最近因为公司准备将数据迁移到Hbase上，所以不得不学习hadoop，我们尝试过将数据迁移到mongodb上面，周末在家里做了一个小试验，使用java+mongodb做爬虫抓取数据，我将mongodb安装在centos6.3虚拟机上，分配1G的内存，开始抓数据，半小时后，虚拟机内存吃光，没有办法解决内存问题，网上很多人说没有32G的内存不要玩mongodb，这样的说法很搞笑，难道我将1T的数据都放在内存上，不是坑么，所以说这是我们为什么选择Hbase的原因之一。
软件环境
使用虚拟机安装，centos 6.3，jdk-7u45-linux-i586.rpm，hadoop-1.1.2.tar.gz，hbase-0.96.1.1-hadoop1-bin.tar.gz，zookeeper-3.4.5-1374045102000.tar.gz
第一步，配置网络
1、配置IP
使用ifconfig 查看IP，可以使用vi /etc/sysconfig/network-scripts/ifcfg-eth0 ，ifcfg-eth0为网卡，根据自己的需求配置，如下：
DEVICE=eth0
BOOTPROTO=static
ONBOOT=yes
IPADDR=192.168.1.110
NETMASK=255.255.255.0
TYPE=Ethernet
GATEWAY=192.168.1.1
里面MAC地址不需要修改。
2、配置主机名network与DNS
使用vi /etc/sysconfig/network，修改主机名与网关，这里的网关可填可不填，如下
NETWORKING=yes
NETWORKING_IPV6=no
HOSTNAME=master
GATEWAY=192.168.1.1
使用vi /etc/resolv.conf ，添加DNS，这里可以添加你省内常用的DNS，什么8.8.8.8就算了，太慢了，添加DNS为了使用yum安装程序，如下
nameserver 202.106.0.20
nameserver 192.168.1.1
配置vi /etc/hosts 文件，如下
192.168.1.110master
192.168.1.111node1
192.168.1.112node2
3、配置yum源与安装基础软件包
配置yum源，为了更方便的安装程序包，我使用的是163的源，国内比较快，当然你也可以不配置yum源，下载CentOS-Base.repo，地址：http://mirrors.163.com/.help/centos.html，上传yum文件夹中的文件到/etc/yum.repos.d中，覆盖文件，然后yum makecache 更新源，使用yum源安装：
# yum -y install lrzsz gcc gcc-c++ libstdc++-devel ntp 安装配置文件
4、同步时间与地区
cp /usr/share/zoneinfo/Asia/Shanghai /etc/localtime
将地区配置为上海，再使用ntpdate更新时间，再使用crontab -e，添加如下
# 30 23 * * * /usr/sbin/ntpdate cn.pool.ntp.org ; hwclock -w
每天晚上23:30更新一下时间
5、关闭防火墙，SELINUX=disable
# service iptables stop
# vi /etc/selinux/config
SELINUX=disabled
第二步，安装java，配置ssh，添加用户
1、安装java1.7
将所有的软件包都放到/opt文件夹中
# chmod -R 777 /opt
# rpm -ivh jdk-7u45-linux-i586.rpm
[root@master opt]# vi /etc/profile.d/java_hadoop.sh
export JAVA_HOME=/usr/java/jdk1.7.0_45/
export PATH=$PATH:$JAVA_HOME/bin
[root@master opt]# source /etc/profile
[root@master opt]# echo $JAVA_HOME
/usr/java/jdk1.7.0_45/
2、添加hadoop用户，配置无ssh登录
# groupadd hadoop
# useradd hadoop -g hadoop
# su hadoop
$ ssh-keygen -t dsa -P '' -f /home/hadoop/.ssh/id_dsa
$ cp id_dsa.pub authorized_keys
$ chmod go-wx authorized_keys
[hadoop@master .ssh]$ ssh master
Last login: Sun Mar 23 23:16:00 2014 from 192.168.1.110
做到了以上工作，就可以开始安装hadoop了。
为要选择使用hadoop1，可以去看看51cto向磊的博客http://slaytanic.blog.51cto.com/2057708/1397396，就是因为产品比较成熟，稳定。
开始安装：
[root@master opt]# ls
hadoop-1.1.2.tar.gz
hbase-0.96.1.1-hadoop1-bin.tar.gz
zookeeper-3.4.5-1374045102000.tar.gz
#将需要的文件与安装包上传，hbase要与hadoop版本一致，在hbase/lib/hadoop-core-1.1.2.jar，为正确。
[root@master opt]# mkdir -p /opt/modules/hadoop/
[root@master hadoop]# chown -R hadoop:hadoop /opt/modules/hadoop/*
[root@master hadoop]# ll
total 148232
-rwxrwxrwx 1 hadoop hadoop 61927560 Oct 29 11:16 hadoop-1.1.2.tar.gz
-rwxrwxrwx 1 hadoop hadoop 73285670 Mar 24 12:57 hbase-0.96.1.1-hadoop1-bin.tar.gz
-rwxrwxrwx 1 hadoop hadoop 16402010 Mar 24 12:57 zookeeper-3.4.5-1374045102000.tar.gz
#新建/opt/modules/hadoop/文件夹，把需要的软件包都复制到文件夹下，将文件夹的权限使用者配置为hadoop用户组和hadoop用户，其实这一步可以最后来做。
[root@master hadoop]# tar -zxvf hadoop-1.1.2.tar.gz
#解压hadoop-1.1.2.tar.gz文件
[root@master hadoop]# cat /etc/profile.d/java_hadoop.sh
export JAVA_HOME=/usr/java/jdk1.7.0_45/
export HADOOP_HOME=/opt/modules/hadoop/hadoop-1.1.2/
export PATH=$PATH:$JAVA_HOME/bin:$HADOOP_HOME/bin
#使用vi，新建一下HADOOP_HOME的变量，名称不能自定义(不然会很麻烦)。
[root@master opt]# source /etc/profile
#更新环境变量。
[root@master hadoop]# echo $HADOOP_HOME
/opt/modules/hadoop/hadoop-1.1.2
#打印一下$HADOOP_HOME，就输出了我们的地址。
[hadoop@master conf]$ vi /opt/modules/hadoop/hadoop-1.1.2/conf/hadoop-env.sh
export HADOOP_HEAPSIZE=64
#这里是内存大少的配置，我的虚拟机配置为64m。
[root@master conf]# mkdir -p /data/
#新建一下data文件夹，准备将所有的数据，存放在这个文件夹下。
[root@master conf]# mkdir -p /data/hadoop/hdfs/data
[root@master conf]# mkdir -p /data/hadoop/hdfs/name
[root@master conf]# mkdir -p /data/hadoop/hdfs/namesecondary/
[hadoop@master bin]$ mkdir -p /data/hadoop/mapred/mrlocal
[hadoop@master bin]$ mkdir -p /data/hadoop/mapred/mrsystem
[root@master conf]#su
#切换到root用户，不然是不可以修改权限的。
[root@master conf]# chown -R hadoop:hadoop /data/
#将data文件夹的权限使用者配置为hadoop用户组和hadoop用户
[root@master conf]# su hadoop
#切换到hadoop用户
[hadoop@master conf]$ chmod go-w /data/hadoop/hdfs/data/
#这里步非常重要，就是去除其他用户写入hdfs数据，可以配置为755
[hadoop@master conf]$ ll /data/hadoop/hdfs/
drwxr-xr-x 2 hadoop hadoop 4096 Mar 24 13:21 data
drwxrwxr-x 2 hadoop hadoop 4096 Mar 24 13:21 name
drwxrwxr-x 2 hadoop hadoop 4096 Mar 24 13:20 namesecondary
接下来就配置xml文件：
[hadoop@master conf]$ cd /opt/modules/hadoop/hadoop-1.1.2/conf
[hadoop@master conf]$ vi core-site.xml
fs.default.name
hdfs://master:9000
fs.checkpoint.dir
/data/hadoop/hdfs/namesecondary
fs.checkpoint.period
1800
fs.checkpoint.size
33554432
io.compression.codecs
org.apache.hadoop.io.compress.DefaultCodec,org.apache.hadoop.io.compress.GzipCodec,org.apache
.hadoop.io.compress.BZip2Codec
fs.trash.interval
1440
[hadoop@master conf]$ vi hdfs-site.xml
dfs.name.dir
/data/hadoop/hdfs/name
dfs.data.dir
/data/hadoop/hdfs/data
dfs.http.address
master:50070
dfs.secondary.http.address
node1:50090
dfs.replication
3
dfs.datanode.du.reserved
873741824
dfs.block.size
134217728
dfs.permissions
false
[hadoop@master conf]$ vi mapred-site.xml
mapred.job.tracker
master:9001
mapred.local.dir
/data/hadoop/mapred/mrlocal
true
mapred.system.dir
/data/hadoop/mapred/mrsystem
true
mapred.tasktracker.map.tasks.maximum
2
true
mapred.tasktracker.reduce.tasks.maximum
1
true
io.sort.mb
32
true
mapred.child.java.opts
-Xmx128M
mapred.compress.map.output
true
#配置文件完成
#切换到master主机，开启
[root@master hadoop]# vi/opt/modules/hadoop/hadoop-1.1.2/conf/masters
node1
node2
[root@master hadoop]# vi/opt/modules/hadoop/hadoop-1.1.2/conf/slaves
master
node1
node2
#节点配置，这里很重要，masters文件不需要把masters加入
hadoop1.1.2的node1，node2配置开始
前面的网络的配置我就不说了：
#登录master，将authorized_keys，发过去
[root@master ~]# scp /home/hadoop/.ssh/authorized_keys root@node1:/home/hadoop/.ssh/
[root@master ~]# scp /home/hadoop/.ssh/authorized_keys root@node1:/home/hadoop/.ssh/
The authenticity of host 'node1 (192.168.1.111)' can't be established.
RSA key fingerprint is 0d:aa:04:89:28:44:b9:e8:bb:5e:06:d0:dc:de:22:85.
Are you sure you want to continue connecting (yes/no)? yes
Warning: Permanently added 'node1,192.168.1.111' (RSA) to the list of known hosts.
root@node1's password:
#切换到node1主机
[root@master ~]# su hadoop
[hadoop@master root]$ ssh master
Last login: Sun Mar 23 23:17:06 2014 from 192.168.1.110
[root@master hadoop]# vi/opt/modules/hadoop/hadoop-1.1.2/conf/masters
node1
node2
[root@master hadoop]# vi/opt/modules/hadoop/hadoop-1.1.2/conf/slaves
master
node1
node2
#切换到master主机，开启
[hadoop@master conf]$ hadoop namenode -format
Warning: $HADOOP_HOME is deprecated.
14/03/24 13:33:52 INFO namenode.NameNode: STARTUP_MSG:
/************************************************************
STARTUP_MSG: Starting NameNode
STARTUP_MSG: host = master/192.168.1.110
STARTUP_MSG: args = [-format]
STARTUP_MSG: version = 1.1.2
STARTUP_MSG: build = https://svn.apache.org/repos/asf/hadoop/common/branches/branch-1.1 -r 1440782; compiled by 'hortonfo' on Thu Jan 31 02:03:24 UTC 2013
************************************************************/
Re-format filesystem in /data/hadoop/hdfs/name ? (Y or N) Y
14/03/24 13:33:54 INFO util.GSet: VM type = 32-bit
14/03/24 13:33:54 INFO util.GSet: 2% max memory = 0.61875 MB
14/03/24 13:33:54 INFO util.GSet: capacity = 2^17 = 131072 entries
14/03/24 13:33:54 INFO util.GSet: recommended=131072, actual=131072
14/03/24 13:33:55 INFO namenode.FSNamesystem: fsOwner=hadoop
14/03/24 13:33:55 INFO namenode.FSNamesystem: supergroup=supergroup
14/03/24 13:33:55 INFO namenode.FSNamesystem: isPermissionEnabled=false
14/03/24 13:33:55 INFO namenode.FSNamesystem: dfs.block.invalidate.limit=100
14/03/24 13:33:55 INFO namenode.FSNamesystem: isAccessTokenEnabled=false accessKeyUpdateInterval=0 min(s), accessTokenLifetime=0 min(s)
14/03/24 13:33:55 INFO namenode.NameNode: Caching file names occuring more than 10 times
14/03/24 13:33:55 INFO common.Storage: Image file of size 112 saved in 0 seconds.
14/03/24 13:33:56 INFO namenode.FSEditLog: closing edit log: position=4, editlog=/data/hadoop/hdfs/name/current/edits
14/03/24 13:33:56 INFO namenode.FSEditLog: close success: truncate to 4, editlog=/data/hadoop/hdfs/name/current/edits
14/03/24 13:33:56 INFO common.Storage: Storage directory /data/hadoop/hdfs/name has been successfully formatted.
14/03/24 13:33:56 INFO namenode.NameNode: SHUTDOWN_MSG:
/************************************************************
SHUTDOWN_MSG: Shutting down NameNode at master/192.168.1.110
************************************************************/
[hadoop@master bin]$ start-all.sh
[hadoop@master bin]$ jps
7603 TaskTracker
7241 DataNode
7119 NameNode
7647 Jps
7473 JobTracker
zookeeper-3.4.5配置开始
#在master机器上安装，为namenode
[root@master hadoop]# tar -zxvf zookeeper-3.4.5-1374045102000.tar.gz
[root@master hadoop]# chown -R hadoop:hadoop zookeeper-3.4.5
[root@master hadoop]# vi /opt/modules/hadoop/zookeeper-3.4.5/conf/zoo.cfg
# The number of milliseconds of each tick
tickTime=2000
# The number of ticks that the initial
# synchronization phase can take
initLimit=10
# The number of ticks that can pass between
# sending a request and getting an acknowledgement
syncLimit=5
# the directory where the snapshot is stored.
# do not use /tmp for storage, /tmp here is just
# example sakes.
dataDir=/data/zookeeper
# the port at which the clients will connect
clientPort=2181
server.1=192.168.1.110:2888:3888
server.2=192.168.1.111:2888:3888
server.3=192.168.1.112:2888:3888
#新建文件myid(在zoo.cfg 配置的dataDir目录下，此处为/home/hadoop/zookeeper)，使得myid中的值与server的编号相同，比如namenode上的myid： 1。datanode1上的myid：2。以此类推。
# Be sure to read the maintenance section of the
# administrator guide before turning on autopurge.
#
# http://zookeeper.apache.org/doc/current/zookeeperAdmin.html#sc_maintenance
#
# The number of snapshots to retain in dataDir
#autopurge.snapRetainCount=3
# Purge task interval in hours
# Set to "0" to disable auto purge feature
#autopurge.purgeInterval=1
#开始配置。
[root@master hadoop]# mkdir -p /data/zookeeper/
[root@master hadoop]# chown -R hadoop:hadoop /data/zookeeper/
[root@master hadoop]# echo "1" > /data/zookeeper/myid
[root@master hadoop]# cat /data/zookeeper/myid
1
[root@node1 zookeeper-3.4.5]# chown -R hadoop:hadoop /data/zookeeper/*
[root@master hadoop]# scp -r /opt/modules/hadoop/zookeeper-3.4.5/ root@node1:/opt/modules/hadoop/
#将/opt/modules/hadoop/zookeeper-3.4.5发送到node1节点，新增一个myid为2
#切换到node1
[root@node1 data]# echo "2" > /data/zookeeper/myid
[root@node1 data]# cat /data/zookeeper/myid
2
[root@node1 zookeeper-3.4.5]# chown -R hadoop:hadoop /opt/modules/hadoop/zookeeper-3.4.5
[root@node1 zookeeper-3.4.5]# chown -R hadoop:hadoop /data/zookeeper/*
#切换到master
[root@master hadoop]# su hadoop
[hadoop@master hadoop]$ cd zookeeper-3.4.5
[hadoop@master bin]$ ./zkServer.sh start
JMX enabled by default
Using config: /opt/modules/hadoop/zookeeper-3.4.5/bin/../conf/zoo.cfg
Starting zookeeper ... STARTED
[hadoop@master bin]$ jps
5507 NameNode
5766 JobTracker
6392 Jps
6373 QuorumPeerMain
5890 TaskTracker
5626 DataNode
[root@node1 zookeeper-3.4.5]# su hadoop
[hadoop@node1 zookeeper-3.4.5]$ cd bin/
[hadoop@node1 bin]$ ./zkServer.sh start
JMX enabled by default
Using config: /opt/modules/hadoop/zookeeper-3.4.5/bin/../conf/zoo.cfg
Starting zookeeper ... STARTED
[hadoop@node1 bin]$ jps
5023 SecondaryNameNode
5120 TaskTracker
5445 Jps
4927 DataNode
5415 QuorumPeerMain
#两边开启之后，就测试一下Mode: follower代表正常
[hadoop@master bin]$ ./zkServer.sh status
JMX enabled by default
Using config: /opt/modules/hadoop/zookeeper-3.4.5/bin/../conf/zoo.cfg
Mode: follower
-----------------------------------zookeeper-3.4.5配置结束-----------------------------------
[root@master ~]# su hadoop
[hadoop@master root]$ cd /opt/modules/hadoop/zookeeper-3.4.5/bin/
[hadoop@master bin]$ ./zkServer.sh start
hbase配置开始，三台机器都需要的
[root@master hadoop]# tar -zxvf hbase-0.96.1.1-hadoop1-bin.tar.gz
#解压文件
[root@master hadoop]# vi /etc/profile.d/java_hadoop.sh
export JAVA_HOME=/usr/java/jdk1.7.0_45/
export HADOOP_HOME=/opt/modules/hadoop/hadoop-1.1.2/
export HBASE_HOME=/opt/modules/hadoop/hbase-0.96.1.1/
export HBASE_CLASSPATH=/opt/modules/hadoop/hadoop-1.1.2/conf/
export HBASE_MANAGES_ZK=true
export PATH=$PATH:$JAVA_HOME/bin:$HADOOP_HOME/bin:$HBASE_HOME/bin
#配置环境变量。
[root@master hadoop]# source /etc/profile
[root@master hadoop]# echo $HBASE_CLASSPATH
/opt/modules/hadoop/hadoop-1.1.2/conf/
[root@master conf]# vi /opt/modules/hadoop/hbase-0.96.1.1/conf/hbase-site.xml
hbase.rootdir
hdfs://master:9000/hbase
hbase.cluster.distributed
true
hbase.zookeeper.quorum
master,node1,node2
hbase.zookeeper.property.dataDir
/data/zookeeper
[root@master conf]# cat /opt/modules/hadoop/hbase-0.96.1.1/conf/regionservers
master
node1
node2
[root@master conf]# chown -R hadoop:hadoop /opt/modules/hadoop/hbase-0.96.1.1
[root@master hadoop]# su hadoop
[hadoop@master hadoop]$ ll
total 148244
drwxr-xr-x 16 hadoop hadoop 4096 Mar 24 13:36 hadoop-1.1.2
-rwxrwxrwx 1 hadoop hadoop 61927560 Oct 29 11:16 hadoop-1.1.2.tar.gz
drwxr-xr-x 7 hadoop hadoop 4096 Mar 24 22:40 hbase-0.96.1.1
-rwxrwxrwx 1 hadoop hadoop 73285670 Mar 24 12:57 hbase-0.96.1.1-hadoop1-bin.tar.gz
drwxr-xr-x 10 hadoop hadoop 4096 Nov 5 2012 zookeeper-3.4.5
-rwxrwxrwx 1 hadoop hadoop 16402010 Mar 24 12:57 zookeeper-3.4.5-1374045102000.tar.gz
[root@master hadoop]$ scp -r hbase-0.96.1.1 node1:/opt/modules/hadoop
[root@master hadoop]$ scp -r hbase-0.96.1.1 node2:/opt/modules/hadoop
[root@node1 hadoop]# chown -R hadoop:hadoop /opt/modules/hadoop/hbase-0.96.1.1
[root@node2 hadoop]# chown -R hadoop:hadoop /opt/modules/hadoop/hbase-0.96.1.1
[root@node2 hadoop]# su hadoop
[hadoop@node2 bin]$ hbase shell
#进入hbase
[root@master conf]# jps
17616 QuorumPeerMain
20282 HRegionServer
20101 HMaster
9858 JobTracker
9712 DataNode
9591 NameNode
29655 Jps
9982 TaskTracker
