时间戳

首页 > 技术文章 > 【Hadoop离线基础总结】Hadoop High Availability\Hadoop基础环境增强

【Hadoop离线基础总结】Hadoop High Availability\Hadoop基础环境增强

zzzsw0412 2020-03-26 22:11 原文

目录

简单介绍

  • Hadoop HA 概述

    HA(High Available) —— 高可用，是保证业务连续性的有效解决方案。一般有两个或两个以上的节点，分为活动节点（Active）及备用节点（Standby）。通常把正在执行业务的称为活动节点，而作为活动节点的一个备份的则称为备用节点。当活动节点出现问题，导致正在运行的业务（任务）不能正常运行时，备用节点此时就会侦测到，并立即接续活动节点来执行业务。从而实现业务的不中断或短暂中断。
    Hadoop1.X版本，NN是HDFS集群的单点故障点，每一个集群只有一个NN,如果这个机器或进程不可用，整个集群就无法使用。为了解决这个问题，出现了一堆针对HDFS HA的解决方案（如：Linux HA, VMware FT, shared NAS+NFS, BookKeeper, QJM/Quorum Journal Manager, BackupNode等）。
    在HA具体实现方法不同情况下，HA框架的流程是一致的, 不一致的就是如何存储、管理、同步edits编辑日志文件。
    在Active NN和Standby NN之间要有个共享的存储日志的地方，Active NN把edit Log写到这个共享的存储日志的地方，Standby NN去读取日志然后执行，这样Active和Standby NN内存中的HDFS元数据保持着同步。一旦发生主从切换Standby NN可以尽快接管Active NN的工作。

  • 集群搭建规划

    在这里插入图片描述

集群搭建

  • 第一步：停止服务

    要停止hadoop集群的所有服务，包括HDFS、yarn、impala、hive、oozie、hue等

    # 停止oozie
cd /export/servers/oozie-4.1.0-cdh5.14.0
bin/oozied.sh stop

    hue	impala	hive在进程中杀死即可

    cd /export/servers/hadoop-2.6.0-cdh5.14.0
sbin/stop-dfs.sh
sbin/stop-yarn.sh
sbin/mr-jobhistory-daemon.sh stop historyserver

  • 第二步：启动所有节点的ZooKeeper

    cd /export/servers/zookeeper-3.4.5-cdh5.14.0
bin/zkServer.sh start

  • 第三步：更改配置文件

    core-site.xml

    <!--
<property>
	<name>fs.default.name</name>
	<value>hdfs://192.168.52.100:8020</value>
</property>
<property>
	<name>hadoop.tmp.dir</name>
	<value>/export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/tempDatas</value>
</property>

<property>
	<name>io.file.buffer.size</name>
	<value>4096</value>
</property>

<property>
	<name>fs.trash.interval</name>
	<value>10080</value>
</property>
-->	

<property>
	<name>ha.zookeeper.quorum</name>
	<value>node01.hadoop.com:2181,node02.hadoop.com:2181,node03.hadoop.com:2181</value>
</property>

<property>
	<name>fs.defaultFS</name>
	<value>hdfs://hann</value>
</property>
<!--  缓冲区大小，实际工作中根据服务器性能动态调整 -->
<property>
	<name>io.file.buffer.size</name>
	<value>4096</value>
</property>

<property>
	<name>hadoop.tmp.dir</name>
	<value>/export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/tempDatas</value>
</property>

<property>
	<name>fs.trash.interval</name>
	<value>10080</value>
</property>
 
<property>
	<name>hadoop.proxyuser.root.hosts</name>
	<value>*</value>
</property>
<property>
	<name>hadoop.proxyuser.root.groups</name>
	<value>*</value>
</property>

    hdfs-site.xml

    <!-- NameNode存储元数据信息的路径，实际工作中，一般先确定磁盘的挂载目录，然后多个目录用，进行分割   --> 
	<!--   集群动态上下线 
	<property>
		<name>dfs.hosts</name>
		<value>/export/servers/hadoop-2.7.4/etc/hadoop/accept_host</value>
	</property>
	
	<property>
		<name>dfs.hosts.exclude</name>
		<value>/export/servers/hadoop-2.7.4/etc/hadoop/deny_host</value>
	</property>
	 -->
	 
	 <!-- 
	 <property>
			<name>dfs.namenode.secondary.http-address</name>
			<value>node01:50090</value>
	</property>

	<property>
		<name>dfs.namenode.http-address</name>
		<value>node01:50070</value>
	</property>
	<property>
		<name>dfs.namenode.name.dir</name>
		<value>file:///export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/namenodeDatas</value>
	</property>

	<property>
		<name>dfs.datanode.data.dir</name>
		<value>file:///export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/datanodeDatas</value>
	</property>
	
	<property>
		<name>dfs.namenode.edits.dir</name>
		<value>file:///export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/dfs/nn/edits</value>
	</property>
	

	<property>
		<name>dfs.namenode.checkpoint.dir</name>
		<value>file:///export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/dfs/snn/name</value>
	</property>
	<property>
		<name>dfs.namenode.checkpoint.edits.dir</name>
		<value>file:///export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/dfs/nn/snn/edits</value>
	</property>

	<property>
		<name>dfs.replication</name>
		<value>1</value>
	</property>

	<property>
		<name>dfs.permissions</name>
		<value>false</value>
	</property>

	<property>
		<name>dfs.blocksize</name>
		<value>134217728</value>
	</property>
	
	<property>
	  <name>dfs.webhdfs.enabled</name>
	  <value>true</value>
	</property>

	<property>
		<name>dfs.client.read.shortcircuit</name>
		<value>true</value>
	</property>
	<property>
		<name>dfs.domain.socket.path</name>
		<value>/var/run/hdfs-sockets/dn</value>
	</property>
	<property>
		<name>dfs.client.file-block-storage-locations.timeout.millis</name>
		<value>10000</value>
	</property>
	<property>
		<name>dfs.datanode.hdfs-blocks-metadata.enabled</name>
		<value>true</value>
	</property>

	<property>
	  <name>dfs.webhdfs.enabled</name>
	  <value>true</value>
	</property>
	
	-->	
	<property>
		<name>dfs.nameservices</name>
		<value>hann</value>
	</property>

		<property>
		<name>dfs.ha.namenodes.hann</name>
		<value>nn1,nn2</value>
	</property>
	<property>
		<name>dfs.namenode.rpc-address.hann.nn1</name>
		<value>node01.hadoop.com:8020</value>
	</property>
	<property>
		<name>dfs.namenode.rpc-address.hann.nn2</name>
		<value>node02.hadoop.com:8020</value>
	</property>
	
	<property>
		<name>dfs.namenode.servicerpc-address.hann.nn1</name>
		<value>node01.hadoop.com:8022</value>
	</property>
	<property>
		<name>dfs.namenode.servicerpc-address.hann.nn2</name>
		<value>node02.hadoop.com:8022</value>
	</property>
	
	
	<property>
		<name>dfs.namenode.http-address.hann.nn1</name>
		<value>node01.hadoop.com:50070</value>
	</property>
	<property>
		<name>dfs.namenode.http-address.hann.nn2</name>
		<value>node02.hadoop.com:50070</value>
	</property>

	
		<property>
		<name>dfs.namenode.shared.edits.dir</name>
		<value>qjournal://node01.hadoop.com:8485;node02.hadoop.com:8485;node03.hadoop.com:8485/hann</value>
	</property>
	
		<property>
		<name>dfs.journalnode.edits.dir</name>
		<value>/export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/dfs/jn</value>
	</property>
	
	<property>
		<name>dfs.client.failover.proxy.provider.hann</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.ha.automatic-failover.enabled</name>
		<value>true</value>
	</property>
	<property>
		<name>dfs.namenode.name.dir</name>
		<value>file:///export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/namenodeDatas</value>
		
	</property>
	<property>
		<name>dfs.namenode.edits.dir</name>
		<value>file:///export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/dfs/nn/edits</value>
	</property>
	
	<property>
		<name>dfs.datanode.data.dir</name>
		<value>file:///export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/datanodeDatas</value>
	</property>



	<property>
		<name>dfs.replication</name>
		<value>3</value>
	</property>

	<property>
		<name>dfs.permissions</name>
		<value>false</value>
	</property>

		<property>
		<name>dfs.blocksize</name>
		<value>134217728</value>
	</property>
	
	<property>
	  <name>dfs.webhdfs.enabled</name>
	  <value>true</value>
	</property>

	
	

	<property>
		<name>dfs.client.read.shortcircuit</name>
		<value>true</value>
	</property>
	<property>
		<name>dfs.domain.socket.path</name>
		<value>/var/run/hdfs-sockets/dn</value>
	</property>
	<property>
		<name>dfs.client.file-block-storage-locations.timeout.millis</name>
		<value>10000</value>
	</property>
	<property>
		<name>dfs.datanode.hdfs-blocks-metadata.enabled</name>
		<value>true</value>
	</property>

	<property>
	  <name>dfs.webhdfs.enabled</name>
	  <value>true</value>
	</property>

    mapred-site.xml

    <!--
	<property>
		<name>mapreduce.framework.name</name>
		<value>yarn</value>
	</property>
	<property>
		<name>mapreduce.job.ubertask.enable</name>
		<value>true</value>
	</property>

	<property>
		<name>mapreduce.jobhistory.address</name>
		<value>node01:10020</value>
	</property>

	<property>
		<name>mapreduce.jobhistory.webapp.address</name>
		<value>node01:19888</value>
	</property>
	-->
<!--
	<property>
		<name>mapreduce.map.output.compress</name>
		<value>true</value>
	</property>
	<property>
		<name>mapreduce.map.output.compress.codec</name>
		<value>org.apache.hadoop.io.compress.SnappyCodec</value>
	</property>


	 <property>
                <name>mapreduce.output.fileoutputformat.compress</name>
                <value>true</value>
        </property>


	 <property>
                <name>mapreduce.output.fileoutputformat.compress.type</name>
                <value>RECORD</value>
        </property>

	 <property>
                <name>mapreduce.output.fileoutputformat.compress.codec</name>
                <value>org.apache.hadoop.io.compress.SnappyCodec</value>
        </property>
-->
<!--指定运行mapreduce的环境是yarn -->
<property>
        <name>mapreduce.framework.name</name>
        <value>yarn</value>
</property>
<!-- MapReduce JobHistory Server IPC host:port -->
<property>
        <name>mapreduce.jobhistory.address</name>
        <value>node03:10020</value>
</property>
<!-- MapReduce JobHistory Server Web UI host:port -->
<property>
        <name>mapreduce.jobhistory.webapp.address</name>
        <value>node03:19888</value>
</property>
<!-- The directory where MapReduce stores control files.默认 ${hadoop.tmp.dir}/mapred/system -->
<property>
        <name>mapreduce.jobtracker.system.dir</name>
        <value>/export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/jobtracker</value>
</property>
<!-- The amount of memory to request from the scheduler for each map task. 默认 1024-->
<property>
        <name>mapreduce.map.memory.mb</name>
        <value>1024</value>
</property>
<!-- <property>
                <name>mapreduce.map.java.opts</name>
                <value>-Xmx1024m</value>
        </property> -->
<!-- The amount of memory to request from the scheduler for each reduce task. 默认 1024-->
<property>
        <name>mapreduce.reduce.memory.mb</name>
        <value>1024</value>
</property>
<!-- <property>
               <name>mapreduce.reduce.java.opts</name>
               <value>-Xmx2048m</value>
        </property> -->
<!-- 用于存储文件的缓存内存的总数量，以兆字节为单位。默认情况下，分配给每个合并流1MB，给个合并流应该寻求最小化。默认值100-->
<property>
        <name>mapreduce.task.io.sort.mb</name>
        <value>100</value>
</property>
 
<!-- <property>
        <name>mapreduce.jobtracker.handler.count</name>
        <value>25</value>
        </property>-->
<!-- 整理文件时用于合并的流的数量。这决定了打开的文件句柄的数量。默认值10-->
<property>
        <name>mapreduce.task.io.sort.factor</name>
        <value>10</value>
</property>
<!-- 默认的并行传输量由reduce在copy(shuffle)阶段。默认值5-->
<property>
        <name>mapreduce.reduce.shuffle.parallelcopies</name>
        <value>25</value>
</property>
<property>
        <name>yarn.app.mapreduce.am.command-opts</name>
        <value>-Xmx1024m</value>
</property>
<!-- MR AppMaster所需的内存总量。默认值1536-->
<property>
        <name>yarn.app.mapreduce.am.resource.mb</name>
        <value>1536</value>
</property>
<!-- MapReduce存储中间数据文件的本地目录。目录不存在则被忽略。默认值${hadoop.tmp.dir}/mapred/local-->
<property>
        <name>mapreduce.cluster.local.dir</name>
        <value>/export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/mapreduce/local</value>
</property>

    yarn-site.xml

    <!--
	<property>
		<name>yarn.resourcemanager.hostname</name>
		<value>node01</value>
	</property>
	<property>
		<name>yarn.nodemanager.aux-services</name>
		<value>mapreduce_shuffle</value>
	</property>
	
	<property>
		<name>yarn.log-aggregation-enable</name>
		<value>true</value>
	</property>
	<property>
		<name>yarn.log-aggregation.retain-seconds</name>
		<value>604800</value>
	</property>
	-->
	
	<!-- Site specific YARN configuration properties -->
<!-- 是否启用日志聚合.应用程序完成后,日志汇总收集每个容器的日志,这些日志移动到文件系统,例如HDFS. -->
<!-- 用户可以通过配置"yarn.nodemanager.remote-app-log-dir"、"yarn.nodemanager.remote-app-log-dir-suffix"来确定日志移动到的位置 -->
<!-- 用户可以通过应用程序时间服务器访问日志 -->

<!-- 启用日志聚合功能，应用程序完成后，收集各个节点的日志到一起便于查看 -->
	<property>
			<name>yarn.log-aggregation-enable</name>
			<value>true</value>
	</property>
 

<!--开启resource manager HA,默认为false--> 
<property>
        <name>yarn.resourcemanager.ha.enabled</name>
        <value>true</value>
</property>
<!-- 集群的Id，使用该值确保RM不会做为其它集群的active -->
<property>
        <name>yarn.resourcemanager.cluster-id</name>
        <value>mycluster</value>
</property>
<!--配置resource manager  命名-->
<property>
        <name>yarn.resourcemanager.ha.rm-ids</name>
        <value>rm1,rm2</value>
</property>
<!-- 配置第一台机器的resourceManager -->
<property>
        <name>yarn.resourcemanager.hostname.rm1</name>
        <value>node03.hadoop.com</value>
</property>
<!-- 配置第二台机器的resourceManager -->
<property>
        <name>yarn.resourcemanager.hostname.rm2</name>
        <value>node02.hadoop.com</value>
</property>

<!-- 配置第一台机器的resourceManager通信地址 -->
<property>
        <name>yarn.resourcemanager.address.rm1</name>
        <value>node03.hadoop.com:8032</value>
</property>
<property>
        <name>yarn.resourcemanager.scheduler.address.rm1</name>
        <value>node03.hadoop.com:8030</value>
</property>
<property>
        <name>yarn.resourcemanager.resource-tracker.address.rm1</name>
        <value>node03.hadoop.com:8031</value>
</property>
<property>
        <name>yarn.resourcemanager.admin.address.rm1</name>
        <value>node03.hadoop.com:8033</value>
</property>
<property>
        <name>yarn.resourcemanager.webapp.address.rm1</name>
        <value>node03.hadoop.com:8088</value>
</property>

<!-- 配置第二台机器的resourceManager通信地址 -->
<property>
        <name>yarn.resourcemanager.address.rm2</name>
        <value>node02.hadoop.com:8032</value>
</property>
<property>
        <name>yarn.resourcemanager.scheduler.address.rm2</name>
        <value>node02.hadoop.com:8030</value>
</property>
<property>
        <name>yarn.resourcemanager.resource-tracker.address.rm2</name>
        <value>node02.hadoop.com:8031</value>
</property>
<property>
        <name>yarn.resourcemanager.admin.address.rm2</name>
        <value>node02.hadoop.com:8033</value>
</property>
<property>
        <name>yarn.resourcemanager.webapp.address.rm2</name>
        <value>node02.hadoop.com:8088</value>
</property>
<!--开启resourcemanager自动恢复功能-->
<property>
        <name>yarn.resourcemanager.recovery.enabled</name>
        <value>true</value>
</property>
<!--在node3上配置rm1,在node2上配置rm2,注意：一般都喜欢把配置好的文件远程复制到其它机器上，但这个在YARN的另一个机器上一定要修改，其他机器上不配置此项-->
	<property>       
		<name>yarn.resourcemanager.ha.id</name>
		<value>rm1</value>
       <description>If we want to launch more than one RM in single node, we need this configuration</description>
	</property>
	   
	   <!--用于持久存储的类。尝试开启-->
<property>
        <name>yarn.resourcemanager.store.class</name>
        <value>org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore</value>
</property>
<property>
        <name>yarn.resourcemanager.zk-address</name>
        <value>node01.hadoop.com:2181,node02.hadoop.com:2181,node03.hadoop.com:2181</value>
        <description>For multiple zk services, separate them with comma</description>
</property>
<!--开启resourcemanager故障自动切换，指定机器--> 
<property>
        <name>yarn.resourcemanager.ha.automatic-failover.enabled</name>
        <value>true</value>
        <description>Enable automatic failover; By default, it is enabled only when HA is enabled.</description>
</property>
<property>
        <name>yarn.client.failover-proxy-provider</name>
        <value>org.apache.hadoop.yarn.client.ConfiguredRMFailoverProxyProvider</value>
</property>
<!-- 允许分配给一个任务最大的CPU核数，默认是8 -->
<property>
        <name>yarn.nodemanager.resource.cpu-vcores</name>
        <value>4</value>
</property>
<!-- 每个节点可用内存,单位MB -->
<property>
        <name>yarn.nodemanager.resource.memory-mb</name>
        <value>512</value>
</property>
<!-- 单个任务可申请最少内存，默认1024MB -->
<property>
        <name>yarn.scheduler.minimum-allocation-mb</name>
        <value>512</value>
</property>
<!-- 单个任务可申请最大内存，默认8192MB -->
<property>
        <name>yarn.scheduler.maximum-allocation-mb</name>
        <value>512</value>
</property>
<!--多长时间聚合删除一次日志 此处-->
<property>
        <name>yarn.log-aggregation.retain-seconds</name>
        <value>2592000</value><!--30 day-->
</property>
<!--时间在几秒钟内保留用户日志。只适用于如果日志聚合是禁用的-->
<property>
        <name>yarn.nodemanager.log.retain-seconds</name>
        <value>604800</value><!--7 day-->
</property>
<!--指定文件压缩类型用于压缩汇总日志-->
<property>
        <name>yarn.nodemanager.log-aggregation.compression-type</name>
        <value>gz</value>
</property>
<!-- nodemanager本地文件存储目录-->
<property>
        <name>yarn.nodemanager.local-dirs</name>
        <value>/export/servers/hadoop-2.6.0-cdh5.14.0/hadoopDatas/yarn/local</value>
</property>
<!-- resourceManager  保存最大的任务完成个数 -->
<property>
        <name>yarn.resourcemanager.max-completed-applications</name>
        <value>1000</value>
</property>
<!-- 逗号隔开的服务列表，列表名称应该只包含a-zA-Z0-9_,不能以数字开始-->
<property>
        <name>yarn.nodemanager.aux-services</name>
        <value>mapreduce_shuffle</value>
</property>

<!--rm失联后重新链接的时间--> 
<property>
        <name>yarn.resourcemanager.connect.retry-interval.ms</name>
        <value>2000</value>
</property>

    把这四个发送到node02,node03，node02的yarn-site.xml要把yarn.resourcemanager.ha.id的值修改为rm2

    scp core-site.xml hdfs-site.xml mapred-site.xml yarn-site.xml node02:$PWD
scp core-site.xml hdfs-site.xml mapred-site.xml yarn-site.xml node03:$PWD

  • 第四步：启动服务

    1.在node01初始化zookeeper

    cd /export/servers/hadoop-2.6.0-cdh5.14.0
bin/hdfs zkfc -formatZK

    2.启动journalNode，三台机器都要执行

    cd /export/servers/hadoop-2.6.0-cdh5.14.0
sbin/hadoop-daemon.sh start journalnode

    3.初始化journalNode在node01执行即可

    cd /export/servers/hadoop-2.6.0-cdh5.14.0
bin/hdfs namenode -initializeSharedEdits -force

    4.在node01启动NameNode

    cd /export/servers/hadoop-2.6.0-cdh5.14.0
sbin/hadoop-daemon.sh start namenode

    5.在node02启动Namenode

    cd /export/servers/hadoop-2.6.0-cdh5.14.0
bin/hdfs namenode -bootstrapStandby
sbin/hadoop-daemon.sh start namenode

    6.在node01启动所有节点的datanode

    cd /export/servers/hadoop-2.6.0-cdh5.14.0
sbin/hadoop-daemons.sh start datanode

    7.在node01和node02启动zkfc

    cd /export/servers/hadoop-2.6.0-cdh5.14.0
sbin/hadoop-daemon.sh start zkfc

    8.在node02和node03启动yarn

    cd /export/servers/hadoop-2.6.0-cdh5.14.0
sbin/start-yarn.sh

    8.在node03启动jobhistoryserver

    cd /export/servers/hadoop-2.6.0-cdh5.14.0
sbin/mr-jobhistory-daemon.sh start historyserver

推荐阅读