使用阿里云服务器搭建这个hadoop集群,踩得坑人已经傻了,浪费了我快一周的时间。
1.准备环节
准备三台阿里云服务器
端口设置(后面会有坑)
Namenode
通过安全组手动开放9000端口
在控制台查看Namenode的私网ip
Datanode1和Datanode2
通过安全组开放50010端口
2.配置环境
我们只需要把Namenode环境配置好,其他datanode采用主传的方式即可
配置Namenode环境
下载并解压所需包
ssh 121.196.224.191 //连接Namenode节点mkdir /home/hadoopcd /home/hadoop/wget /cloudcomputing/download/hadoop.tar.gztar -zxvf hadoop.tar.gzmkdir /home/jdkcd /home/jdk/wget /cloudcomputing/download/jdk.tar.gztar -zxvf jdk.tar.gz
设置bash profile
vi ~/.bash_profile
将原有的配置替换为下面代码
#export PATHexport JAVA_HOME=/home/jdk/jdk1.7.0_75export JAVA_JRE=/home/jdk/jdk1.7.0_75/jreexport HADOOP_HOME=/home/hadoop/hadoop-2.6.0# pathexport PATH=$HADOOP_HOME/bin:$HADOOP_HOME/sbin:$PATHexport PATH=$JAVA_HOME/bin:$PATH
source ~/.bash_profile //保存配置
修改Hadoop配置文件(以下都是将原有的配置进行替换)
cd $HADOOP_HOME
mkdir namenode
mkdir datanode
cd etc/hadoop/
vi core-site.xml
<configuration><property><name>fs.default.name</name><value>hdfs://Namenode:9000</value></property><property><name>hadoop.tmp.dir</name><value>/home/hadoop/hadoop-2.6.0/tmp/hadoop-${user.name}</value></property></configuration>
vi hdfs-site.xml
<configuration><property><name>dfs.replication</name><value>2</value></property><property><name>dfs.namenode.name.dir</name><value>/home/hadoop/hadoop-2.6.0/namenode/name_1, /home/hadoop/hadoop-2.6.0/namenode/name_2</value></property><property><name>dfs.datanode.data.dir</name><value>/home/hadoop/hadoop-2.6.0/datanode/data_1, /home/hadoop/hadoop-2.6.0/datanode/data_2</value></property></configuration>
vi mapred-site.xml
<configuration><property><name>mapred.job.tracker</name><value>Namenode:9001</value></property></configuration>
vi hadoop-env.sh
export JAVA_HOME=/home/jdk/jdk1.7.0_75
vi slaves
设置数据节点
Datanode1Datanode2
设置3个节点间的无密码访问
更改主机名称
ssh 121.196.224.191
vi /etc/hostname
Namenode
vi /etc/hosts
121.196.224.191Namenode 121.196.226.12Datanode147.96.42.166Datanode2
ssh 121.196.226.12
vi /etc/hostname
Datanode1
ssh 47.96.42.166
vi /etc/hostname
Datanode2
ssh Namenode vi /etc/hosts 将Namenode改为私网ip
将Namenode节点配置复制到datanode上
scp /etc/hosts root@121.196.226.12:/etc/hostsscp /etc/hosts root@47.96.42.166:/etc/hostsscp ~/.bash_profile root@121.196.226.12:~/.bash_profilescp ~/.bash_profile root@47.96.42.166:~/.bash_profile
进行密钥生成
ssh-keygen -t rsa //点击三次enter即可
ssh Datanode1
ssh-keygen -t rsa //点击三次enter即可
scp /root/.ssh/id_rsa.pub root@Namenode:/root/.ssh/id_rsa.pub.Datanode1
ssh Datanode2
ssh-keygen -t rsa //点击三次enter即可
scp /root/.ssh/id_rsa.pub root@Namenode:/root/.ssh/id_rsa.pub.Datanode2
密钥交互
ssh Namenode
cd /root/.ssh
cat id_rsa.pub >> authorized_keys
cat id_rsa.pub.Datanode1 >> authorized_keys
cat id_rsa.pub.Datanode2 >> authorized_keys
chmod 644 authorized_keys
scp ~/.ssh/authorized_keys root@Datanode1:/root/.ssh/authorized_keys
scp ~/.ssh/authorized_keys root@Datanode2:/root/.ssh/authorized_keys
测试连接
ssh Datanode1
ssh Datanode2
ssh Namenode
启动及测试hadoop
把hadoop,jdk文件和配置从namenode复制到datanode中
scp -r /home/hadoop/ root@Datanode1:/home/hadoopscp -r /home/hadoop/ root@Datanode2:/home/hadoopscp -r /home/jdk/ root@Datanode1:/home/jdkscp -r /home/jdk/ root@Datanode2:/home/jdk
cd $HADOOP_HOMEcd etc/hadoophdfs namenode -format //初始化hadoop,不要多次初始化,可能会导致生成id不唯一start-all.sh
查看HDFS文件系统状态
hdfs dfsadmin -report
jps
测试HDFS文件系统
hadoop fs -ls /
vi aaa.txt
hadoop fs -put aaa.txt /aaa.txt
恭喜我们到了这里,接下来让我们看下到这一步的比较难受的坑
namenode没有启动,jps找不到namenode,运行hdfs dfsadmin -report报如下错误
Retrying connect to server: hadoop/121.196.224.191:9000. Already tried 0 time(s);
原因:完成到datanode节点配置复制后,没有将namenode中 /etc/hosts 文件中namenode的ip改为私网ip
详细见这篇博客datanode未启动,查看datanode日志文件,发现报如下错误
Retrying connect to server: Namenode/121.196.224.191:9000. Already tried 0 time(s);
原因:没有通过安全组规则设置打开namenode的9000端口hadoop fs -put aaa.txt /aaa.txt 这步时,相应很长时间,报这种错误
INFO hdfs.DFSClient: Excluding datanode 47.96.42.166:50010
原因:没有打开47.96.42.166 这个datanode2的50010端口,通过安全组规则设置将其打开即可。
4. 如果datanode此时未启动可用 hadoop-daemon.sh start datanode 启动