hadoop2.5.2配置httpfs服务

wstlwl

hadoop2.5.2配置httpfs服务

 
httpfs hadoop hdfs



测试环境

• ubuntu 14.04 单机
  
• hadoop2.5.2 伪分布式
  
• jdk1.7
  

作用

• 通过HttpFs你可以在浏览器里面管理HDFS上的文件，功能同hadoop shell相似
  
• HttpFs还提供了一套REST 风格的API可以用来管理HDFS
  

缺点

• 网络安全性
  
• 需要配置acl
  

修改core-site.xml

 

    <property>
<name>hadoop.proxyuser.root.hosts</name>
<value>localhost</value>
</property>
<property>
<name>hadoop.proxyuser.root.groups</name>
<value>*</value>
</property>

[size=1em]添加上述两个配置hadoop.proxyuser.root.hosts允许通过httpfs方式访问hdfs的主机名或者域名；hadoop.proxyuser.root.groups允许访问的客户端的用户组

配置httpfs环境变量

 

export CATALINA_BASE=/opt/work/hadoop/share/hadoop/httpfs/tomcat

[size=1em]临时配置，也可以在配置文件配置环境变量

重启hadoop集群

 

root@localhost:/opt/nfs#/opt/work/hadoop/sbin//stop-all.sh
root@localhost:/opt/nfs#/opt/work/hadoop/sbin//start-all.sh


启动httpfs服务

 

root@localhost:/opt/git/hadoop_dev/hdfsToInfoBright# /opt/work/hadoop/sbin/httpfs.sh start
Setting HTTPFS_HOME:          /opt/work/hadoop
Setting HTTPFS_CONFIG:        /opt/work/hadoop/etc/hadoop
Sourcing:                    /opt/work/hadoop/etc/hadoop/httpfs-env.sh
Setting HTTPFS_LOG:           /opt/work/hadoop/logs
Setting HTTPFS_TEMP:           /opt/work/hadoop/temp
Setting HTTPFS_HTTP_PORT:     14000
Setting HTTPFS_ADMIN_PORT:     14001
Setting HTTPFS_HTTP_HOSTNAME: localhost
Setting HTTPFS_SSL_ENABLED: false
Setting HTTPFS_SSL_KEYSTORE_FILE:     /root/.keystore
Setting HTTPFS_SSL_KEYSTORE_PASS:     password
Using   CATALINA_BASE:       /opt/work/hadoop/share/hadoop/httpfs/tomcat/
Setting HTTPFS_CATALINA_HOME:       /opt/work/hadoop/share/hadoop/httpfs/tomcat/
Setting CATALINA_OUT:        /opt/work/hadoop/logs/httpfs-catalina.out
Setting CATALINA_PID:        /tmp/httpfs.pid
Using   CATALINA_OPTS:      
Adding to CATALINA_OPTS:     -Dhttpfs.home.dir=/opt/work/hadoop -Dhttpfs.config.dir=/opt/work/hadoop/etc/hadoop -Dhttpfs.log.dir=/opt/work/hadoop/logs -Dhttpfs.temp.dir=/opt/work/hadoop/temp -Dhttpfs.admin.port=14001 -Dhttpfs.http.port=14000 -Dhttpfs.http.hostname=localhost -Dhttpfs.ssl.enabled=false -Dhttpfs.ssl.keystore.file=/root/.keystore -Dhttpfs.ssl.keystore.pass=password
Using CATALINA_BASE:   /opt/work/hadoop/share/hadoop/httpfs/tomcat/
Using CATALINA_HOME:   /opt/work/hadoop/share/hadoop/httpfs/tomcat
Using CATALINA_TMPDIR: /opt/work/hadoop/share/hadoop/httpfs/tomcat//temp
Using JRE_HOME:        /usr/local/jdk1.7.0
Using CLASSPATH:       /opt/work/hadoop/share/hadoop/httpfs/tomcat//bin/tomcat-juli.jar:/opt/work/hadoop/share/hadoop/httpfs/tomcat/bin/bootstrap.jar
Using CATALINA_PID:    /tmp/httpfs.pid
Existing PID file found during start.
Removing/clearing stale PID file.
root@localhost:/opt/git/hadoop_dev/hdfsToInfoBright#

[size=1em]并查看启动日志，并确定环境变量的正确CATALINA_BASE的值

浏览器访问httpfs

http://localhost:14000/ 
页面显示：HttpFs service, service base URL at /webhdfs/v1. 根目录为：/webhdfs/v1



curl访问httpfs

 

root@localhost:/opt/nfs# curl -i -X PUT -T /opt/test.json "http://localhost:14000/webhdfs/v1/tmp/test.json?op=CREATE&data=true&user.name=root" -H "Content-Type:application/octet-stream"

[size=1em]将文件上传到hdfs 
http访问文件： 
http://localhost:14000/webhdfs/v1/tmp/test.json?user.name=root&op=open 
即可下载test.json文件

考虑

• httpfs的安全性
  
• httpfs的性能：httpfs是一个tomcat启动的web服务，当数据文件非常大，上传下载的性能将骤减
  
• httpfs的并发访问：tomcat的并发问题
  

参考文献

官网：http://hadoop.apache.org/docs/r2.5.2/hadoop-hdfs-httpfs/ServerSetup.html