HDFS学习二：客户端操作

一、环境准备及测试

1、HDFS客户端环境准备

将编译后的Hadoop的包放到指定路径，然后环境遍历配置好Hadoop_HOME和PATH

2、创建一个Maven工程

pom文件如下

<dependencies>
  <dependency>
    <groupId>junit</groupId>
    <artifactId>junit</artifactId>
    <version>4.11</version>
    <scope>test</scope>
  </dependency>

  <dependency>
    <groupId>org.apache.logging.log4j</groupId>
    <artifactId>log4j-core</artifactId>
    <version>2.11.0</version>
  </dependency>

  <dependency>
    <groupId>org.apache.hadoop</groupId>
    <artifactId>hadoop-common</artifactId>
    <version>3.3.1</version>
  </dependency>

  <dependency>
    <groupId>org.apache.hadoop</groupId>
    <artifactId>hadoop-client</artifactId>
    <version>3.3.1</version>
  </dependency>

  <dependency>
    <groupId>org.apache.hadoop</groupId>
    <artifactId>hadoop-hdfs</artifactId>
    <version>3.3.1</version>
  </dependency>

</dependencies>

 创建一个log4j.properties

log4j.rootLogger=INFO, stdout
log4j.appender.stdout=org.apache.log4j.ConsoleAppender
log4j.appender.stdout.layout=org.apache.log4j.PatternLayout
log4j.appender.stdout.layout.ConversionPattern=%d %p [%c] - %m%n
log4j.appender.logfile=org.apache.log4j.FileAppender
log4j.appender.logfile.File=target/spring.log
log4j.appender.logfile.layout=org.apache.log4j.PatternLayout
log4j.appender.logfile.layout.ConversionPattern=%d %p [%c] -%m%n

写个启动类

public class HDFSClient {
    public static void main(String[] args) throws IOException, URISyntaxException, InterruptedException {
        Configuration conf = new Configuration();

        //获取客户端对象
        FileSystem fs = FileSystem.get(new URI("hdfs://002:9000"),conf,"user");

        //修改路径
        fs.mkdirs(new Path("/loong/1231"));

        //关闭资源
        fs.close();

        System.out.println("over");
    }
}

二、HDFS的API操作

2.1文件上传

@Test
public void testCopyFromLocal() throws IOException, URISyntaxException, InterruptedException {
    Configuration conf = new Configuration();

    //获取客户端对象
    FileSystem fs = FileSystem.get(new URI("hdfs://002:9000"), conf, "user");

    //执行上传操作
    fs.copyFromLocalFile(new Path("D:/WORK/tianlongbabu.txt"), new Path("/tianlongbabu.txt"));

    //关闭资源
    fs.close();

    System.out.println("over");
}

2.2测试参数优先级

设置的参数优先级关系如下：

代码中设置的参数 > resources文件夹下的xml > 集群中etc目录下的xml文件中的设置 > hadoop的参数默认值

2.3文件下载

fs.copyToLocalFile(new Path("/tianlongbabu.txt"), new Path("d:/tianlongbau.txt"));

若是修改一下代码

 fs.copyToLocalFile(true,new Path("/tianlongbabu.txt"), new Path("d:/tianlongbau.txt"),true);

第一个true：是否删除hdfs上文件（设置true就相当于剪切）

第二个true：设置true就相当于同意下载下来的文件能够允许在本地修改（关闭了校验，就会少下一个crc文件）

2.4文件删除

delete中boolean的参数主要针对目录，意思是，true的话就会递归删除，false的话文件不为空会报异常。

fs.delete(new Path("/loong"), true);

2.5文件更名

fs.rename(new Path("/loong/tianlongbabu.txt"),new Path("/loong/kuihuabaodian.txt"));

2.6文件详情查看

@Test
public void testListFiles() throws IOException, URISyntaxException, InterruptedException {
    Configuration conf = new Configuration();
    conf.set("dfs.client.use.datanode.hostname", "true");
    //获取客户端对象
    FileSystem fs = FileSystem.get(new URI("hdfs://LOONG002:9000"), conf, "root");
    RemoteIterator<LocatedFileStatus> listFiles = fs.listFiles(new Path("/"), true);
    while (listFiles.hasNext()) {
        LocatedFileStatus fileStatus = listFiles.next();

        //查看文件名称，权限，长度，块信息
        System.out.println(fileStatus.getPath().getName());
        System.out.println(fileStatus.getPermission());
        System.out.println(fileStatus.getLen());
        BlockLocation[] blockLocations = fileStatus.getBlockLocations();
        for (BlockLocation blockLocation : blockLocations) {
            String[] hosts = blockLocation.getHosts();
            for (String host : hosts) {
                System.out.println(host);
            }

        }
        System.out.println("--------分割线----------");
    }

    //关闭资源
    fs.close();
    System.out.println("over");
}

2.7判断是文件还是文件夹

@Test
public void testListStatus() throws IOException, URISyntaxException, InterruptedException {
    Configuration conf = new Configuration();
    conf.set("dfs.client.use.datanode.hostname", "true");
    //获取客户端对象
    FileSystem fs = FileSystem.get(new URI("hdfs://LOONG002:9000"), conf, "root");

    FileStatus[] listStatus=fs.listStatus(new Path("/"));
    for (FileStatus fileStatus:listStatus){
        if(fileStatus.isFile()){
            System.out.println("f:"+fileStatus.getPath().getName());
        }else {
            System.out.println("d:"+fileStatus.getPath().getName());
        }
    }

    //关闭资源
    fs.close();

    System.out.println("over");
}

三、HDFS的I/O流操作

3.1HDFS文件上传

@Test
public void putFileToHDFS() throws IOException, URISyntaxException, InterruptedException {
    Configuration conf = new Configuration();
    conf.set("dfs.client.use.datanode.hostname", "true");
    //获取客户端对象
    FileSystem fs = FileSystem.get(new URI("hdfs://LOONG002:9000"), conf, "root");

    //获取输入流
    FileInputStream fis = new FileInputStream(new File("d:/work/pen.txt"));

    //获取输出流
    FSDataOutputStream fos = fs.create(new Path("/pen.txt"));

    //流的对拷
    IOUtils.copyBytes(fis, fos, conf);

    //关闭资源
    IOUtils.closeStream(fos);
    IOUtils.closeStream(fis);
    fs.close();

    System.out.println("over");
}

3.2文件下载

@Test
public void getFileFromHDFS() throws IOException, URISyntaxException, InterruptedException {
    Configuration conf = new Configuration();
    conf.set("dfs.client.use.datanode.hostname", "true");
    //获取客户端对象
    FileSystem fs = FileSystem.get(new URI("hdfs://LOONG002:9000"), conf, "root");

    //获取输入流
    FSDataInputStream fis = fs.open(new Path("/pen.txt"));;

    //获取输出流
    FileOutputStream fos =new FileOutputStream(new File("d:/work/pen.txt"));

    //流的对拷
    IOUtils.copyBytes(fis, fos, conf);

    //关闭资源
    IOUtils.closeStream(fos);
    IOUtils.closeStream(fis);
    fs.close();

    System.out.println("over");
}

3.3定位读取文件

下载第一块block文件

@Test
public void readFileSeek1() throws IOException, URISyntaxException, InterruptedException {
    Configuration conf = new Configuration();
    conf.set("dfs.client.use.datanode.hostname", "true");
    //获取客户端对象
    FileSystem fs = FileSystem.get(new URI("hdfs://LOONG002:9000"), conf, "root");

    //获取输入流
    FSDataInputStream fis = fs.open(new Path("/jdk-8u311-linux-x64.tar.gz"));

    //获取输出流
    FileOutputStream fos = new FileOutputStream(new File("d:/jdk-8u311-linux-x64.tar.gz.part1"));

    //流的对拷
    byte[] buf = new byte[1024];
    for (int i = 0; i < 1024 * 128; i++) {
        fis.read(buf);
        fos.write(buf);
    }

    //关闭资源
    IOUtils.closeStream(fos);
    IOUtils.closeStream(fis);
    fs.close();

    System.out.println("over");
}

下载第二块

@Test
public void readFileSeek2() throws IOException, URISyntaxException, InterruptedException {
    Configuration conf = new Configuration();
    conf.set("dfs.client.use.datanode.hostname", "true");
    //获取客户端对象
    FileSystem fs = FileSystem.get(new URI("hdfs://LOONG002:9000"), conf, "root");

    //获取输入流
    FSDataInputStream fis = fs.open(new Path("/jdk-8u311-linux-x64.tar.gz"));

    //设置读取的起点
    fis.seek(1024 * 1024 * 128);

    //获取输出流
    FileOutputStream fos = new FileOutputStream(new File("d:/jdk-8u311-linux-x64.tar.gz.part2"));

    //流的对拷
    IOUtils.copyBytes(fis, fos, conf);

    //关闭资源
    IOUtils.closeStream(fos);
    IOUtils.closeStream(fis);
    fs.close();

    System.out.println("over");
}

四、HDFS读写数据流机制

1.写数据流流程

1.2 网络拓扑-节点距离计算

HDFS写数据的时候，Name会选择距离待上传数据距离最近的Datanode接收数据。

这里涉及到计算节点距离；

节点距离：两个节点到达最近公共祖先的距离总和。

 

1.3机架感知（副本存储节点选择）

副本节点选择如图，这是2.7.2版本的

 2.HDFS读数据流程