Linux ->> UBuntu 14.04 LTE下安装Hadoop 1.2.1（伪分布模式）

Hadoop的运行模式可分为单机模式、伪分布模式和分布模式。

首先无论哪种模式都需要安装JDK的，这一步之前的随笔Ubuntu 14.04 LTE下安装JDK 1.8中已经做了。这里就不多说了。

其次是安装SSH。安装SSH是为了每次可以免密码登陆数据节点服务器。因为集群的环境下，每次登陆到数据节点服务器不可能每次都输入密码。这一步在前面的随笔Ubuntu 14.04 LTE下配置SSH免密码登录中已经做了。这里也不多说了。

伪分布模式安装：

首先下载Hadoop 1.2.1到本机，再解压到用户目录下。

jerry@ubuntu:~/Downloads$ tar zxf hadoop-1.2.1.tar.gz -C ~/hadoop_1.2.1
jerry@ubuntu:~/Downloads$ cd ~/hadoop_1.2.1/
jerry@ubuntu:~/hadoop_1.2.1$ ls
hadoop-1.2.1
jerry@ubuntu:~/hadoop_1.2.1$ cd hadoop-1.2.1/
jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1$ ls
bin          hadoop-ant-1.2.1.jar          ivy          sbin
build.xml    hadoop-client-1.2.1.jar       ivy.xml      share
c++          hadoop-core-1.2.1.jar         lib          src
CHANGES.txt  hadoop-examples-1.2.1.jar     libexec      webapps
conf         hadoop-minicluster-1.2.1.jar  LICENSE.txt
contrib      hadoop-test-1.2.1.jar         NOTICE.txt
docs         hadoop-tools-1.2.1.jar        README.txt

然后配置hadoop的几个配置文件，都是XML格式。

首先是core-default.xml。这里配置hadoop分布式文件系统的地址和端口，以及Hadoop临时文件目录（/tmp/hadoop-${user.name}）。

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ cat core-site.xml
<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>

<!-- Put site-specific property overrides in this file. -->

<configuration>
<property>
<name>fs.default.name</name>
<value>hdfs://localhost:9000</value>
</property>
<property>
<name>hadoop.tmp.dir</name>
<value>/hadoop/hadooptmp</value>
</property>
</configuration>
jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$

修改hadoop系统环境配置文件，告诉hadoop安装好的jdk的主目录路径

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1$ cd conf/
jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ ls
capacity-scheduler.xml      hadoop-policy.xml      slaves
configuration.xsl           hdfs-site.xml          ssl-client.xml.example
core-site.xml               log4j.properties       ssl-server.xml.example
fair-scheduler.xml          mapred-queue-acls.xml  taskcontroller.cfg
hadoop-env.sh               mapred-site.xml        task-log4j.properties
hadoop-metrics2.properties  masters
jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ sudo vim hadoop-env.sh n
[sudo] password for jerry:
2 files to edit
jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ sudo vim hadoop-env.sh
jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ tail -n 1 hadoop-env.sh
export JAVA_HOME=/usr/lib/jvm/jdk

然后是hdfs-site.xml 。修改hdfs的文件备份数量为1，dfs命名节点的主目录，dfs数据节点的目录。

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ cat hdfs-site.xml
<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>

<!-- Put site-specific property overrides in this file. -->

<configuration>
<property>
<name>dfs.replication</name>
<value>1</value>
</property>
<property>
<name>dfs.name.dir</name>
<value>/hadoop/hdfs/name</value>
</property>
<property>
<name>dfs.data.dir</name>
<value>/hadoop/hdfs/data</value>
</property>
</configuration>
jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$

最后配置mapreduce的job tracker的地址和端口

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ cat mapred-site.xml
<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>

<!-- Put site-specific property overrides in this file. -->

<configuration>
<property>
<name>mapred.job.tracker</name>
<value>localhost:9001</value>
</property>
</configuration>
jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$

配置masters文件和slaves文件，这里因为我们是伪分布式，命名节点和数据节点其实都是一样。

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ cat masters
localhost
192.168.2.100

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ cat slaves
localhost
192.168.2.100
jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$

编辑/etc/hosts文件，配置主机名和IP地址的映射关系

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ cat /etc/hosts
127.0.0.1    localhost
127.0.1.1    ubuntu

# The following lines are desirable for IPv6 capable hosts
::1     ip6-localhost ip6-loopback
fe00::0 ip6-localnet
ff00::0 ip6-mcastprefix
ff02::1 ip6-allnodes
ff02::2 ip6-allrouters
192.168.2.100 master
192.168.2.100 slave
jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$

创建好core-default.xml，hdfs-site.xml，mapred-site.xml 三个配置文件里面写到的目录

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ mkdir -p /hadoop/hadooptmp

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ mkdir -p /hadoop/hdfs/name

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ mkdir -p /hadoop/hdfs/data

格式化HDFS

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/bin$ ./hadoop namenode -format

启动所有Hadoop服务，包括JobTracker，TaskTracker，Namenode

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/bin$ ./start-all.sh
starting namenode, logging to /home/jerry/hadoop_1.2.1/hadoop-1.2.1/libexec/../logs/hadoop-jerry-namenode-ubuntu.out
192.168.68.130: starting datanode, logging to /home/jerry/hadoop_1.2.1/hadoop-1.2.1/libexec/../logs/hadoop-jerry-datanode-ubuntu.out
localhost: starting datanode, logging to /home/jerry/hadoop_1.2.1/hadoop-1.2.1/libexec/../logs/hadoop-jerry-datanode-ubuntu.out
localhost: ulimit -a for user jerry
localhost: core file size          (blocks, -c) 0
localhost: data seg size           (kbytes, -d) unlimited
localhost: scheduling priority             (-e) 0
localhost: file size               (blocks, -f) unlimited
localhost: pending signals                 (-i) 7855
localhost: max locked memory       (kbytes, -l) 64
localhost: max memory size         (kbytes, -m) unlimited
localhost: open files                      (-n) 1024
localhost: pipe size            (512 bytes, -p) 8
localhost: starting secondarynamenode, logging to /home/jerry/hadoop_1.2.1/hadoop-1.2.1/libexec/../logs/hadoop-jerry-secondarynamenode-ubuntu.out
192.168.68.130: secondarynamenode running as process 10689. Stop it first.
starting jobtracker, logging to /home/jerry/hadoop_1.2.1/hadoop-1.2.1/libexec/../logs/hadoop-jerry-jobtracker-ubuntu.out
192.168.68.130: starting tasktracker, logging to /home/jerry/hadoop_1.2.1/hadoop-1.2.1/libexec/../logs/hadoop-jerry-tasktracker-ubuntu.out
localhost: starting tasktracker, logging to /home/jerry/hadoop_1.2.1/hadoop-1.2.1/libexec/../logs/hadoop-jerry-tasktracker-ubuntu.out
localhost: ulimit -a for user jerry
localhost: core file size          (blocks, -c) 0
localhost: data seg size           (kbytes, -d) unlimited
localhost: scheduling priority             (-e) 0
localhost: file size               (blocks, -f) unlimited
localhost: pending signals                 (-i) 7855
localhost: max locked memory       (kbytes, -l) 64
localhost: max memory size         (kbytes, -m) unlimited
localhost: open files                      (-n) 1024
localhost: pipe size            (512 bytes, -p) 8
jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/bin$

查看Hadoop服务是否启动成功

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$ jps
3472 JobTracker
3604 TaskTracker
3084 NameNode
5550 Jps
3247 DataNode
3391 SecondaryNameNode
jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/conf$

查看hadoop群集的状态

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/bin$ ./hadoop dfsadmin -report
Configured Capacity: 41083600896 (38.26 GB)
Present Capacity: 32723169280 (30.48 GB)
DFS Remaining: 32723128320 (30.48 GB)
DFS Used: 40960 (40 KB)
DFS Used%: 0%
Under replicated blocks: 0
Blocks with corrupt replicas: 0
Missing blocks: 0

-------------------------------------------------
Datanodes available: 1 (1 total, 0 dead)

Name: 127.0.0.1:50010
Decommission Status : Normal
Configured Capacity: 41083600896 (38.26 GB)
DFS Used: 40960 (40 KB)
Non DFS Used: 8360431616 (7.79 GB)
DFS Remaining: 32723128320(30.48 GB)
DFS Used%: 0%
DFS Remaining%: 79.65%
Last contact: Sat Dec 26 12:22:07 PST 2015

jerry@ubuntu:~/hadoop_1.2.1/hadoop-1.2.1/bin$

过程中遇到不少问题，这里贴下一些有用的链接：

Hadoop伪分布模式安装

hadoop配置、运行错误总结

hadoop环境配置过程中可能遇到问题的解决方案

Hadoop的datanode无法启动

Hadoop 添加删除datanode及tasktracker

hadoop datanode启动不起来