码迷,mamicode.com
首页 > 编程语言 > 详细

使用HDFS客户端java api读取hadoop集群上的信息

时间:2017-07-22 16:49:09      阅读:173      评论:0      收藏:0      [点我收藏+]

标签:exist   log   int   form   efault   pid   file   his   hdfs   

本文介绍使用hdfs java api的配置方法。

1、先解决依赖,pom

<dependency>
            <groupId>org.apache.hadoop</groupId>
            <artifactId>hadoop-client</artifactId>
            <version>2.7.2</version>
            <scope>provided</scope>
        </dependency>

2、配置文件,存放hdfs集群配置信息,基本都是来源于core-site.xml和hdfs-site.xml,可以根据hdfs集群client端配置文件里的信息进行填写

#============== hadoop ===================
hdfs.fs.defaultFS=hdfs://mycluster-tj
hdfs.ha.zookeeper.quorum=XXXX-apache00.XX01,XXXX-apache01.XX01,XXXX-apache02.XX01
hdfs.dfs.nameservices=XXXX
hdfs.dfs.ha.namenodes.mycluster-tj=XX1,XX2
hdfs.dfs.namenode.rpc-address.mycluster-tj.nn1=XXXX-apachenn01.XX01:8020
hdfs.dfs.namenode.rpc-address.mycluster-tj.nn2=XXXX-apachenn02.XX01:8020

3、java client api

import java.io.IOException;
import java.net.URI;
import org.apache.hadoop.fs.FileStatus;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.conf.Configuration;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import java.text.SimpleDateFormat;
import java.util.Date;


public class HadoopClient {
    protected final Logger logger = LoggerFactory.getLogger(this.getClass());
    private FileSystem fs;
    private String defaultFS;
    private String zKQuorum;
    private String nameServices;
    private String nameNodes;
    private String rpcAddressNN1;
    private String rpcAddressNN2;

    public void setDefaultFS(String defaultFS) {
        this.defaultFS = defaultFS;
    }
    public String getDefaultFS() {
        return defaultFS;
    }
    public void setZKQuorum(String zKQuorum) {
        this.zKQuorum = zKQuorum;
    }
    public String getzKQuorum() {
        return zKQuorum;
    }
    public void setNameServices(String nameServices) {
        this.nameServices = nameServices;
    }
    public String getNameServices() {
        return nameServices;
    }
    public void setNameNodes(String nameNodes) {
        this.nameNodes = nameNodes;
    }
    public String getNameNodes() {
        return nameNodes;
    }
    public void setRpcAddressNN1(String rpcAddressNN1) {
        this.rpcAddressNN1 = rpcAddressNN1;
    }
    public String getRpcAddressNN1() {
        return rpcAddressNN1;
    }
    public void setRpcAddressNN2(String rpcAddressNN2) {
        this.rpcAddressNN2 = rpcAddressNN2;
    }
    public String getRpcAddressNN2() {
        return rpcAddressNN2;
    }

    public void init() {
        try {
            Configuration conf = new Configuration();
            conf.set("fs.defaultFS", defaultFS);
            conf.set("ha.zookeeper.quorum", zKQuorum);
            conf.set("dfs.nameservice", nameServices);
            conf.set("dfs.ha.namenodes.mycluster-tj", nameNodes);
            conf.set("dfs.namenode.rpc-address.mycluster-tj.nn1", rpcAddressNN1);
            conf.set("dfs.namenode.rpc-address.mycluster-tj.nn2", rpcAddressNN2);
            fs = FileSystem.get(new URI(defaultFS), conf);
        } catch (Exception ex) {
            ex.printStackTrace();
        }
    }

    public void stop() {
        try {
            fs.close();
        } catch(Exception e) {

        }
    }

    public boolean exists(String path) {
        boolean isExists = false;
        try {
            Path hdfsPath = new Path(path);
            isExists = fs.exists(hdfsPath);
        } catch (Exception ex) {
            logger.error("exists error: {}", ex.getMessage());
        }
        return isExists;
    }

    public String getModificationTime(String path) throws IOException {
        String modifyTime = null;
        try {
            Path hdfsPath = new Path(path);
            FileStatus fileStatus = fs.getFileStatus(hdfsPath);
            long modifyTimestamp = fileStatus.getModificationTime();
            SimpleDateFormat simpleDateFormat = new SimpleDateFormat("yyyyMMddHHmmss");
            Date date = new Date(modifyTimestamp);
            modifyTime = simpleDateFormat.format(date);
        } catch(Exception ex) {
            logger.error("getModificationTime error: {}", ex.getMessage());
        }
        return modifyTime;
    }


}

4、configuration

import com.xiaoju.dqa.prometheus.client.hadoop.HadoopClient;
import org.springframework.beans.factory.annotation.Value;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;

@Configuration
public class HadoopConfiguration {
    @Value("${hdfs.fs.defaultFS}")
    private String defaultFS;
    @Value("${hdfs.ha.zookeeper.quorum}")
    private String zKQuorum;
    @Value("${hdfs.dfs.nameservices}")
    private String nameServices;
    @Value("${hdfs.dfs.ha.namenodes.mycluster-tj}")
    private String nameNodes;
    @Value("${hdfs.dfs.namenode.rpc-address.mycluster-tj.nn1}")
    private String rpcAddressNN1;
    @Value("${hdfs.dfs.namenode.rpc-address.mycluster-tj.nn2}")
    private String rpcAddressNN2;

    @Bean(initMethod = "init", destroyMethod = "stop")
    public HadoopClient hadoopClient() {
        HadoopClient hadoopClient = new HadoopClient();
        hadoopClient.setDefaultFS(defaultFS);
        hadoopClient.setZKQuorum(zKQuorum);
        hadoopClient.setNameServices(nameServices);
        hadoopClient.setNameNodes(nameNodes);
        hadoopClient.setRpcAddressNN1(rpcAddressNN1);
        hadoopClient.setRpcAddressNN2(rpcAddressNN2);
        return hadoopClient;
    }
}

 

使用HDFS客户端java api读取hadoop集群上的信息

标签:exist   log   int   form   efault   pid   file   his   hdfs   

原文地址:http://www.cnblogs.com/kangoroo/p/7221527.html

(0)
(0)
   
举报
评论 一句话评论(0
登录后才能评论!
© 2014 mamicode.com 版权所有  联系我们:gaon5@hotmail.com
迷上了代码!