Cassandra集群配置

版本：apache-cassandra-1.0.6

机器1:10.16.233.27
机器2:10.16.233.41
机器3:10.16.233.45
机器4:10.16.233.36
机器5:10.16.233.39
机器6:10.16.233.66

机器1和机器2作为seed_provider
java jdk已经安装并且设置过JAVA_HOME
ssh已经设置
1.解压 apache-cassandra-1.0.6-bin.tar.gz 到/usr/cassandra 目录
设置CASSANDRA_HOME,打开profile文件
命令：vim /etc/profile
设置如下：
    JAVA_HOME=/usr/java/jdk1.6.0_30
    CASSANDRA_HOME=/usr/cassandra
    CLASSPATH=.:$JAVA_HOME/lib:$CASSANDRA_HOME/lib
    PATH=$PATH:$JAVA_HOME/bin:$JAVA_HOME/jre/bin:$CASSANDRA_HOME/bin

    export JAVA_HOME
    export CASSANDRA_HOME
    export PATH
    export CLASSPATH

2.按需要修改6台机器的conf/cassandra.yaml文件

修改conf/cassandra.yaml
2.1设置cluster_name 这里我们命名为E3E4
    # The name of the cluster. This is mainly used to prevent machines in
    # one logical cluster from joining another.
    cluster_name: 'E3E4'

2.2设置数据文件 提交日志 缓存文件 路径
    # directories where Cassandra should store data on disk.
    data_file_directories:
        - /usr/cassandra/data
    # commit log
    commitlog_directory: /usr/cassandra/commitlog
    # saved caches
    saved_caches_directory: /usr/cassandra/saved_caches

2.3设置endpoint_snitch
    # endpoint_snitch -- Set this to a class that implements
    # IEndpointSnitch, which will let Cassandra know enough
    # org.apache.cassandra.locator.PropertyFileSnitch:
    #  - Proximity is determined by rack and data center, which are
    #    explicitly configured in cassandra-topology.properties.
    endpoint_snitch: org.apache.cassandra.locator.PropertyFileSnitch

2.4设置init_token，注：这里要通过算法程序来计算
算法代码如下：
    #! /usr/bin/python
    import sys
    if(len(sys.argv)>1):
        num=int(sys.argv[1])
    else:
        num=int(raw_input(""))
    for i in range(0,num):
        print 'token %d:%d'%(i,(i*(2**127)/num))
计算结果（6台机器）：
    token 0:0
    token 1:28356863910078205288614550619314017621
    token 2:56713727820156410577229101238628035242
    token 3:85070591730234615865843651857942052864
    token 4:113427455640312821154458202477256070485
    token 5:141784319550391026443072753096570088106
算好后，依据输出设置initial_token,PS initial_token冒号后有个空格。
    # If blank, Cassandra will request a token bisecting the range of
    # the heaviest-loaded existing node.  If there is no load information
    # available, such as is the case with a new cluster, it will pick
    # a random token, which will lead to hot spots.
    initial_token: 85070591730234615865843651857942052864

2.5设置seed_provider
    seed_provider:
        # Addresses of hosts that are deemed contact points.
        # Cassandra nodes use this list of hosts to find each other and learn
        # the topology of the ring.  You must change this if you are running
        # multiple nodes!
        - class_name: org.apache.cassandra.locator.SimpleSeedProvider
          parameters:
          # seeds is actually a comma-delimited list of addresses.
          # Ex: "<ip1>,<ip2>,<ip3>"
          - seeds: "10.16.233.41,10.16.233.27"


2.6 设置listen_address rpc_address为本机IP，这里机器3，机器4的设置如下
      机器3:
      listen_address: 10.16.233.36
      rpc_address: 10.16.233.36
      机器4:
      listen_address: 10.16.233.39
      rpc_address: 10.16.233.39

2.7修改6台机器:conf/cassandra-topology.properties
    # Cassandra Node IP=Data Center:Rack
    10.16.233.41=DC1:RAC1
    10.16.233.27=DC2:RAC1
    10.16.233.45=DC1:RAC1
    10.16.233.39=DC2:RAC1
    10.16.233.36=DC1:RAC2
    10.16.233.66=DC2:RAC2
    # default for unknown nodes
    default=DC1:RAC1

2.8 修改conf/log4j-server.properties 设置日志目录：/usr/cassandra/log/system.log
    # rolling log file
    log4j.appender.R=org.apache.log4j.RollingFileAppender
    log4j.appender.R.maxFileSize=20MB
    log4j.appender.R.maxBackupIndex=50
    log4j.appender.R.layout=org.apache.log4j.PatternLayout
    log4j.appender.R.layout.ConversionPattern=%5p [%t] %d{ISO8601} %F (line %L) %m%n
    # Edit the next line to point to your logs directory
    log4j.appender.R.File=/usr/cassandra/log/system.log

3设置hosts
[root@tt bin]# vim /etc/hosts
内容如下:
127.0.0.1       localhost.localdomain   localhost
::1             localhost6.localdomain6 localhost6
10.16.233.41 tt.cassandra.xa.01
10.16.233.27 tt.cassandra.xa.02
10.16.233.45 tt.cassandra.xa.03
10.16.233.39 tt.cassandra.xa.04
10.16.233.36 tt.cassandra.xa.05
10.16.233.66 tt.cassandra.xa.06

4设置network
[root@tt bin]# vim /etc/sysconfig/network
内容如下:
NETWORKING=yes
HOSTNAME=tt.cassandra.xa.04

5 清空数据并且重启机器
rm -rf /usr/cassandra/data/*
rm -rf /usr/cassandra/saved_caches/*
rm -rf /usr/cassandra/commitlog/*
rm -rf /usr/cassandra/log/*
6一定要记住关闭防火墙
[root@tt bin]# iptables -F:关闭防火

启动Cassandra 并查看：
#在cassandra的bin目录，10.16.233.39是本机IP
[root@tt bin]# ./nodetool -h 10.16.233.39 ring