hadoop 2.x HA(QJM)安装部署规划

一、主机服务规划：

db01                                             db02                                        db03                                      db04                                             db05

namenode                                namenode

journalnode                             journalnode                        journalnode

datanode                                  datanode                              datanode                              datanode                                   datanode

zookeeper                                zookeeper                           zookeeper

ZKFC                                           ZKFC

二、环境配置

1、创建hadoop用户用于安装软件

groupadd hadoop

useradd -g hadoop hadoop

echo "dbking588" | passwd --stdin hadoop

配置环境变量：

export HADOOP_HOME=/opt/cdh-5.3.6/hadoop-2.5.0

export PATH=$HADOOP_HOME/bin:$HADOOP_HOME/sbin:$PATH:$HOME/bin

2、配置ssh免密码登录

--配置方法：

$ ssh-keygen -t rsa

$ ssh-copy-id db07.chavin.king

（ssh-copy-id方式只能用于rsa加密秘钥配置，测试对于dsa加密配置无效）

--验证：

[hadoop@db01 ~]$ ssh db02 date

Wed Apr 19 09:57:34 CST 2017

3、设置hadoop用户sudo权限

chmod u+w /etc/sudoers

echo "hadoop ALL=(root)NOPASSWD:ALL" >> /etc/sudoers

chmod u-w /etc/sudoers

4、关闭防火墙并且禁用selinux

sed -i '/SELINUX=enforcing/d' /etc/selinux/config

sed -i '/SELINUX=disabled/d' /etc/selinux/config

echo "SELINUX=disabled" >> /etc/selinux/config

sed -e 's/SELINUX=enforcing/SELINUX=disabled/d' /etc/selinux/config

service iptables stop

chkconfig iptables off

5、设置文件打开数量及最大进程数

cp /etc/security/limits.conf /etc/security/limits.conf.bak

echo "* soft nproc 32000" >>/etc/security/limits.conf

echo "* hard nproc 32000" >>/etc/security/limits.conf

echo "* soft nofile 65535" >>/etc/security/limits.conf

echo "* hard nofile 65535" >>/etc/security/limits.conf

6、配置集群时间同步服务

cp /etc/ntp.conf /etc/ntp.conf.bak

cp /etc/sysconfig/ntpd /etc/sysconfig/ntpd.bak

echo "restrict 192.168.100.0 mask 255.255.255.0 nomodify notrap" >> /etc/ntp.conf

echo "SYNC_HWCLOCK=yes" >> /etc/sysconfig/ntpd

service ntpd restart

0-59/10 * * * * /opt/scripts/sync_time.sh

# cat /opt/scripts/sync_time.sh

/sbin/service ntpd stop

/usr/sbin/ntpdate db01.chavin.king

/sbin/service ntpd start

7、安装java

[root@master ~]# vim /etc/profile

在末尾添加环境变量：

export JAVA_HOME=/usr/java/jdk1.7.0_67-cloudera

export PATH=$JAVA_HOME/bin:$JAVA_HOME/jre/bin:$PATH

export CLASSPATH=.:$JAVA_HOME/lib:$JAVA_HOME/jre/lib:$CLASSPATH

检查java是否安装成功：

[root@master ~]# java -version

8、安装hadoop软件

# cd /opt/software

# tar -zxvf hadoop-2.5.0.tar.gz -C /opt/cdh-5.3.6/

# chown -R hadoop:hadoop /opt/cdh-5.3.6/hadoop-2.5.0

三、编辑hadoop配置文件

Hadoop HA需要配置的文件主要包括以下两类加粗内容，其他部分安装hadoop完全分布式部署方法搭建就可以了：

HDFS配置文件：

etc/hadoop/hadoop-env.sh

etc/hadoop/core-site.xml

etc/hadoop/hdfs-site.xml

etc/haoop/slaves

YARN配置文件：

etc/hadoop/yarn-env.sh

etc/hadoop/yarn-site.xml

etc/haoop/slaves

MapReduce配置文件：

etc/hadoop/mapred-env.sh

etc/hadoop/mapred-site.xml

HA相关配置文件内容如下：

[hadoop@db01 hadoop]$ cat core-site.xml

<?xml version="1.0" encoding="UTF-8"?>

<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>

<!--

  Licensed under the Apache License, Version 2.0 (the "License");

  you may not use this file except in compliance with the License.

  You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

  Unless required by applicable law or agreed to in writing, software

  distributed under the License is distributed on an "AS IS" BASIS,

  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

  See the License for the specific language governing permissions and

  limitations under the License. See accompanying LICENSE file.

-->

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

<configuration>

<property>

<name>fs.defaultFS</name>

<value>hdfs://ns1</value>

</property>

        <property>

                <name>hadoop.tmp.dir</name>

                <value>/usr/local/hadoop-2.5.0/data/tmp</value>

        </property>

        <property>

                <name>fs.trash.interval</name>

                <value>7000</value>

        </property>

</configuration>

[hadoop@db01 hadoop]$ cat hdfs-site.xml

<?xml version="1.0" encoding="UTF-8"?>

<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>

<!--

  Licensed under the Apache License, Version 2.0 (the "License");

  you may not use this file except in compliance with the License.

  You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

  Unless required by applicable law or agreed to in writing, software

  distributed under the License is distributed on an "AS IS" BASIS,

  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

  See the License for the specific language governing permissions and

  limitations under the License. See accompanying LICENSE file.

-->

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

<configuration>

<property>

        <name>dfs.nameservices</name>

        <value>ns1</value>

    </property>

<property>

<name>dfs.ha.namenodes.ns1</name>

<value>nn1,nn2</value>

</property>

<property>

<name>dfs.namenode.rpc-address.ns1.nn1</name>

<value>db01:8020</value>

</property>

<property>

<name>dfs.namenode.rpc-address.ns1.nn2</name>

<value>db02:8020</value>

</property>

<property>

<name>dfs.namenode.http-address.ns1.nn1</name>

<value>db01:50070</value>

</property>

<property>

<name>dfs.namenode.http-address.ns1.nn2</name>

<value>db02:50070</value>

</property>

<property>

<name>dfs.namenode.shared.edits.dir</name>

<value>qjournal://db01:8485;db02:8485;db03:8485/ns1</value>

</property>

    <property>

        <name>dfs.journalnode.edits.dir</name>

        <value>/usr/local/hadoop-2.5.0/data/dfs/jn</value>

    </property>

<property>

<name>dfs.client.failover.proxy.provider.ns1</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>/home/hadoop/.ssh/id_rsa</value>

</property>

</configuration>

[hadoop@db01 hadoop-2.5.0]$ cat etc/hadoop/yarn-site.xml

<?xml version="1.0"?>

<!--

  Licensed under the Apache License, Version 2.0 (the "License");

  you may not use this file except in compliance with the License.

  You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

  Unless required by applicable law or agreed to in writing, software

  distributed under the License is distributed on an "AS IS" BASIS,

  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

  See the License for the specific language governing permissions and

  limitations under the License. See accompanying LICENSE file.

-->

<configuration>

<property>

<name>yarn.nodemanager.aux-services</name>

                <value>mapreduce_shuffle</value>

</property>

        <property>

                <name>yarn.resourcemanager.hostname</name>

                <value>db02</value>

        </property>

<property>

                <name>yarn.log-aggregation-enable</name>

                <value>true</value>

        </property>

        <property>

                <name>yarn.log-aggregation.retain-seconds</name>

                <value>600000</value>

        </property>

</configuration>

[hadoop@db01 hadoop-2.5.0]$ cat etc/hadoop/mapred-site.xml

<?xml version="1.0"?>

<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>

<!--

  Licensed under the Apache License, Version 2.0 (the "License");

  you may not use this file except in compliance with the License.

  You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

  Unless required by applicable law or agreed to in writing, software

  distributed under the License is distributed on an "AS IS" BASIS,

  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

  See the License for the specific language governing permissions and

  limitations under the License. See accompanying LICENSE file.

-->

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

<configuration>

<property>

<name>mapreduce.framework.name</name>

                <value>yarn</value>

</property>

        <property>

                <name>mapreduce.jobhistory.address</name>

                <value>db01:10020</value>

        </property>

        <property>

                <name>mapreduce.jobhistory.webapp.address</name>

                <value>db01:19888</value>

        </property>

</configuration>

[hadoop@db01 hadoop-2.5.0]$ cat etc/hadoop/slaves

db01

db02

db03

db04

db05

在以下文件中修改Java环境变量：

etc/hadoop/hadoop-env.sh

etc/hadoop/yarn-env.sh

etc/hadoop/mapred-env.sh

创建数据目录：

/opt/cdh-5.3.6/hadoop-2.5.0/data/tmp

/opt/cdh-5.3.6/hadoop-2.5.0/data/dfs/jn

同步文件到其他节点：

$ scp /opt/cdh-5.3.6/hadoop-2.5.0 hadoop@db02:/opt/cdh-5.3.6/hadoop-2.5.0

$ scp /opt/cdh-5.3.6/hadoop-2.5.0 hadoop@db03:/opt/cdh-5.3.6/hadoop-2.5.0

$ scp /opt/cdh-5.3.6/hadoop-2.5.0 hadoop@db04:/opt/cdh-5.3.6/hadoop-2.5.0

$ scp /opt/cdh-5.3.6/hadoop-2.5.0 hadoop@db05:/opt/cdh-5.3.6/hadoop-2.5.0

四、第一次启动集群

1、启动journalnode服务

[db01]$ sbin/hadoop-daemon.sh start journalnode

[db02]$ sbin/hadoop-daemon.sh start journalnode

[db03]$ sbin/hadoop-daemon.sh start journalnode

2、格式化hdfs文件系统

[db01]$ bin/hdfs namenode -format

3、在nn1上启动namenode

[db01]$ sbin/hadoop-daemon.sh start namenode

4、在nn2节点上同步nn1节点元数据（也可以直接cp元数据）

[db02]$ bin/hdfs namenode -bootstrapStandby

5、启动nn2上的namenode服务

[db02]$ sbin/hadoop-daemon.sh start namenode

6、启动所有的datanode服务

[db01]$ sbin/hadoop-daemon.sh start datanode

[db02]$ sbin/hadoop-daemon.sh start datanode

[db03]$ sbin/hadoop-daemon.sh start datanode

[db04]$ sbin/hadoop-daemon.sh start datanode

[db05]$ sbin/hadoop-daemon.sh start datanode

7、将nn1切换成active状态

[db01]$ bin/hdfs haadmin -transitionToActive nn1

[db01]$ bin/hdfs haadmin -getServiceState nn1

[db01]$ bin/hdfs haadmin -getServiceState nn2

至此、HDFS集群启动成功。

8、对HDFS文件系统进行基本测试

文件的创建、删除、上传、读取等等

五、手工方式验证namenode active和standby节点切换

[db01]$ bin/hdfs haadmin -transitionToStandby nn1

[db01]$ bin/hdfs haadmin -transitionToActive nn2

[db01]$ bin/hdfs haadmin -getServiceState nn1

standby

[db01]$ bin/hdfs haadmin -getServiceState nn2

active

进行HDFS基本功能测试。

六、使用zookeeper实现HDFS自动故障转移

1、根据服务规划安装zookeeper集群

安装zkserver集群：

$ tar -zxvf zookeeper-3.4.5.tar.gz -C /usr/local/

$ chown -R hadoop:hadoop zookeeper-3.4.5/

$ cp zoo_sample.cfg zoo.cfg

$vi zoo.cfg    --在文件中添加下面内容

dataDir=/usr/local/zookeeper-3.4.5/data

server.1=db01:2888:3888

server.2=db02:2888:3888

server.3=db03:2888:3888

配置myid文件：

$cd /usr/local/zookeeper-3.4.5/data/

$vi myid

输入上述对应server编号1

同步安装文件到其他两个节点上：

# scp -r zookeeper-3.4.5/ db02:/usr/local/

# scp -r zookeeper-3.4.5/ db03:/usr/local/

修改各个服务器的myid文件。

分别启动zk集群服务器：

db01$ bin/zkServer.sh start

db02$ bin/zkServer.sh start

db03$ bin/zkServer.sh start

2、修改core-site.xml和hdfs-site.xml配置文件：

core-site.xml file, 添加以下内容:

<property>

   <name>ha.zookeeper.quorum</name>

   <value>db01:2181,db02:2181,db03:2181</value>

</property>

hdfs-site.xml file, 添加以下内容:

<property>

   <name>dfs.ha.automatic-failover.enabled</name>

   <value>true</value>

</property>

修改后的core-site.cml和hdfs-site.xml文件主要内容如下：

core-site.xml：

<configuration>

<property>

<name>fs.defaultFS</name>

<value>hdfs://ns1</value>

</property>

        <property>

                <name>hadoop.tmp.dir</name>

                <value>/usr/local/hadoop-2.5.0/data/tmp</value>

        </property>

        <property>

                <name>fs.trash.interval</name>

                <value>7000</value>

        </property>

        <property>

                <name>ha.zookeeper.quorum</name>

                <value>db01:2181,db02:2181,db03:2181</value>

        </property>

</configuration>

Hdfs-site.xml：

<configuration>

<property>

        <name>dfs.nameservices</name>

        <value>ns1</value>

    </property>

<property>

<name>dfs.ha.namenodes.ns1</name>

<value>nn1,nn2</value>

</property>

<property>

<name>dfs.namenode.rpc-address.ns1.nn1</name>

<value>db01:8020</value>

</property>

<property>

<name>dfs.namenode.rpc-address.ns1.nn2</name>

<value>db02:8020</value>

</property>

<property>

<name>dfs.namenode.http-address.ns1.nn1</name>

<value>db01:50070</value>

</property>

<property>

<name>dfs.namenode.http-address.ns1.nn2</name>

<value>db02:50070</value>

</property>

<property>

<name>dfs.namenode.shared.edits.dir</name>

<value>qjournal://db01:8485;db02:8485;db03:8485/ns1</value>

</property>

    <property>

        <name>dfs.journalnode.edits.dir</name>

        <value>/usr/local/hadoop-2.5.0/data/dfs/jn</value>

    </property>

<property>

<name>dfs.client.failover.proxy.provider.ns1</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>/home/hadoop/.ssh/id_rsa</value>

</property>

    <property>

       <name>dfs.ha.automatic-failover.enabled</name>

       <value>true</value>

    </property>

</configuration>

配置文件修改完成后，关闭hdfs集群，并且同步文件到其他节点：

[db01]$ sbin/stop-dfs.sh

[db01]$ scp -r etc/hadoop/core-site.xml etc/hadoop/hdfs-site.xml hadoop@db02:/opt/cdh-5.3.6/hadoop-2.5.0/etc/hadoop/

[db01]$ scp -r etc/hadoop/core-site.xml etc/hadoop/hdfs-site.xml hadoop@db03:/opt/cdh-5.3.6/hadoop-2.5.0/etc/hadoop/

[db01]$ scp -r etc/hadoop/core-site.xml etc/hadoop/hdfs-site.xml hadoop@db04:/opt/cdh-5.3.6/hadoop-2.5.0/etc/hadoop/

[db01]$ scp -r etc/hadoop/core-site.xml etc/hadoop/hdfs-site.xml hadoop@db05:/opt/cdh-5.3.6/hadoop-2.5.0/etc/hadoop/

3、hadoop初始化zookeeper

[db01]$ bin/hdfs zkfc -formatZK

可以在zkCli客户端下看到hadoop-ha的文件：

[zk: localhost:2181(CONNECTED) 3] ls /

[hadoop-ha, zookeeper]

4、启动hdfs集群

[db01]$ sbin/start-dfs.sh

七、测试自动故障转移功能

[hadoop@db01 hadoop-2.5.0]$ bin/hdfs haadmin -getServiceState nn1

standby

[hadoop@db01 hadoop-2.5.0]$ bin/hdfs haadmin -getServiceState nn2

Active

[hadoop@db02 hadoop-2.5.0]$ kill -9 25121

[hadoop@db01 hadoop-2.5.0]$ bin/hdfs haadmin -getServiceState nn1

active

[hadoop@db01 hadoop-2.5.0]$ bin/hdfs haadmin -getServiceState nn2

17/03/12 14:24:50 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable

17/03/12 14:24:51 INFO ipc.Client: Retrying connect to server: db02/192.168.100.232:8020. Already tried 0 time(s); retry policy is RetryUpToMaximumCountWithFixedSleep

(maxRetries=1, sleepTime=1000 MILLISECONDS)

Operation failed: Call From db01/192.168.100.231 to db02:8020 failed on connection exception: java.net.ConnectException: Connection refused; For more details see: 

http://wiki.apache.org/hadoop/ConnectionRefused

自动转移功能配置成功，基于QJM方式搭建的hadoop HA方式大功告成。