : liubin
      * date : 2019/5/8
      * Description : 利用kafkaDS工具类实现将数据存储到Hive[kerberos]，数据源为kafka[kerberos]
      */
    object HiveSinkDemoKerberosYarnCluster {
    
      // kafka conf
      val sinkTopic = "kafkaSink"
      val sourceTopic = "kafkaSource"
      val bootstrapServers = "node1:9092,node2:9092,node3:9092"
      val autoOffsetReset = "latest"
      val groupId = "test-kerberos"
    
      // kerberos conf
      val krb5Debug = "true"
      val krb5Path = args(0)
      val principal = args(1)
      val keytab = args(2)
      val kafkaKerberos = args(3)
    
      def main(args: Array[String]): Unit = {
    
        // set global kerberos conf
        System.setProperty("java.security.krb5.conf", krb5Path)
        System.setProperty("sun.security.krb5.debug", krb5Debug)
        System.setProperty("java.security.auth.login.config", kafkaKerberos)
    
        val conf = new SparkConf().setAppName(this.getClass.getSimpleName)
        val session = SparkSession.builder().config(conf)      .enableHiveSupport().getOrCreate()
        // 配置hive支持动态分区
        session.sqlContext.setConf("hive.exec.dynamic.partition", "true")
        // 配置hive动态分区为非严格模式
        session.sqlContext.setConf("hive.exec.dynamic.partition.mode", "nonstrict")
        // 配置是否合并小文件
        session.sqlContext.setConf("hive.merge.sparkfiles", "true")
          
        val ssc = new StreamingContext(session.sparkContext, Seconds(5))
    
        // kafka source config
        val kafkaParams = Map[String, Object](
          "key.deserializer" -> classOf[StringDeserializer],
          "value.deserializer" -> classOf[StringDeserializer],
          "group.id" -> groupId,
          "bootstrap.servers" -> bootstrapServers,
          "enable.auto.commit" -> (true: java.lang.Boolean),
          "auto.offset.reset" -> autoOffsetReset,
          //在kerberos环境下，以下配置需要增加
          "security.protocol" -> "SASL_PLAINTEXT",
          "sasl.kerberos.service.name" -> "kafka",
          "sasl.mechanism" -> "GSSAPI"
        )
    
        // kafka sink config
        val kafkaProducer: Broadcast[KafkaSink[String, String]] = {
          val kafkaProducerConfig = {
            val p = new Properties()
            p.setProperty("bootstrap.servers", bootstrapServers)
            p.setProperty("key.serializer", classOf[StringSerializer].getName)
            p.setProperty("value.serializer", classOf[StringSerializer].getName)
            //在kerberos环境下，以下配置需要增加
            p.setProperty("security.protocol", "SASL_PLAINTEXT")
            p.setProperty("sasl.mechanism", "GSSAPI")
            p.setProperty("sasl.kerberos.service.name", "kafka")
            p
          }
          ssc.sparkContext.broadcast(KafkaSink[String, String](kafkaProducerConfig))
        }
          
        val hadoopConf = new Configuration()
        hbaseConf.addResource("hive-site.xml")
    
        // Hbase操作:根据userPrincipal和keytabFile获取连接
        UserGroupInformation.setConfiguration(hadoopConf)
        UserGroupInformation.loginUserFromKeytab(principal, keytab)
    
        // kafka数据源
        val kafkaDStream = KafkaSource.createDirectStream[String, String](ssc, sourceTopic, kafkaParams)
    
        try {
          dstream.foreachRDD(rdd => {
            
          })
    
        } catch {
          case e: RuntimeException => e.printStackTrace()
        }
    
        ssc.start()
        ssc.awaitTermination()
    
      }
    }