hadoop 2.8.5完全分散式環境搭建

zookeeper叢集文件:

namenode HA文件:

1.下載zookeeper

[root@node1 ~]# wget

[root@node1 ~]# tar xvf zookeeper-3.4.13.tar.gz -C /opt/

[root@node1 ~]# cd /opt/zookeeper-3.4.13/conf/

[root@node1 conf]# vim zoo.cfg 

tickTime=2000
dataDir=/opt/zookeeper-3.4.13/data
clientPort=2181
initLimit=5
syncLimit=2
server.1=node1:2888:3888
server.2=node2:2888:3888
server.3=node3:2888:3888

[root@node1 conf]# mkdir /opt/zookeeper-3.4.13/data

[root@node1 conf]# cd /opt/zookeeper-3.4.13/data       --myid必須要在data目錄下面,否則會報錯

[root@node1 data]# cat myid 

1

[root@node1 zookeeper-3.4.13]# cd ..

[root@node1 opt]# scp -r zookeeper-3.4.13 node2:/opt/

[root@node1 opt]# scp -r zookeeper-3.4.13 node3:/opt/

2.在node2修改myid檔案

[root@node2 opt]# cat /opt/zookeeper-3.4.13/data/myid 

2

[root@node2 opt]# 

3.在node3修改myid檔案

[root@node3 ~]# cat /opt/zookeeper-3.4.13/data/myid 

3

[root@node3 ~]# zkServer.sh start        --每個節點都要啟動zookeeper服務

ZooKeeper JMX enabled by default

Using config: /opt/zookeeper-3.4.13/bin/../conf/zoo.cfg

Starting zookeeper ... STARTED

[root@node3 opt]# zkCli.sh     --使用客戶端登陸

4.下載解壓hadoop安裝包(在各節點上安裝jdk和hadoop)

[root@node1 ~]# wget

[root@node1 ~]# wget

[root@node1 ~]# scp jdk-8u202-linux-x64.rpm  node2:/root/    --將jdk傳到各節點上

[root@node1 ~]# rpm -ivh jdk-8u202-linux-x64.rpm             --在各節點安裝jdk

[root@node1 ~]# tar xvf hadoop-2.8.5.tar.gz  -C /opt/

[root@node1 ~]# cd /opt/hadoop-2.8.5/etc/hadoop/

[root@node1 hadoop]# vim hadoop-env.sh 

export JAVA_HOME=/usr/java/jdk1.8.0_202-amd64/

[root@node1 hadoop]# cat slaves     --配置datanode節點(datanode節點都需要)

node2
node3
node4

[root@node1 hadoop]# 

5.配置環境變數

[root@node1 opt]# vim /etc/profile    --其它node節點一樣新增

export JAVA_HOME=/usr/java/jdk1.8.0_202-amd64

export HADOOP_HOME=/opt/hadoop-2.8.5

export ZOOKEEPER_HOME=/opt/zookeeper-3.4.13

export PATH=$PATH:$HADOOP_HOME/bin:$HADOOP_HOME/sbin:$JAVA_HOME/bin:$ZOOKEEPER_HOME/bin

[root@node1 opt]# source  /etc/profile

[root@node1 opt]# scp /etc/profile node2:/etc/

profile                                                                          100% 2037     1.4MB/s   00:00    

[root@node1 opt]# scp /etc/profile node3:/etc/

profile                                                                           100% 2037   961.7KB/s   00:00    

[root@node1 opt]# 

5.使用zookeeper做namenode高可用

[root@node1 ~]# cd /opt/hadoop-2.8.5/etc/hadoop/

[root@node1 hadoop]# vim hdfs-site.xml 

  dfs.nameservices     --指定叢集名
  mycluster


  dfs.ha.namenodes.mycluster   --指定叢集兩個namenode成員
  nn1,nn2


  dfs.namenode.rpc-address.mycluster.nn1   --nn1成員是node1
  node1:8020


  dfs.namenode.rpc-address.mycluster.nn2   --nn2成員是node4
  node4:8020


  dfs.namenode.http-address.mycluster.nn1
  node1:50070


  dfs.namenode.http-address.mycluster.nn2
  node4:50070


  dfs.namenode.shared.edits.dir
  qjournal://node2:8485;node3:8485;node4:8485/mycluster


  dfs.client.failover.proxy.provider.mycluster
  org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider


  dfs.ha.fencing.methods
  sshfence


  dfs.ha.fencing.ssh.private-key-files     --各個節點透過/root/.ssh/id_rsa檔案信任關係
  /root/.ssh/id_rsa


  dfs.ha.automatic-failover.enabled        --開啟HA功能
  true

[root@node1 hadoop]# vim core-site.xml 

  fs.defaultFS
  hdfs://mycluster

 
   ha.zookeeper.quorum
   node1:2181,node2:2181,node3:2181      --定義zookeeper
 

[root@node1 hadoop]# scp hdfs-site.xml core-site.xml node2:/opt/hadoop-2.8.5/etc/hadoop/

[root@node1 hadoop]# scp hdfs-site.xml core-site.xml node3:/opt/hadoop-2.8.5/etc/hadoop/

[root@node1 hadoop]# scp hdfs-site.xml core-site.xml node4:/opt/hadoop-2.8.5/etc/hadoop/

6.啟用journalnode(node1,2,3,4都要啟動)

[root@node2 hadoop]# hadoop-daemon.sh  start journalnode

starting journalnode, logging to /opt/hadoop-2.8.5/logs/hadoop-root-journalnode-node2.out

[root@node2 hadoop]# 

7.格式化磁碟和啟動服務

[root@node1 hadoop]# hdfs namenode -format     --在任意一個datanode就可以了

[root@node1 hadoop]# scp -r /tmp/hadoop-root/dfs node4:/tmp/hadoop-root/     --將格式化好後的資料複製到另一個namenode上(node4)

[root@node1 hadoop]# hdfs zkfc -formatZK       --初使化zookeeper

[root@node1 hadoop]# start-dfs.sh              --啟動服務

Starting namenodes on [node1 node4]

node1: starting namenode, logging to /opt/hadoop-2.8.5/logs/hadoop-root-namenode-node1.out

node4: starting namenode, logging to /opt/hadoop-2.8.5/logs/hadoop-root-namenode-node4.out

node4: starting datanode, logging to /opt/hadoop-2.8.5/logs/hadoop-root-datanode-node4.out

node2: starting datanode, logging to /opt/hadoop-2.8.5/logs/hadoop-root-datanode-node2.out

node3: starting datanode, logging to /opt/hadoop-2.8.5/logs/hadoop-root-datanode-node3.out

Starting journal nodes [node2 node3 node4]

node2: journalnode running as process 10429. Stop it first.

node4: journalnode running as process 9923. Stop it first.

node3: journalnode running as process 10198. Stop it first.

Starting ZK Failover Controllers on NN hosts [node1 node4]

node1: starting zkfc, logging to /opt/hadoop-2.8.5/logs/hadoop-root-zkfc-node1.out

node4: starting zkfc, logging to /opt/hadoop-2.8.5/logs/hadoop-root-zkfc-node4.out

[root@node1 hadoop]# jps

12728 QuorumPeerMain        --zookeeper程式

3929 DataNode               --datanode程式

15707 JournalNode           --JournalNode程式

16907 DFSZKFailoverController

16556 NameNode

17471 Jps

[root@node1 hadoop]# 

8.datanode程式

[root@node2 hadoop]# jps 

11282 DataNode

11867 Jps

10429 JournalNode

8590 QuorumPeerMain

[root@node2 hadoop]# 

9.使用web訪問