Cannot connect to Cassandra from Spark (Contact points contain multiple data centers)

问题

I am trying to run my first spark job (a Scala job that accesses Cassandra) which is failing and showing the following error :

java.io.IOException: Failed to open native connection to Cassandra at {<ip>}:9042
at com.datastax.spark.connector.cql.CassandraConnector$.com$datastax$spark$connector$cql$CassandraConnector$$createSession(CassandraConnector.scala:164)
at com.datastax.spark.connector.cql.CassandraConnector$$anonfun$2.apply(CassandraConnector.scala:150)
at com.datastax.spark.connector.cql.CassandraConnector$$anonfun$2.apply(CassandraConnector.scala:150)
at com.datastax.spark.connector.cql.RefCountedCache.createNewValueAndKeys(RefCountedCache.scala:31)
...........
............
Caused by: java.lang.IllegalArgumentException: Contact points contain multiple data centers: 
at com.datastax.spark.connector.cql.LocalNodeFirstLoadBalancingPolicy.init(LocalNodeFirstLoadBalancingPolicy.scala:47)
at com.datastax.driver.core.Cluster$Manager.init(Cluster.java:1099)
at com.datastax.driver.core.Cluster.getMetadata(Cluster.java:271)
at com.datastax.spark.connector.cql.CassandraConnector$.com$datastax$spark$connector$cql$CassandraConnector$$createSession(CassandraConnector.scala:157)

What are we doing wrong here?

I am using :

• Spark 1.5.2
• Apache Cassandra 2.1.10
• spark-cassandra connector 1.3.1 /1.5.0-M2 (tried both connectors)
• Scala version 2.10.4

回答1:

--> According to author there a work in progress to fix this. See comments below this answer.

I found this in the documentation, I hope it will help you :

override def init(cluster: Cluster, hosts: JCollection[Host]) {
    nodes = hosts.toSet
    // use explicitly set DC if available, otherwise see if all contact points have same DC
    // if so, use that DC; if not, throw an error
    dcToUse = localDC match { 
      case Some(local) => local
      case None => 
        val dcList = dcs(nodesInTheSameDC(contactPoints, hosts.toSet))
        if (dcList.size == 1) 
            dcList.head
        else 
            throw new IllegalArgumentException(s"Contact points contain multiple data centers: ${dcList.mkString(", ")}")
    }
    clusterMetadata = cluster.getMetadata
}

回答2:

I was facing the same issue while trying to connect two Cassandra data center using Apache Spark 2.x.x.

public class  SparkCassandraTest {  
  private static final String CASSANDRA_ENDPOINTS = "DC1_node1,DC1_node2,DC1_node3,DC2_node1,DC2_node2,DC2_node3";

  public static void main(String[] args) {
        sparkConf = new SparkConf().setAppName(APP_NAME);
        sparkConf.set("spark.cassandra.connection.host", CASSANDRA_ENDPOINTS);
        sparkConf.set("spark.cassandra.auth.username", CASSANDRA_USERNAME);
        sparkConf.set("spark.cassandra.auth.password", CASSANDRA_PASSWORD);
        sparkSession = SparkSession.builder().config(sparkConf).enableHiveSupport().getOrCreate();

        //.....................
        //.....................
        //.....................
       }
}

Caused by: java.lang.IllegalArgumentException: requirement failed: Contact points contain multiple data centers: DC2-XXXXX2, DC1-XXXXX1

I resolve this issue by connecting any one Cassandra data center (DC1_node1,DC1_node2,DC1_node3) or (DC2_node1,DC2_node2,DC2_node3).

来源：https://stackoverflow.com/questions/34004959/cannot-connect-to-cassandra-from-spark-contact-points-contain-multiple-data-cen