Apache Hadoop 2.9.2 的集群管理之服役和退役
作者：尹正杰
版权声明：原创作品，谢绝转载！否则将追究法律责任。
随着公司业务的发展，客户量越来越多，产生的日志自然也就越来越大来，可能我们现有集群的DataNode节点的容量依旧不能满足存储数据的需求，因此需要在现有的集群基础之上动态添加DataNode在生成环境中也是很有可能的。
一.添加新节点的过程（服役）
1>.查看需要加入进群的节点的IP地址
[root@node110.yinzhengjie.org.cn ~]# ifconfig
eth0: flags=<UP,BROADCAST,RUNNING,MULTICAST> mtu
inet 172.30.1.110 netmask 255.255.255.0 broadcast 172.30.1.255
inet6 fe80::21c:42ff:fe11: prefixlen scopeid 0x20<link>
ether :1c:::: txqueuelen (Ethernet)
RX packets bytes (7.6 KiB)
RX errors dropped overruns frame
TX packets bytes (7.3 KiB)
TX errors dropped overruns carrier collisions lo: flags=<UP,LOOPBACK,RUNNING> mtu
inet 127.0.0.1 netmask 255.0.0.0
inet6 :: prefixlen scopeid 0x10<host>
loop txqueuelen (Local Loopback)
RX packets bytes (0.0 B)
RX errors dropped overruns frame
TX packets bytes (0.0 B)
TX errors dropped overruns carrier collisions [root@node110.yinzhengjie.org.cn ~]#
[root@node110.yinzhengjie.org.cn ~]# ifconfig
[root@node110.yinzhengjie.org.cn ~]# cat /etc/hosts | grep yinzhengjie
172.30.1.101 node101.yinzhengjie.org.cn
172.30.1.102 node102.yinzhengjie.org.cn
172.30.1.103 node103.yinzhengjie.org.cn
172.30.1.110 node110.yinzhengjie.org.cn
[root@node110.yinzhengjie.org.cn ~]#
[root@node110.yinzhengjie.org.cn ~]# cat /etc/hosts | grep yinzhengjie
[root@node110.yinzhengjie.org.cn ~]# yum -y install java-1.8.-openjdk-devel
Loaded plugins: fastestmirror, langpacks
Loading mirror speeds from cached hostfile
* base: mirrors.aliyun.com
* extras: mirrors.aliyun.com
* updates: mirrors.aliyun.com
Resolving Dependencies
--> Running transaction check
---> Package java-1.8.-openjdk-devel.x86_64 :1.8.0.201.b09-.el7_6 will be installed
--> Processing Dependency: java-1.8.-openjdk(x86-) = :1.8.0.201.b09-.el7_6 for package: :java-1.8.-openjdk-devel-1.8.0.201.b09-.el7_6.x86_64
--> Running transaction check
---> Package java-1.8.-openjdk.x86_64 :1.8.0.201.b09-.el7_6 will be installed
--> Processing Dependency: java-1.8.-openjdk-headless(x86-) = :1.8.0.201.b09-.el7_6 for package: :java-1.8.-openjdk-1.8.0.201.b09-.el7_6.x86_64
--> Running transaction check
---> Package java-1.8.-openjdk-headless.x86_64 :1.8.0.201.b09-.el7_6 will be installed
--> Finished Dependency Resolution Dependencies Resolved ==========================================================================================================================================================================================================================================
Package Arch Version Repository Size
==========================================================================================================================================================================================================================================
Installing:
java-1.8.-openjdk-devel x86_64 :1.8.0.201.b09-.el7_6 updates 9.8 M
Installing for dependencies:
java-1.8.-openjdk x86_64 :1.8.0.201.b09-.el7_6 updates k
java-1.8.-openjdk-headless x86_64 :1.8.0.201.b09-.el7_6 updates M Transaction Summary
==========================================================================================================================================================================================================================================
Install Package (+ Dependent packages) Total download size: M
Installed size: M
Downloading packages:
(/): java-1.8.-openjdk-1.8.0.201.b09-.el7_6.x86_64.rpm | kB ::
(/): java-1.8.-openjdk-devel-1.8.0.201.b09-.el7_6.x86_64.rpm | 9.8 MB ::
(/): java-1.8.-openjdk-headless-1.8.0.201.b09-.el7_6.x86_64.rpm | MB ::
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Total 2.3 MB/s | MB ::
Running transaction check
Running transaction test
Transaction test succeeded
Running transaction
Installing : :java-1.8.-openjdk-headless-1.8.0.201.b09-.el7_6.x86_64 /
Installing : :java-1.8.-openjdk-1.8.0.201.b09-.el7_6.x86_64 /
Installing : :java-1.8.-openjdk-devel-1.8.0.201.b09-.el7_6.x86_64 /
Verifying : :java-1.8.-openjdk-1.8.0.201.b09-.el7_6.x86_64 /
Verifying : :java-1.8.-openjdk-headless-1.8.0.201.b09-.el7_6.x86_64 /
Verifying : :java-1.8.-openjdk-devel-1.8.0.201.b09-.el7_6.x86_64 / Installed:
java-1.8.-openjdk-devel.x86_64 :1.8.0.201.b09-.el7_6 Dependency Installed:
java-1.8.-openjdk.x86_64 :1.8.0.201.b09-.el7_6 java-1.8.-openjdk-headless.x86_64 :1.8.0.201.b09-.el7_6 Complete!
[root@node110.yinzhengjie.org.cn ~]#
[root@node110.yinzhengjie.org.cn ~]# yum -y install java-1.8.0-openjdk-devel
2>.将NameNode节点配置环境拷贝到需要加入进群的节点的服务器
[root@node101.yinzhengjie.org.cn ~]# cat /yinzhengjie/softwares/hadoop-2.9./etc/hadoop/hdfs-site.xml
<?xml version="1.0" encoding="UTF-8"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?> <configuration>
<property>
<name>dfs.namenode.checkpoint.period</name>
<value></value>
</property> <property>
<name>dfs.namenode.name.dir</name>
<value>file:///${hadoop.tmp.dir}/dfs/namenode1,file:///${hadoop.tmp.dir}/dfs/namenode2,file:///${hadoop.tmp.dir}/dfs/namenode3</value>
</property> <property>
<name>dfs.datanode.data.dir</name>
<value>file:///${hadoop.tmp.dir}/dfs/data1,file:///${hadoop.tmp.dir}/dfs/data2</value>
</property> <property>
<name>dfs.replication</name>
<value></value>
</property> <property>
<name>dfs.namenode.heartbeat.recheck-interval</name>
<value></value>
</property> <property>
<name> dfs.heartbeat.interval </name>
<value></value>
</property> <!--
<property>
<name>dfs.hosts</name>
<value>file:///${hadoop.tmp.dir}/DataNodesHostname.txt</value>
</property>
-->
<property>
<name>dfs.hosts.exclude</name>
<value>file:///${hadoop.tmp.dir}/dfs.hosts.exclude.txt</value>
</property>
</configuration> <!--
hdfs-site.xml 配置文件的作用：
#HDFS的相关设定，如文件副本的个数、块大小及是否使用强制权限等，此中的参数定义会覆盖hdfs-default.xml文件中的默认配置. dfs.namenode.checkpoint.period 参数的作用：
#两个定期检查点之间的秒数,默认是3600，即1小时。 dfs.namenode.name.dir 参数的作用：
#指定namenode的工作目录，默认是file://${hadoop.tmp.dir}/dfs/name,namenode的本地目录可以配置成多个，且每个目录存放内容相同，增加了可靠性。建议配置的多目录用不同磁盘挂在，这样可以提升IO性能！ dfs.datanode.data.dir 参数的作用：
#指定datanode的工作目录，议配置的多目录用不同磁盘挂在，这样可以提升IO性能！但是多个目录存储的数据并不相同哟！而是把数据存放在不同的目录，当namenode存储数据时效率更高！ 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.heartbeat.interval 参数的作用：
#设置心跳检测时间 dfs.namenode.heartbeat.recheck-interval和dfs.heartbeat.interval 参数的作用：
#设置HDFS中NameNode和DataNode的超时时间，计算公式为:timeout = * dfs.namenode.heartbeat.recheck-interval + * dfs.heartbeat.interval。 dfs.hosts 参数的作用：
#添加白名单，功能和黑名单（dfs.hosts.exclude）相反。我这里是将其注释掉了。 dfs.hosts.exclude 参数的作用:
#这是我们添加的黑名单，该属性的value定义的是一个配置文件，这个配置文件的主机就是需要退役的节点。 --> [root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# cat /yinzhengjie/softwares/hadoop-2.9.2/etc/hadoop/hdfs-site.xml
[root@node101.yinzhengjie.org.cn ~]# ssh-copy-id node110.yinzhengjie.org.cn
/usr/bin/ssh-copy-id: INFO: Source of key(s) to be installed: "/root/.ssh/id_rsa.pub"
The authenticity of host 'node110.yinzhengjie.org.cn (172.30.1.110)' can't be established.
ECDSA key fingerprint is SHA256:7fUJXbFRvzvhON0FKnxvIivwkEZsiQ1jXE3V3i7ONEg.
ECDSA key fingerprint is MD5:7b::fb:ed:::ed:7b::db:::::0c:5b.
Are you sure you want to continue connecting (yes/no)? yes
/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: key(s) remain to be installed -- if you are prompted now it is to install the new keys
root@node110.yinzhengjie.org.cn's password: Number of key(s) added: Now try logging into the machine, with: "ssh 'node110.yinzhengjie.org.cn'"
and check to make sure that only the key(s) you wanted were added. [root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# ssh-copy-id node110.yinzhengjie.org.cn
[root@node101.yinzhengjie.org.cn ~]# scp -r /yinzhengjie/softwares/hadoop-2.9./ node110.yinzhengjie.org.cn:/yinzhengjie/softwares/ .......
README.txt % .8MB/s :
mapred-config.cmd % .8MB/s :
hdfs-config.sh % .9MB/s :
httpfs-config.sh % .4MB/s :
hdfs-config.cmd % .3MB/s :
yarn-config.cmd % .0MB/s :
mapred-config.sh % .6MB/s :
hadoop-config.cmd % .3MB/s :
hadoop-config.sh % 11KB .3MB/s :
yarn-config.sh % .4MB/s :
kms-config.sh % .3MB/s :
libhadooppipes.a % 1594KB .3MB/s :
libhadooputils.a % 465KB .0MB/s :
libhdfs.so.0.0. % 278KB .4MB/s :
wordcount-part % 933KB .5MB/s :
wordcount-simple % 924KB .3MB/s :
pipes-sort % 910KB .9MB/s :
wordcount-nopipe % 972KB .8MB/s :
libhdfs.so % 278KB .1MB/s :
libhdfs.a % 444KB .7MB/s :
libhadoop.so % 822KB .9MB/s :
libhadoop.a % 1406KB .3MB/s :
libhadoop.so.1.0. % 822KB .6MB/s :
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# scp -r /yinzhengjie/softwares/hadoop-2.9.2/ node110.yinzhengjie.org.cn:/yinzhengjie/softwares/
[root@node101.yinzhengjie.org.cn ~]# scp /etc/profile node110.yinzhengjie.org.cn:/etc/
profile % .8MB/s :
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# scp /etc/profile node110.yinzhengjie.org.cn:/etc/ profile
3>.在新加入的datanode节点上单独启动datanode进程
[root@node110.yinzhengjie.org.cn ~]# hadoop-daemon.sh start datanode
starting datanode, logging to /yinzhengjie/softwares/hadoop-2.9./logs/hadoop-root-datanode-node110.yinzhengjie.org.cn.out
[root@node110.yinzhengjie.org.cn ~]#
[root@node110.yinzhengjie.org.cn ~]# jps
DataNode
Jps
[root@node110.yinzhengjie.org.cn ~]#
[root@node110.yinzhengjie.org.cn ~]# hadoop-daemon.sh start datanode
4>.编辑黑名单，将node102.yinzhengjie.org.cn节点加入我们定义好的配置文件
[root@node101.yinzhengjie.org.cn ~]# cat /data/hadoop/hdfs/dfs.hosts.exclude.txt
node102.yinzhengjie.org.cn
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# cat /data/hadoop/hdfs/dfs.hosts.exclude.txt
[root@node101.yinzhengjie.org.cn ~]# hdfs dfsadmin -refreshNodes
Refresh nodes successful
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# hdfs dfsadmin -refreshNodes #如果有同时不存在文件的话可以重启一下集群！
5>.关于服役相关说明
根据上述的步骤，我们可以成功将一个节点加入到集另一个群中。也可以禁用某个节点，只需要将该主机加入到我们自定义的黑名单配置文件中即可。其实我们将节点加入集群成功后，我们最好将slaves文件也修改为最新的状态，从而达到便于管理集群的DataNode的目的，不用每次启动集群时都新加入的集群需要单独启动！
[root@node101.yinzhengjie.org.cn ~]# cat /yinzhengjie/softwares/hadoop-2.9./etc/hadoop/slaves
node101.yinzhengjie.org.cn
node102.yinzhengjie.org.cn
node103.yinzhengjie.org.cn
node110.yinzhengjie.org.cn
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# cat /yinzhengjie/softwares/hadoop-2.9.2/etc/hadoop/slaves
二.删除旧节点的过程（退役）
1>.修改NameNode的配置文件并同步到其他节点上去
[root@node101.yinzhengjie.org.cn ~]# cat /yinzhengjie/softwares/hadoop-2.9./etc/hadoop/hdfs-site.xml
<?xml version="1.0" encoding="UTF-8"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?> <configuration>
<property>
<name>dfs.namenode.checkpoint.period</name>
<value></value>
</property> <property>
<name>dfs.namenode.name.dir</name>
<value>file:///${hadoop.tmp.dir}/dfs/namenode1,file:///${hadoop.tmp.dir}/dfs/namenode2,file:///${hadoop.tmp.dir}/dfs/namenode3</value>
</property> <property>
<name>dfs.datanode.data.dir</name>
<value>file:///${hadoop.tmp.dir}/dfs/data1,file:///${hadoop.tmp.dir}/dfs/data2</value>
</property> <property>
<name>dfs.replication</name>
<value></value>
</property> <property>
<name>dfs.namenode.heartbeat.recheck-interval</name>
<value></value>
</property> <property>
<name> dfs.heartbeat.interval </name>
<value></value>
</property> <property>
<name>dfs.hosts</name>
<value>/data/hadoop/hdfs/DataNodesHostname.txt</value>
</property> <!--
<property>
<name>dfs.hosts.exclude</name>
<value>/data/hadoop/hdfs/dfs.hosts.exclude.txt</value>
</property>
--> </configuration> <!--
hdfs-site.xml 配置文件的作用：
#HDFS的相关设定，如文件副本的个数、块大小及是否使用强制权限等，此中的参数定义会覆盖hdfs-default.xml文件中的默认配置. dfs.namenode.checkpoint.period 参数的作用：
#两个定期检查点之间的秒数,默认是3600，即1小时。 dfs.namenode.name.dir 参数的作用：
#指定namenode的工作目录，默认是file://${hadoop.tmp.dir}/dfs/name,namenode的本地目录可以配置成多个，且每个目录存放内容相同，增加了可靠性。建议配置的多目录用不同磁盘挂在，这样可以提升IO性能！ dfs.datanode.data.dir 参数的作用：
#指定datanode的工作目录，议配置的多目录用不同磁盘挂在，这样可以提升IO性能！但是多个目录存储的数据并不相同哟！而是把数据存放在不同的目录，当namenode存储数据时效率更高！ 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.heartbeat.interval 参数的作用：
#设置心跳检测时间 dfs.namenode.heartbeat.recheck-interval和dfs.heartbeat.interval 参数的作用：
#设置HDFS中NameNode和DataNode的超时时间，计算公式为:timeout = * dfs.namenode.heartbeat.recheck-interval + * dfs.heartbeat.interval。 dfs.hosts 参数的作用：
#添加白名单，功能和黑名单（dfs.hosts.exclude）相反。 dfs.hosts.exclude 参数的作用:
#这是我们添加的黑名单，该属性的value定义的是一个配置文件，这个配置文件的主机就是需要退役的节点。我这里是将其注释掉了。 --> [root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# cat /yinzhengjie/softwares/hadoop-2.9.2/etc/hadoop/hdfs-site.xml
[root@node101.yinzhengjie.org.cn ~]# cat /yinzhengjie/softwares/hadoop-2.9./etc/hadoop/slaves
node101.yinzhengjie.org.cn
node102.yinzhengjie.org.cn
node103.yinzhengjie.org.cn
node110.yinzhengjie.org.cn
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# cat /yinzhengjie/softwares/hadoop-2.9.2/etc/hadoop/slaves
[root@node101.yinzhengjie.org.cn ~]# stop-dfs.sh
Stopping namenodes on [node101.yinzhengjie.org.cn]
node101.yinzhengjie.org.cn: stopping namenode
node101.yinzhengjie.org.cn: stopping datanode
node102.yinzhengjie.org.cn: stopping datanode
node103.yinzhengjie.org.cn: stopping datanode
node110.yinzhengjie.org.cn: stopping datanode
Stopping secondary namenodes [account.jetbrains.com]
account.jetbrains.com: stopping secondarynamenode
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# stop-dfs.sh
[root@node101.yinzhengjie.org.cn ~]# scp -r /yinzhengjie/softwares/hadoop-2.9./ node102.yinzhengjie.org.cn:/yinzhengjie/softwares/ ........
hadoop-yarn-server-applicationhistoryservice-2.9..jar % 528KB .6MB/s :
mssql-jdbc-6.2..jre7.jar % 774KB .0MB/s :
hadoop-streaming-2.9..jar % 132KB .4MB/s :
commons-math3-3.1..jar % 1562KB .9MB/s :
json-io-2.5..jar % 73KB .5MB/s :
ojalgo-43.0.jar % 1625KB .2MB/s :
hadoop-azure-datalake-2.9..jar % 52KB .3MB/s :
README.txt % .0MB/s :
mapred-config.cmd % .0MB/s :
hdfs-config.sh % .0MB/s :
httpfs-config.sh % .9MB/s :
hdfs-config.cmd % .0MB/s :
yarn-config.cmd % .2MB/s :
mapred-config.sh % .9MB/s :
hadoop-config.cmd % .1MB/s :
hadoop-config.sh % 11KB .5MB/s :
yarn-config.sh % .1MB/s :
kms-config.sh % .0MB/s :
libhadooppipes.a % 1594KB .4MB/s :
libhadooputils.a % 465KB .5MB/s :
libhdfs.so.0.0. % 278KB .1MB/s :
wordcount-part % 933KB .2MB/s :
wordcount-simple % 924KB .8MB/s :
pipes-sort % 910KB .9MB/s :
wordcount-nopipe % 972KB .6MB/s :
libhdfs.so % 278KB .1MB/s :
libhdfs.a % 444KB .7MB/s :
libhadoop.so % 822KB .9MB/s :
libhadoop.a % 1406KB .3MB/s :
libhadoop.so.1.0. % 822KB .9MB/s :
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# scp -r /yinzhengjie/softwares/hadoop-2.9.2/ node102.yinzhengjie.org.cn:/yinzhengjie/softwares/
[root@node101.yinzhengjie.org.cn ~]# scp -r /yinzhengjie/softwares/hadoop-2.9./ node103.yinzhengjie.org.cn:/yinzhengjie/softwares/ ......
hadoop-yarn-server-applicationhistoryservice-2.9..jar % 528KB .0MB/s :
mssql-jdbc-6.2..jre7.jar % 774KB .6MB/s :
hadoop-streaming-2.9..jar % 132KB .0MB/s :
commons-math3-3.1..jar % 1562KB .3MB/s :
json-io-2.5..jar % 73KB .8MB/s :
ojalgo-43.0.jar % 1625KB .0MB/s :
hadoop-azure-datalake-2.9..jar % 52KB .3MB/s :
README.txt % .0MB/s :
mapred-config.cmd % .6MB/s :
hdfs-config.sh % .2MB/s :
httpfs-config.sh % .0MB/s :
hdfs-config.cmd % .8MB/s :
yarn-config.cmd % .4MB/s :
mapred-config.sh % .4MB/s :
hadoop-config.cmd % .4MB/s :
hadoop-config.sh % 11KB .1MB/s :
yarn-config.sh % .6MB/s :
kms-config.sh % .5MB/s :
libhadooppipes.a % 1594KB .6MB/s :
libhadooputils.a % 465KB .3MB/s :
libhdfs.so.0.0. % 278KB .6MB/s :
wordcount-part % 933KB .3MB/s :
wordcount-simple % 924KB .2MB/s :
pipes-sort % 910KB .2MB/s :
wordcount-nopipe % 972KB .2MB/s :
libhdfs.so % 278KB .6MB/s :
libhdfs.a % 444KB .3MB/s :
libhadoop.so % 822KB .9MB/s :
libhadoop.a % 1406KB .2MB/s :
libhadoop.so.1.0. % 822KB .3MB/s :
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# scp -r /yinzhengjie/softwares/hadoop-2.9.2/ node103.yinzhengjie.org.cn:/yinzhengjie/softwares/
[root@node101.yinzhengjie.org.cn ~]# scp -r /yinzhengjie/softwares/hadoop-2.9./ node110.yinzhengjie.org.cn:/yinzhengjie/softwares/ .......
mssql-jdbc-6.2..jre7.jar % 774KB .9MB/s :
hadoop-streaming-2.9..jar % 132KB .1MB/s :
commons-math3-3.1..jar % 1562KB .7MB/s :
json-io-2.5..jar % 73KB .7MB/s :
ojalgo-43.0.jar % 1625KB .2MB/s :
hadoop-azure-datalake-2.9..jar % 52KB .5MB/s :
README.txt % .5MB/s :
mapred-config.cmd % .8MB/s :
hdfs-config.sh % .6KB/s :
httpfs-config.sh % .9MB/s :
hdfs-config.cmd % .1MB/s :
yarn-config.cmd % .3MB/s :
mapred-config.sh % .0MB/s :
hadoop-config.cmd % .9MB/s :
hadoop-config.sh % 11KB .7MB/s :
yarn-config.sh % .7MB/s :
kms-config.sh % .6MB/s :
libhadooppipes.a % 1594KB .1MB/s :
libhadooputils.a % 465KB .7MB/s :
libhdfs.so.0.0. % 278KB .3MB/s :
wordcount-part % 933KB .5MB/s :
wordcount-simple % 924KB .3MB/s :
pipes-sort % 910KB .6MB/s :
wordcount-nopipe % 972KB .4MB/s :
libhdfs.so % 278KB .4MB/s :
libhdfs.a % 444KB .2MB/s :
libhadoop.so % 822KB .9MB/s :
libhadoop.a % 1406KB .2MB/s :
libhadoop.so.1.0. % 822KB .3MB/s :
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# scp -r /yinzhengjie/softwares/hadoop-2.9.2/ node110.yinzhengjie.org.cn:/yinzhengjie/softwares/
2>.编辑白名单
[root@node101.yinzhengjie.org.cn ~]# cat /data/hadoop/hdfs/DataNodesHostname.txt
node101.yinzhengjie.org.cn
node103.yinzhengjie.org.cn
node110.yinzhengjie.org.cn
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# cat /data/hadoop/hdfs/DataNodesHostname.txt
[root@node101.yinzhengjie.org.cn ~]# start-dfs.sh
Starting namenodes on [node101.yinzhengjie.org.cn]
node101.yinzhengjie.org.cn: starting namenode, logging to /yinzhengjie/softwares/hadoop-2.9./logs/hadoop-root-namenode-node101.yinzhengjie.org.cn.out
node102.yinzhengjie.org.cn: starting datanode, logging to /yinzhengjie/softwares/hadoop-2.9./logs/hadoop-root-datanode-node102.yinzhengjie.org.cn.out
node101.yinzhengjie.org.cn: starting datanode, logging to /yinzhengjie/softwares/hadoop-2.9./logs/hadoop-root-datanode-node101.yinzhengjie.org.cn.out
node103.yinzhengjie.org.cn: starting datanode, logging to /yinzhengjie/softwares/hadoop-2.9./logs/hadoop-root-datanode-node103.yinzhengjie.org.cn.out
node110.yinzhengjie.org.cn: starting datanode, logging to /yinzhengjie/softwares/hadoop-2.9./logs/hadoop-root-datanode-node110.yinzhengjie.org.cn.out
Starting secondary namenodes [account.jetbrains.com]
account.jetbrains.com: starting secondarynamenode, logging to /yinzhengjie/softwares/hadoop-2.9./logs/hadoop-root-secondarynamenode-node101.yinzhengjie.org.cn.out
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# jps
Jps
NameNode
DataNode
SecondaryNameNode
[root@node101.yinzhengjie.org.cn ~]#
[root@node101.yinzhengjie.org.cn ~]# start-dfs.sh
3>.关于yarn的服役和退役
配置的套路和上面配置DataNode的套路几乎一致！详情请参考：https://www.cnblogs.com/yinzhengjie/p/9101070.html。