记一次本机使用 Java API 读取虚拟机 HDFS 文件出错：Could not obtain block…

背景

• 虚拟机安装 Hadoop 3.1.3，并运行了 HDFS。

• 使用本机 Java API 读取文件出错，在虚拟机中正常。
  

• 代码

package com.ggkkdd.hdfs.example;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.*;

import java.io.IOException;
import java.io.PrintStream;
import java.net.URI;

public class MergeFile {

    Path inputPath = null;
    Path outputPath = null;

    public MergeFile(Path inputPath, Path outputPath) {
        this.inputPath = inputPath;
        this.outputPath = outputPath;
    }

    public MergeFile(String inputPath, String outputPath) {
        this.inputPath = new Path(inputPath);
        this.outputPath = new Path(outputPath);
    }

    public void doMerge() throws IOException {
        Configuration conf = new Configuration();
        System.setProperty("HADOOP_USER_NAME", "hadoop");

        conf.set("fs.defaultFS", "hdfs://hadoop-master:9000");
        conf.set("fs.hdfs.impl", "org.apache.hadoop.hdfs.DistributedFileSystem");

        FileSystem fsSource = FileSystem.get(URI.create(inputPath.toString()), conf);
        FileSystem fsDst = FileSystem.get(URI.create(outputPath.toString()), conf);

        FileStatus[] sourceStatus = fsSource.listStatus(inputPath, new MyPathFilter(".*\\.abc"));
        System.out.println(sourceStatus.length);

        FSDataOutputStream fsDos = fsDst.create(outputPath);
        PrintStream ps = new PrintStream(System.out);

        for (FileStatus sta : sourceStatus) {
            System.out.println("路径：" + sta.getPath());
            System.out.println("文件大小：" + sta.getLen());
            System.out.println("权限：" + sta.getPermission());
            System.out.println("内容 ：");

            FSDataInputStream fsDis = fsSource.open(sta.getPath());
            byte[] data = new byte[1024];
            int read;
            while ((read = fsDis.read(data)) > 0) {
                System.out.println(read);
                ps.write(data, 0, read);
                fsDos.write(data, 0, read);
            }

            fsDis.close();
        }
    }

    public static void main(String[] args) throws IOException {
        MergeFile merge = new MergeFile("hdfs://hadoop-master:9000/user/hadoop/input",
                "hdfs://hadoop-master:9000/user/hadoop/output/merge.txt");
        // MergeFile merge = new MergeFile("/user/hadoop/input",
        //         "/user/hadoop/output/merge.txt");
        merge.doMerge();
    }

}

class MyPathFilter implements PathFilter {

    String reg = null;

    public MyPathFilter(String reg) {
        this.reg = reg;
    }

    @Override
    public boolean accept(Path path) {
        return !path.toString().matches(reg);
    }

}

• 错误信息

...
Connected to the target VM, address: '127.0.0.1:50938', transport: 'socket'
log4j:WARN No appenders could be found for logger (org.apache.hadoop.util.Shell).
log4j:WARN Please initialize the log4j system properly.
log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info.
9
路径：hdfs://hadoop-master:9000/user/hadoop/input/capacity-scheduler.xml
文件大小：8260
权限：rw-r--r--
内容 ：
Exception in thread "main" org.apache.hadoop.hdfs.BlockMissingException: Could not obtain block: BP-1745224033-127.0.1.1-1639451530131:blk_1073741825_1001 file=/user/hadoop/input/capacity-scheduler.xml
	at org.apache.hadoop.hdfs.DFSInputStream.refetchLocations(DFSInputStream.java:875)
	at org.apache.hadoop.hdfs.DFSInputStream.chooseDataNode(DFSInputStream.java:858)
	at org.apache.hadoop.hdfs.DFSInputStream.chooseDataNode(DFSInputStream.java:837)
	at org.apache.hadoop.hdfs.DFSInputStream.blockSeekTo(DFSInputStream.java:566)
	at org.apache.hadoop.hdfs.DFSInputStream.readWithStrategy(DFSInputStream.java:756)
	at org.apache.hadoop.hdfs.DFSInputStream.read(DFSInputStream.java:825)
	at java.io.DataInputStream.read(DataInputStream.java:100)
	at com.ggkkdd.hdfs.example.MergeFile.doMerge(MergeFile.java:53)
	at com.ggkkdd.hdfs.example.MergeFile.main(MergeFile.java:68)
Disconnected from the target VM, address: '127.0.0.1:50938', transport: 'socket'

Process finished with exit code 1

原因

由于是外部访问远程 HDFS，读取文件的时候是直接访问 NameNode，能正常返回元数据等，但读取数据的时候，NameNode 返回的 DataNode 节点的地址 127.0.0.1 在本机无法访问。

解决方法

需要在代码中开启 客户端使用 DataNote 的主机名。

conf.set("dfs.client.use.datanode.hostname", "true");

或者在文件中配置

<configuration>
    ...
    <property>
        <name>dfs.client.use.datanode.hostname</name>
        <value>true</value>
    </property>
</configuration>

最后在本机中绑定 DataNote 的主机名和 IP：

sudo echo "x.x.x.x	xxxhostname" >> /etc/hosts

可以在服务器使用 hostname 查看主机名。

---

如果不知道 hostname、无配置 host 的话，会出现无法解决网络地址问题。

那么可以用 debug 模式打断点进去看主机名。

成功结果

本机 host 配置

10.211.55.3	hadoop-master
10.211.55.3	ubuntu-linux-20-04-desktop

Java API 调用

conf.set("dfs.client.use.datanode.hostname", "true");

参考

在 Google 寻找解决方法半天没头绪，最后回百度看到了一条类似的问题，于是乎参考并实践了一下，完美解决！

eclipse连接hadoop报错Could not obtain block-Java-CSDN问答