节点的分配如下
node-1：NN、ZKFC、JN、ZK、DN
node-2：NN、ZKFC、JN、ZK、DN
node-3：JN、ZK、DN
环境准备
1、修改对应host
2、设置各节点ssh相互免密登陆
3、安装fuser命令
yum install -y psmisc
4、下载安装包(下载到/usr/local/src/)
jdk8 官网：https://www.oracle.com/
下载地址： https://www.oracle.com/java/technologies/javase/javase-jdk8-downloads.html
zookeeper 3.4.6 官网：https://zookeeper.apache.org/
下载地址： https://archive.apache.org/dist/zookeeper/zookeeper-3.4.6/zookeeper-3.4.6.tar.gz
hadoop-3.1.3 官网：http://hadoop.apache.org/
下载地址：https://downloads.apache.org/hadoop/common/hadoop-3.1.3/hadoop-3.1.3.tar.gz
注意 没有特别说的都在node-1上操作
部署 jdk（所有节点都要执行）
cd /usr/local/src/
tar zxf jdk-8u241-linux-x64.tar.gz
mv jdk1.8.0_241/ /usr/local/
echo 'export JAVA_HOME=/usr/local/jdk1.8.0_241' >> /etc/profile.d/ecm_env.sh
echo 'export PATH=$PATH:$JAVA_HOME/bin' >> /etc/profile
source /etc/profile部署ZooKeeper
ssh node-1 "mkdir -p /data/ZooKeeper"
ssh node-2 "mkdir -p /data/ZooKeeper"
ssh node-3 "mkdir -p /data/ZooKeeper"
tar zxf zookeeper-3.4.6.tar.gz
mv zookeeper-3.4.6 /usr/local/
# 修改conf文件夹下的zoo.cfg
[root@node-1 src]# cat /usr/local/zookeeper-3.4.6/conf/zoo.cfg
tickTime=2000
initLimit=10
syncLimit=5
dataDir=/data/ZooKeeper
clientPort=2181
server.1=node-1:2888:3888
server.2=node-2:2888:3888
server.3=node-3:2888:3888
# 创建myid
ssh node-1 "echo 1 > /data/ZooKeeper/myid"
ssh node-2 "echo 2 > /data/ZooKeeper/myid"
ssh node-3 "echo 3 > /data/ZooKeeper/myid"
# 拷贝ZooKeeper目录到其它两个节点
scp -r /usr/local/zookeeper-3.4.6/ node-2:/usr/local/
scp -r /usr/local/zookeeper-3.4.6/ node-3:/usr/local/
# 启动ZooKeeper
ssh node-1 "/usr/local/zookeeper-3.4.6/bin/zkServer.sh start"
ssh node-2 "/usr/local/zookeeper-3.4.6/bin/zkServer.sh start"
ssh node-3 "/usr/local/zookeeper-3.4.6/bin/zkServer.sh start"
# 查看ZK启动结果
[root@node-1 src]# jps
2529 QuorumPeerMain
2561 Jps
[root@node-2 src]# jps
2504 QuorumPeerMain
2540 Jps
[root@node-3 data]# jps
2518 Jps
2488 QuorumPeerMain部署HDFS
# 以下 “两行” 所有节点都要执行
echo 'export HADOOP_HOME=/usr/local/hadoop-3.1.3' >> /etc/profile.d/hdfs.sh
source /etc/profile
tar zxf hadoop-3.1.3.tar.gz
mv hadoop-3.1.3 /usr/local/
cd /usr/local/hadoop-3.1.3/etc/hadoop/
# hadoop-env.sh 文件底部添加配置如下
export HDFS_NAMENODE_USER=root
export HDFS_DATANODE_USER=root
export HDFS_JOURNALNODE_USER=root
export HDFS_ZKFC_USER=root
export HADOOP_SHELL_EXECNAME=root
# core-site.xml 配置如下
[root@node-1 hadoop]# cat core-site.xml
省略...
<configuration>
<property>
<name>fs.defaultFS</name>
<value>hdfs://firstcluster</value>
</property>
<property>
<name>hadoop.tmp.dir</name>
<value>/data/hadoop</value>
</property>
<property>
<name>hadoop.http.staticuser.user</name>
<value>root</value>
</property>
<property>
<name>hadoop.http.staticuser.user</name>
<value>root</value>
</property>
<property>
<name>ha.zookeeper.quorum</name>
<value>node-1:2181,node-2:2181,node-3:2181</value>
</property>
</configuration>
# hdfs-site.xml 配置如下
[root@node-1 hadoop]# cat hdfs-site.xml
省略...
<configuration>
<property>
<name>dfs.nameservices</name>
<value>firstcluster</value>
</property>
<property>
<name>dfs.ha.namenodes.firstcluster</name>
<value>nn1,nn2</value>
</property>
<property>
<name>dfs.namenode.rpc-address.firstcluster.nn1</name>
<value>node-1:8020</value>
</property>
<property>
<name>dfs.namenode.rpc-address.firstcluster.nn2</name>
<value>node-2:8020</value>
</property>
<property>
<name>dfs.namenode.http-address.firstcluster.nn1</name>
<value>node-1:9870</value>
</property>
<property>
<name>dfs.namenode.http-address.firstcluster.nn2</name>
<value>node-2:9870</value>
</property>
<property>
<name>dfs.namenode.shared.edits.dir</name>
<value>qjournal://node-1:8485;node-2:8485;node-3:8485/ns</value>
</property>
<property>
<name>dfs.client.failover.proxy.provider.firstcluster</name>
<value>org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider</value>
</property>
<property>
<name>dfs.ha.fencing.methods</name>
<value>sshfence</value>
</property>
<property>
<name>dfs.ha.fencing.ssh.private-key-files</name>
<value>/root/.ssh/id_rsa</value>
</property>
<property>
<name>dfs.journalnode.edits.dir</name>
<value>/data/journalnode/data</value>
</property>
<property>
<name>dfs.ha.automatic-failover.enabled</name>
<value>true</value>
</property>
</configuration>
# workers 配置如下
[root@node-1 hadoop]# cat workers
node-1
node-2
node-3
# 创建数据目录
ssh node-1 "mkdir /data/hadoop && mkdir -p /data/journalnode/data"
ssh node-2 "mkdir /data/hadoop && mkdir -p /data/journalnode/data"
ssh node-3 "mkdir /data/hadoop && mkdir -p /data/journalnode/data"
# 拷贝 hadoop-3.1.3 到其它两台节点
scp -r /usr/local/hadoop-3.1.3/ node-2:/usr/local/
scp -r /usr/local/hadoop-3.1.3/ node-3:/usr/local/
# 启动 JN
ssh node-1 "/usr/local/hadoop-3.1.3/bin/hdfs --daemon start journalnode"
ssh node-2 "/usr/local/hadoop-3.1.3/bin/hdfs --daemon start journalnode"
ssh node-3 "/usr/local/hadoop-3.1.3/bin/hdfs --daemon start journalnode"
# 查看 JN 状态
[root@node-1 src]# jps
2529 QuorumPeerMain
11795 Jps
11727 JournalNode
[root@node-2 src]# jps
11667 JournalNode
2504 QuorumPeerMain
11711 Jps
[root@node-3 data]# jps
11602 JournalNode
2488 QuorumPeerMain
11646 Jps
# 格式化 HDFS
/usr/local/hadoop-3.1.3/bin/hdfs namenode -format
# 拷贝元数据到 node-2，(只拷贝到node-2，因为node-1和node-2是nameNode)
[root@node-1 src]# scp -r /data/hadoop/ node-2:/data/
# 格式化 ZKFC
[root@node-1 src]# /usr/local/hadoop-3.1.3/bin/hdfs zkfc -formatZK
# 启动集群
[root@node-1 hadoop]# /usr/local/hadoop-3.1.3/sbin/start-dfs.sh
Starting namenodes on [node-1 node-2]
上一次登录：一 4月 6 21:14:34 CST 2020pts/0 上
Starting datanodes
上一次登录：一 4月 6 21:18:37 CST 2020pts/0 上
Starting journal nodes [node-1 node-2 node-3]
上一次登录：一 4月 6 21:18:40 CST 2020pts/0 上
Starting ZK Failover Controllers on NN hosts [node-1 node-2]
上一次登录：一 4月 6 21:19:08 CST 2020pts/0 上查看集群状态
node-1
node-2
测试高可用
目前 node-1 是 active，kill掉node-1上的 NameNode，如下图
再次查看发现 node-2 已经 active
在启动node-1上的 NameNode
[root@node-1 logs]# /usr/local/hadoop-3.1.3/bin/hdfs --daemon start namenode
Comments Closed.