Apache Cassandra Failover and load balancing

Apache Cassandra Failover and Load Balancing

I have 3-node cluster, my RF=3, W=ONE and R=QUORUM, once a node is down, another two node can work properly, if the downed node comes back, it also get update.

To solve load balancing, only need to change one line of the codes(from last tutorial) to :
//load balancing build a connection pool
cassandraHostConfigurator = new CassandraHostConfigurator("192.168.0.1:9160,192.168.0.2:9160,192.168.0.3:9160");
Hector will do load balancing internally.

To solve failover problem: please check here.
Be aware of the relationship between Replication factor and Consistency Level, please check here.

W+R > N.  N is replication factor

I'm using cassandra(1.1.5), so I modified
http://ac31004.blogspot.ca/2010/08/consistencylevel-in-hector-and.html's code to:



import me.prettyprint.cassandra.service.OperationType;
import me.prettyprint.hector.api.ConsistencyLevelPolicy;
import me.prettyprint.hector.api.HConsistencyLevel;

// W+R > N .  where N is Replication Factor
public final class MyConsistencyLevel implements ConsistencyLevelPolicy
{
    @Override
    public  HConsistencyLevel get(OperationType op)
    {
       switch (op)
       {
          case READ:
              return HConsistencyLevel.QUORUM;
          case WRITE:
              return HConsistencyLevel.ONE;
          default:
              return HConsistencyLevel.QUORUM;
       }
    }

    @Override
    public HConsistencyLevel get(OperationType op, String cfName)
    {
       return HConsistencyLevel.QUORUM;
    }
}

Add or modify those code in CassandraTest.java(you can find it in last tutorial.)

       ConsistencyLevelPolicy consistencyLevelPolicy = new MyConsistencyLevel();

        Keyspace kpo = HFactory.createKeyspace(keySpace, cluster);
        // set CL
        kpo.setConsistencyLevelPolicy(consistencyLevelPolicy);