hadoop高可用集群(HA模式)
一、安装前
1. 集群规划
hostname | ip | NN | JN | DN | ZKFC | ZK | RM | NM |
---|---|---|---|---|---|---|---|---|
hadoopHA01 | 192.168.10.201 | yes | yes | yes | ||||
hadoopHA02 | 192.168.10.202 | yes | yes | yes | yes | yes | yes | |
hadoopHA03 | 192.168.10.203 | yes | yes | yes | yes | yes | ||
hadoopHA04 | 192.168.10.204 | yes | yes | yes | yes |
2. 安装前配置
1)创建第一台虚拟机(hadoopHA01):
虚拟机系统镜像:CentOS-7.5-x86_64-DVD-1804.iso
系统版本:centos7.5(GNU/Linux)
内存:5G
处理器:4U
硬盘:20G
[root@hadoopHA01 hadoop]# cat /etc/os-release
NAME="CentOS Linux"
VERSION="7 (Core)"
…………
[root@hadoopHA01 hadoop]# uname -a
Linux hadoopHA01 3.10.0-862.el7.x86_64 #1 SMP Fri Apr 20 16:44:24 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux
[root@hadoopHA01 hadoop]# cat /proc/cpuinfo | grep processor | wc -l
4
[root@hadoopHA01 hadoop]# free -m
total used free shared buff/cache available
Mem: 3773 147 1548 11 2076 3283
Swap: 255 0 255
[root@hadoopHA01 hadoop]# lsblk | grep disk
sda 8:0 0 20G 0 disk
2)基本配置
- 网络
[root@hadoopHA01 hadoop]# cat /etc/sysconfig/network-scripts/ifcfg-ens33
…………
IPADDR=192.168.10.201
NETMASK=255.255.255.0
GATEWAY=192.168.10.2
DNS1=192.168.10.2
- 修改主机名
[root@hadoopHA01 hadoop]# cat /etc/hostname
hadoopHA01
- hosts文件,添加集群所有节点的ip/主机名映射
[root@hadoopHA01 hadoop]# cat /etc/hosts
…………
192.168.10.201 hadoopHA01
192.168.10.202 hadoopHA02
192.168.10.203 hadoopHA03
192.168.10.204 hadoopHA04
- 创建分发脚本
[root@hadoopHA01 opt]# cat /root/bin/xsync
#!/bin/bash
#1. 判断参数个数
if [ $# -lt 1 ]
then
echo Not Enough Arguement!
exit;
fi
#2. 遍历集群所有机器
for host in hadoopHA02 hadoopHA03 hadoopHA04
do
echo ==================== $host ====================
#3. 遍历所有目录,挨个发送
for file in $@
do
#4. 判断文件是否存在
if [ -e $file ]
then
#5. 获取父目录
pdir=$(cd -P $(dirname $file); pwd)
#6. 获取当前文件的名称
fname=$(basename $file)
ssh $host "mkdir -p $pdir"
rsync -av $pdir/$fname $host:$pdir
else
echo $file does not exists!
fi
done
done
3. 安装jdk与hadoop
- 传包
将jdk与hadoop安装包上传到新建的/opt/packages目录下
[root@hadoopHA01 hadoop]# mkdir /opt/packages/
[root@hadoopHA01 hadoop]# ll /opt/packages/
总用量 520600
-rw-r--r-- 1 root root 338075860 4月 30 16:57 hadoop-3.1.3.tar.gz
-rw-r--r-- 1 root root 195013152 4月 30 16:57 jdk-8u212-linux-x64.tar.gz
- 解压即安装
分别解压到新建的/opt/software目录下
[root@hadoopHA01 packages]# mkdir /opt/software/
[root@hadoopHA01 packages]# tar xvf jdk-8u212-linux-x64.tar.gz -C /opt/software/
[root@hadoopHA01 packages]# tar xvf hadoop-3.1.3.tar.gz -C /opt/software/
- 创建软连接
[root@hadoopHA01 software]# ln -s hadoop-3.1.3/ hadoop
[root@hadoopHA01 software]# ll
总用量 0
lrwxrwxrwx 1 root root 13 4月 30 17:46 hadoop -> hadoop-3.1.3/
drwxr-xr-x 9 atguigu atguigu 149 9月 12 2019 hadoop-3.1.3
drwxr-xr-x 7 10 143 245 4月 2 2019 jdk1.8.0_212
- 配置环境变量
新建/etc/profile.d/my_env.sh文件写入以下内容
[root@hadoopHA01 packages]# cat /etc/profile.d/my_env.sh
#JAVA_HOME
export JAVA_HOME=/opt/software/jdk1.8.0_212
export PATH=$PATH:$JAVA_HOME/bin
#HADOOP_HOME
export HADOOP_HOME=/opt/software/hadoop-3.1.3
export PATH=$PATH:$HADOOP_HOME/bin
export PATH=$PATH:$HADOOP_HOME/sbin
- source使环境变量生效
[root@hadoopHA01 packages]# source /etc/profile
- 验证安装
[root@hadoopHA01 packages]# java -version
java version "1.8.0_212"
Java(TM) SE Runtime Environment (build 1.8.0_212-b10)
Java HotSpot(TM) 64-Bit Server VM (build 25.212-b10, mixed mode)
[root@hadoopHA01 packages]# hadoop version
Hadoop 3.1.3
Source code repository https://gitbox.apache.org/repos/asf/hadoop.git -r ba631c436b806728f8ec2f54ab1e289526c90579
Compiled by ztang on 2019-09-12T02:47Z
Compiled with protoc 2.5.0
From source with checksum ec785077c385118ac91aadde5ec9799
This command was run using /opt/software/hadoop-3.1.3/share/hadoop/common/hadoop-common-3.1.3.jar
4. 克隆虚拟机与互信配置
从配置好的hadoopHA01克隆出hadoopHA02、hadoopHA03、hadoopHA04,分别修改ip与hostname,重启
以hadoopHA02为例:
[root@hadoopHA02 ~]# ifconfig ens33
ens33: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500
inet 192.168.10.202 netmask 255.255.255.0 broadcast 192.168.10.255
…………
[root@hadoopHA02 ~]# hostname
hadoopHA02
- 互信配置
原则:
hadoopHA01 与所有节点免密方便管理集群
hadoopHA02 与各zk节点免密方便管理zk
nn主备之间需要免密
hadoopHA01 节点所有节点免密
[root@hadoopHA01 hadoop]# history | grep ssh
289 ssh-keygen -t rsa
290 ssh-copy-id hadoopHA02
291 ssh-copy-id hadoopHA03
292 ssh-copy-id hadoopHA04
296 ssh-copy-id hadoopHA01
hadoopHA02与本机以及hadoopHA01 的免密
[root@hadoopHA02 .ssh]# history | grep ssh
294 ssh-keygen -t rsa
295 ssh-copy-id hadoopHA02
301 ssh-copy-id hadoopHA01
302 ssh-copy-id hadoopHA03
303 ssh-copy-id hadoopHA04
5. 搭建zookeeper集群
1)先配置hadoopHA02 单机
- 上传zk安装包到hadoopHA02 ,解压,创建软连接
[root@hadoopHA02 zookeeper]# history
306 mkdir pacakges
307 cd pacakges/
309 tar xvf apache-zookeeper-3.5.7-bin.tar.gz -C ../software/
312 ln -s apache-zookeeper-3.5.7-bin/ zookeeper
- zk配置:
cd zookeeper/conf/
[root@hadoopHA02 zookeeper]# cat conf/zoo.cfg
…………
dataDir=/opt/software/zookeeper/data
# the port at which the clients will connect
clientPort=2181
server.1=hadoopHA02:2888:3888
server.2=hadoopHA03:2888:3888
server.3=hadoopHA04:2888:3888
…………
- 添加环境变量
[root@hadoopHA02 zookeeper]# tail -5 /etc/profile.d/my_env.sh
#ZOOkEEPER
export ZOOKEEPER_HOME=/opt/software/zookeeper
export PATH=$PATH:$ZOOKEEPER_HOME/bin
[root@hadoopHA02 zookeeper]# source /etc/profile
- 在zoo.cfg中dataDir指定的目录下 添加mid文件
[root@hadoopHA02 zookeeper]# cat /opt/software/zookeeper/data/myid
1
2)分发到hadoopHA03、hadoopHA04
[root@hadoopHA02 software]# cd /opt/software/
[root@hadoopHA02 software]# xsync apache-zookeeper-3.5.7-bin/
[root@hadoopHA02 software]# xsync zookeeper
hadoopHA03、hadoopHA04分别修改myid文件
[root@hadoopHA02 software]# xsync /opt/software/hadoop/zk
[root@hadoopHA03 zk]# cat /opt/software/zookeeper/data/myid
2
[root@hadoopHA04 zk]# cat /opt/software/zookeeper/data/myid
3
3) 启动zk集群
三个zk节点分别启动zk,以hadoopHA04为例
[root@hadoopHA04 data]# zkServer.sh start
二、HDFS
1. 配置hdfs
- core-site.xml
<property>
<!-- 配置hadoop使用的hdfs-->
<name>fs.defaultFS</name>
<value>hdfs://hacluster</value>
</property>
<property>
<!-- 配置zkfc使用的zookeeper地址-->
<name>ha.zookeeper.quorum</name>
<value>hadoopHA02:2181,hadoopHA03:2181,hadoopHA04:2181</value>
</property>
- hdfs-site.xml
<property>
<!-- namenode数据存放位置-->
<name>dfs.namenode.data.dir</name>
<value>/opt/software/hadoop/hadata/dfs/name</value>
</property>
<property>
<!-- datanode数据存放位置-->
<name>dfs.datanode.data.dir</name>
<value>/opt/software/hadoop/hadata/dfs/data</value>
</property>
<property>
<!-- journalnode数据存放位置 -->
<name>dfs.journalnode.edits.dir</name>
<value>/opt/software/hadoop/hadata/dfs/journal</value>
</property>
<property>
<!-- 配置hdfs的名字-->
<name>dfs.replication</name>
<value>2</value>
</property>
<property>
<!-- 配置hdfs的名字-->
<name>dfs.nameservices</name>
<value>hacluster</value>
</property>
<property>
<!-- 配置hdfs的nn,这里只是一个名字,下面会定义每个nn与物理主机的映射-->
<name>dfs.ha.namenodes.hacluster</name>
<value>nn1,nn2</value>
</property>
<property>
<!-- 配置nn,与物理主机的映射-->
<name>dfs.namenode.rpc-address.hacluster.nn1</name>
<value>hadoopHA01:8020</value>
</property>
<property>
<!-- 配置nn,与物理主机的映射-->
<name>dfs.namenode.rpc-address.hacluster.nn2</name>
<value>hadoopHA02:8020</value>
</property>
<property>
<!-- 配置nn的前台访问地址-->
<name>dfs.namenode.http-address.hacluster.nn1</name>
<value>hadoopHA01:9870</value>
</property>
<property>
<!-- 配置nn的前台访问地址-->
<name>dfs.namenode.http-address.hacluster.nn2</name>
<value>hadoopHA02:9870</value>
</property>
<property>
<!-- qjournal为协议名,后面ip:port是在配置journalnode的主机,hacluster是jn上的目录-->
<name>dfs.namenode.shared.edits.dir</name>
<value>qjournal://hadoopHA02:8485;hadoopHA03:8485;hadoopHA04:8485/hacluster</value>
</property>
<!-- 代理配置 -->
<property>
<name>dfs.client.failover.proxy.provider.hacluster</name>
<value>org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider</value>
</property>
<property>
<!-- 免密方式 -->
<name>dfs.ha.fencing.methods</name>
<value>sshfence</value>
</property>
<property>
<!-- 免密密钥位置 -->
<name>dfs.ha.fencing.ssh.private-key-files</name>
<value>/root/.ssh/id_rsa</value>
</property>
<property>
<!-- 自动故障转移,true时在启动nn的时候自动启动zkfc(与nn同节点) -->
<name>dfs.ha.automatic-failover.enabled</name>
<value>true</value>
</property>
2. 初始化启动hdfs集群
- 启动所有journalnode
以hadoopHA04 为例
[root@hadoopHA04 ~]# hdfs --daemon start journalnode
- 任选一个namenode格式化,并启动namenode,第一个启动为active节点
[root@hadoopHA01 sbin]# hdfs namenode -format
[root@hadoopHA01 sbin]# hdfs --daemon start namenode
- 另一个namenode同步主namenode,为standby节点
[root@hadoopHA02 current]# hdfs namenode -bootstrapStandby
- 设置集群用户
将以下内容加入到start-dfs.sh stop-dfs.sh文件开头
HDFS_DATANODE_USER=root
HDFS_NAMENODE_USER=root
HDFS_JOURNALNODE_USER=root
HDFS_ZKFC_USER=root
- 格式化zkfc,启动hdfs集群
[root@hadoopHA01 sbin]# hdfs zkfc -formatZK
[root@hadoopHA01 sbin]# start-dfs.sh
三、MapReduce与Yarn
1. 配置MapReduce
[hdfsuser01@hadoopHA01 hadoop]$ cat mapred-site.xml
<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
</configuration>
<property>
<!-- 指定MapReduce运行在Yarn上 -->
<name>mapreduce.framework.name</name>
<value>yarn</value>
</property>
<property>
<!-- 指定mr的classpath -->
<name>mapreduce.application.classpath</name>
<value>$HADOOP_MAPRED_HOME/share/hadoop/mapreduce/*:$HADOOP_MAPRED_HOME/share/hadoop/mapreduce/lib/*</value>
</property>
</configuration>
<!-- 分发配置到其他节点-->
[hdfsuser01@hadoopHA01 hadoop]$ xsync mapred-site.xml
2. 配置yarn
[hdfsuser01@hadoopHA01 hadoop]$ cat yarn-site.xml
<?xml version="1.0"?>
<configuration>
<!-- Site specific YARN configuration properties -->
<property>
<name>yarn.nodemanager.aux-services</name>
<value>mapreduce_shuffle</value>
</property>
<property>
<name>yarn.nodemanager.env-whitelist</name>
<value>JAVA_HOME,HADOOP_COMMON_HOME,HADOOP_HDFS_HOME,HADOOP_CONF_DIR,CLASSPATH_PREPEND_DISTCACHE,HADOOP_YARN_HOME,HADOOP_HOME,PATH,LANG,TZ,HADOOP_MAPRED_HOME</value>
</property>
<property>
<!-- 开启ha模式 -->
<name>yarn.resourcemanager.ha.enabled</name>
<value>true</value>
</property>
<property>
<!-- 集群名称,作为zk上RN抢锁的路径 -->
<name>yarn.resourcemanager.cluster-id</name>
<value>yarncluster</value>
</property>
<property>
<!-- RM节点名称 -->
<name>yarn.resourcemanager.ha.rm-ids</name>
<value>yarn03,yarn04</value>
</property>
<property>
<!-- RM节点的主机名 -->
<name>yarn.resourcemanager.hostname.yarn03</name>
<value>hadoopHA03</value>
</property>
<property>
<name>yarn.resourcemanager.hostname.yarn04</name>
<value>hadoopHA04</value>
</property>
<property>
<!-- RM前台访问地址 -->
<name>yarn.resourcemanager.webapp.address.yarn03</name>
<value>hadoopHA03:8088</value>
</property>
<property>
<name>yarn.resourcemanager.webapp.address.yarn04</name>
<value>hadoopHA04:8088</value>
</property>
<property>
<!-- zk节点地址 -->
<name>hadoop.zk.address</name>
<value>hadoopHA02:2181,hadoopHA03:2181,hadoopHA04:2181</value>
</property>
</configuration>
<!-- 分发配置到其他节点-->
[hdfsuser01@hadoopHA01 hadoop]$ xsync yarn-site.xml
3. 启动yarn
在hadoopHA01上执行:
[hdfsuser01@hadoopHA01 hadoop]$ start-yarn.sh
Starting resourcemanagers on [ hadoopHA03 hadoopHA04]
Starting nodemanagers
四、验证
1. 查看java进程
- hadoopHA01:
[hdfsuser01@hadoopHA01 hadoop]$ jps | grep -v Jps
3105 DFSZKFailoverController
2763 NameNode
- hadoopHA02:
[root@hadoopHA02 data]# jps | grep -v Jps
2802 DFSZKFailoverController
1443 QuorumPeerMain
2697 JournalNode
8410 NodeManager
2589 DataNode
2763 NameNode
- hadoopHA03:
[root@hadoopHA03 ~]# jps | grep -v Jps
1440 QuorumPeerMain
1910 DataNode
7421 ResourceManager
2014 JournalNode
7502 NodeManager
- hadoopHA04:
[root@hadoopHA04 ~]# jps | grep -v Jps
2005 JournalNode
7292 ResourceManager
1901 DataNode
7373 NodeManager
1439 QuorumPeerMain
2. hdfs与yarn前台页面
-
hadoopHA01:hdfs-nanenode
-
hadoopHA02:hdfs-nanenode
-
hadoopHA03:yarn-resourcemanager
-
hadoopHA04:yarn-resourcemanager