版权声明:原创作品,谢绝转载!否则将追究法律责任。
本篇博客的高可用集群是建立在完全分布式基础之上的,详情请参考:https://www.cnblogs.com/yinzhengjie/p/9065191.html。并且需要新增一台Linux服务器,用于Namenode的备份节点。
一.实验环境准备
需要准备五台Linux操作系统的服务器,配置参数最好一样,由于我的虚拟机是之前完全分布式部署而来的,因此我的环境都一致。
1>.NameNode服务器(s101)
2>.DataNode服务器(s102)
3>.DataNode服务器(s103)
4>.DataNode服务器(s104)
5>.DataNode服务器(s105)
二.在s101上修改配置文件并分发到其它节点
关于配置hadoop高可用的参数,参考官网链接:http://hadoop.apache.org/docs/r2.7.3/hadoop-project-dist/hadoop-hdfs/HDFSHighAvailabilityWithQJM.html
1>.在s101上拷贝配置目录并修改符号链接
[yinzhengjie@s101 ~]$ ll /soft/hadoop/etc/ total 12 drwxr-xr-x. 2 yinzhengjie yinzhengjie 4096 Jun 8 05:36 full lrwxrwxrwx. 1 yinzhengjie yinzhengjie 21 Jun 8 05:54 hadoop -> /soft/hadoop/etc/full drwxr-xr-x. 2 yinzhengjie yinzhengjie 4096 May 25 09:15 local drwxr-xr-x. 2 yinzhengjie yinzhengjie 4096 May 25 20:51 pseudo [yinzhengjie@s101 ~]$ cp -r /soft/hadoop/etc/full /soft/hadoop/etc/ha [yinzhengjie@s101 ~]$ ln -sfT /soft/hadoop/etc/ha /soft/hadoop/etc/hadoop [yinzhengjie@s101 ~]$ ll /soft/hadoop/etc/ total 16 drwxr-xr-x. 2 yinzhengjie yinzhengjie 4096 Jun 8 05:36 full drwxr-xr-x. 2 yinzhengjie yinzhengjie 4096 Jun 8 05:54 ha lrwxrwxrwx. 1 yinzhengjie yinzhengjie 19 Jun 8 05:54 hadoop -> /soft/hadoop/etc/ha drwxr-xr-x. 2 yinzhengjie yinzhengjie 4096 May 25 09:15 local drwxr-xr-x. 2 yinzhengjie yinzhengjie 4096 May 25 20:51 pseudo [yinzhengjie@s101 ~]$
2>.配置s105ssh免密登陆
[yinzhengjie@s101 ~]$ ssh-copy-id yinzhengjie@s105 /usr/bin/ssh-copy-id: INFO: attempting to log in with the new key(s), to filter out any that are already installed /usr/bin/ssh-copy-id: INFO: 1 key(s) remain to be installed -- if you are prompted now it is to install the new keys yinzhengjie@s105's password: Number of key(s) added: 1 Now try logging into the machine, with: "ssh 'yinzhengjie@s105'" and check to make sure that only the key(s) you wanted were added. [yinzhengjie@s101 ~]$ who yinzhengjie pts/0 2018-06-08 05:29 (172.16.30.1) [yinzhengjie@s101 ~]$ [yinzhengjie@s101 ~]$ ssh s105 Last login: Fri Jun 8 05:37:20 2018 from 172.16.30.1 [yinzhengjie@s105 ~]$ [yinzhengjie@s105 ~]$ who yinzhengjie pts/0 2018-06-08 05:37 (172.16.30.1) yinzhengjie pts/1 2018-06-08 05:56 (s101) [yinzhengjie@s105 ~]$ exit logout Connection to s105 closed. [yinzhengjie@s101 ~]$
3>.编辑core-site.xml配置文件
[yinzhengjie@s101 ~]$ more /soft/hadoop/etc/hadoop/core-site.xml <?xml version="1.0" encoding="UTF-8"?> <configuration> <property> <name>fs.defaultFS</name> <value>hdfs://mycluster</value> </property> <property> <name>hadoop.tmp.dir</name> <value>/home/yinzhengjie/ha</value> </property> <property> <name>hadoop.http.staticuser.user</name> <value>yinzhengjie</value> </property> </configuration> <!-- core-site.xml配置文件的作用: 用于定义系统级别的参数,如HDFS URL、Hadoop的临时 目录以及用于rack-aware集群中的配置文件的配置等,此中的参 数定义会覆盖core-default.xml文件中的默认配置。 fs.defaultFS 参数的作用: #fs.defaultFS 客户端连接HDFS时,默认的路径前缀。如果前面配置了nameservice ID的值是mycluster,那么这里可以配置为授权 信息的一部分 hadoop.tmp.dir 参数的作用: #声明hadoop工作目录的地址。 hadoop.http.staticuser.user 参数的作用: #在网页界面访问数据使用的用户名。 --> [yinzhengjie@s101 ~]$
4>.编辑hdfs-site.xml配置文件
[yinzhengjie@s101 ~]$ more /soft/hadoop/etc/hadoop/hdfs-site.xml <?xml version="1.0" encoding="UTF-8"?> <?xml-stylesheet type="text/xsl" href="configuration.xsl"?> <configuration> <property> <name>dfs.replication</name> <value>3</value> </property> <property> <name>dfs.namenode.name.dir</name> <value>/home/yinzhengjie/ha/dfs/name1,/home/yinzhengjie/ha/dfs/name2</value> </property> <property> <name>dfs.datanode.data.dir</name> <value>/home/yinzhengjie/ha/dfs/data1,/home/yinzhengjie/ha/dfs/data2</value> </property> <!-- 高可用配置 --> <property> <name>dfs.nameservices</name> <value>mycluster</value> </property> <property> <name>dfs.ha.namenodes.mycluster</name> <value>nn1,nn2</value> </property> <property> <name>dfs.namenode.rpc-address.mycluster.nn1</name> <value>s101:8020</value> </property> <property> <name>dfs.namenode.rpc-address.mycluster.nn2</name> <value>s105:8020</value> </property> <property> <name>dfs.namenode.http-address.mycluster.nn1</name> <value>s101:50070</value> </property> <property> <name>dfs.namenode.http-address.mycluster.nn2</name> <value>s105:50070</value> </property> <property> <name>dfs.namenode.shared.edits.dir</name> <value>qjournal://s102:8485;s103:8485;s104:8485/mycluster</value> </property> <property> <name>dfs.client.failover.proxy.provider.mycluster</name> <value>org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider</value> </property> <!-- 在容灾发生时,保护活跃的namenode --> <property> <name>dfs.ha.fencing.methods</name> <value> sshfence shell(/bin/true) </value> </property> <property> <name>dfs.ha.fencing.ssh.private-key-files</name> <value>/home/yinzhengjie/.ssh/id_rsa</value> </property> </configuration> <!-- hdfs-site.xml 配置文件的作用: #HDFS的相关设定,如文件副本的个数、块大小及是否使用强制权限 等,此中的参数定义会覆盖hdfs-default.xml文件中的默认配置. dfs.replication 参数的作用: #为了数据可用性及冗余的目的,HDFS会在多个节点上保存同一个数据 块的多个副本,其默认为3个。而只有一个节点的伪分布式环境中其仅用 保存一个副本即可,这可以通过dfs.replication属性进行定义。它是一个 软件级备份。 dfs.namenode.name.dir 参数的作用: #本地磁盘目录,NN存储fsimage文件的地方。可以是按逗号分隔的目录列表, fsimage文件会存储在全部目录,冗余安全。这里多个目录设定,最好在多个磁盘, 另外,如果其中一个磁盘故障,不会导致系统故障,会跳过坏磁盘。由于使用了HA, 建议仅设置一个。如果特别在意安全,可以设置2个 dfs.datanode.data.dir 参数的作用: #本地磁盘目录,HDFS数据应该存储Block的地方。可以是逗号分隔的目录列表 (典型的,每个目录在不同的磁盘)。这些目录被轮流使用,一个块存储在这个目录, 下一个块存储在下一个目录,依次循环。每个块在同一个机器上仅存储一份。不存在 的目录被忽略。必须创建文件夹,否则被视为不存在。 dfs.nameservices 参数的作用: #nameservices列表。逗号分隔。 dfs.ha.namenodes.mycluster 参数的作用: #dfs.ha.namenodes.[nameservice ID] 命名空间中所有NameNode的唯一标示名称。 可以配置多个,使用逗号分隔。该名称是可以让DataNode知道每个集群的所有NameNode。 当前,每个集群最多只能配置两个NameNode。 dfs.namenode.rpc-address.mycluster.nn1 参数的作用: #dfs.namenode.rpc-address.[nameservice ID].[name node ID] 每个namenode监听的RPC地址。 dfs.namenode.http-address.mycluster.nn1 参数的作用: #dfs.namenode.http-address.[nameservice ID].[name node ID] 每个namenode监听的http地址。 dfs.namenode.shared.edits.dir 参数的作用: #这是NameNode读写JNs组的uri。通过这个uri,NameNodes可以读写edit log内容。URI的格式"qjournal://host1:port1;host2:port 2;host3:port3/journalId"。这里的host1、host2、host3指的是Journal Node的地址,这里必须是奇数个,至少3个;其中journalId是集群 的唯一标识符,对于多个联邦命名空间,也使用同一个journalId。 dfs.client.failover.proxy.provider.mycluster 参数的作用: #这里配置HDFS客户端连接到Active NameNode的一个java类 dfs.ha.fencing.methods 参数的作用: #dfs.ha.fencing.methods 配置active namenode出错时的处理类。当active namenode出错时,一般需要关闭该进程。处理方式可以 是ssh也可以是shell。 dfs.ha.fencing.ssh.private-key-files 参数的作用: #使用sshfence时,SSH的私钥文件。 使用了sshfence,这个必须指定 --> [yinzhengjie@s101 ~]$
5>.分发配置文件
[yinzhengjie@s101 ~]$ more `which xrsync.sh` #!/bin/bash #@author :yinzhengjie #blog:http://www.cnblogs.com/yinzhengjie #EMAIL:y1053419035@qq.com #判断用户是否传参 if [ $# -lt 1 ];then echo "请输入参数"; exit fi #获取文件路径 file=$@ #获取子路径 filename=`basename $file` #获取父路径 dirpath=`dirname $file` #获取完整路径 cd $dirpath fullpath=`pwd -P` #同步文件到DataNode for (( i=102;i<=105;i++ )) do #使终端变绿色 tput setaf 2 echo =========== s$i %file =========== #使终端变回原来的颜色,即白灰色 tput setaf 7 #远程执行命令 rsync -lr $filename `whoami`@s$i:$fullpath #判断命令是否执行成功 if [ $? == 0 ];then echo "命令执行成功" fi done [yinzhengjie@s101 ~]$ xrsync.sh /soft/hadoop/etc/ =========== s102 %file =========== 命令执行成功 =========== s103 %file =========== 命令执行成功 =========== s104 %file =========== 命令执行成功 =========== s105 %file =========== 命令执行成功 [yinzhengjie@s101 ~]$
三.启动HDFS分布式系统
1>.启动journalnode进程
[yinzhengjie@s101 ~]$ hadoop-daemons.sh start journalnode s104: starting journalnode, logging to /soft/hadoop-2.7.3/logs/hadoop-yinzhengjie-journalnode-s104.out s103: starting journalnode, logging to /soft/hadoop-2.7.3/logs/hadoop-yinzhengjie-journalnode-s103.out s102: starting journalnode, logging to /soft/hadoop-2.7.3/logs/hadoop-yinzhengjie-journalnode-s102.out [yinzhengjie@s101 ~]$ xcall.sh jps ============= s101 jps ============ 2855 Jps 命令执行成功 ============= s102 jps ============ 2568 Jps 2490 JournalNode 命令执行成功 ============= s103 jps ============ 2617 Jps 2539 JournalNode 命令执行成功 ============= s104 jps ============ 2611 Jps 2532 JournalNode 命令执行成功 ============= s105 jps ============ 2798 Jps 命令执行成功 [yinzhengjie@s101 ~]$ more `which xcall.sh` #!/bin/bash #@author :yinzhengjie #blog:http://www.cnblogs.com/yinzhengjie #EMAIL:y1053419035@qq.com #判断用户是否传参 if [ $# -lt 1 ];then echo "请输入参数" exit fi #获取用户输入的命令 cmd=$@ for (( i=101;i<=105;i++ )) do #使终端变绿色 tput setaf 2 echo ============= s$i $cmd ============ #使终端变回原来的颜色,即白灰色 tput setaf 7 #远程执行命令 ssh s$i $cmd #判断命令是否执行成功 if [ $? == 0 ];then echo "命令执行成功" fi done [yinzhengjie@s101 ~]$
2>.格式化名称节点
3>.将s101中的工作目录同步到s105
[yinzhengjie@s101 ~]$ scp -r /home/yinzhengjie/ha yinzhengjie@s105:~ VERSION 100% 205 0.2KB/s 00:00 seen_txid 100% 2 0.0KB/s 00:00 fsimage_0000000000000000000.md5 100% 62 0.1KB/s 00:00 fsimage_0000000000000000000 100% 358 0.4KB/s 00:00 VERSION 100% 205 0.2KB/s 00:00 seen_txid 100% 2 0.0KB/s 00:00 fsimage_0000000000000000000.md5 100% 62 0.1KB/s 00:00 fsimage_0000000000000000000 100% 358 0.4KB/s 00:00 [yinzhengjie@s101 ~]$
4>.启动hdfs进程
[yinzhengjie@s101 ~]$ start-dfs.sh Starting namenodes on [s101 s105] s105: starting namenode, logging to /soft/hadoop-2.7.3/logs/hadoop-yinzhengjie-namenode-s105.out s101: starting namenode, logging to /soft/hadoop-2.7.3/logs/hadoop-yinzhengjie-namenode-s101.out s104: starting datanode, logging to /soft/hadoop-2.7.3/logs/hadoop-yinzhengjie-datanode-s104.out s103: starting datanode, logging to /soft/hadoop-2.7.3/logs/hadoop-yinzhengjie-datanode-s103.out s102: starting datanode, logging to /soft/hadoop-2.7.3/logs/hadoop-yinzhengjie-datanode-s102.out Starting journal nodes [s102 s103 s104] s102: journalnode running as process 2490. Stop it first. s104: journalnode running as process 2532. Stop it first. s103: journalnode running as process 2539. Stop it first. [yinzhengjie@s101 ~]$ xcall.sh jps ============= s101 jps ============ 3377 Jps 3117 NameNode 命令执行成功 ============= s102 jps ============ 2649 DataNode 2490 JournalNode 2764 Jps 命令执行成功 ============= s103 jps ============ 2539 JournalNode 2700 DataNode 2815 Jps 命令执行成功 ============= s104 jps ============ 2532 JournalNode 2693 DataNode 2809 Jps 命令执行成功 ============= s105 jps ============ 3171 NameNode 3254 Jps 命令执行成功 [yinzhengjie@s101 ~]$
5>.手动将s101转换成激活状态
[yinzhengjie@s101 ~]$ hdfs haadmin -transitionToActive nn1 //手动将s101转换成激活状态 [yinzhengjie@s101 ~]$
配置到这里基本上高可用就配置好了,但是美中不足的是需要字节手动切换NameNode模式,这就比较麻烦了。索性的是:Hadoop生态圈有专门维护的工具叫做zookeeper工具,我们可以用该工具对集群继续管理就相当方便啦!详情请参考:https://www.cnblogs.com/yinzhengjie/p/9154265.html