复制本地文件到HDFS本地测试异常

项目中需要将本地文件拷贝到hdfs上，由于本人比较懒，于是使用擅长的Java程序通过Hadoop.FileSystem.CopyFromLocalFile方法来实现。在本地(Window 7 环境)本地模式下运行却遇到了下述异常：

An exception or error caused a run to abort: org.apache.hadoop.io.nativeio.NativeIO$Windows.createFileWithMode0(Ljava/lang/String;JJJI)Ljava/io/FileDescriptor; 
java.lang.UnsatisfiedLinkError: org.apache.hadoop.io.nativeio.NativeIO$Windows.createFileWithMode0(Ljava/lang/String;JJJI)Ljava/io/FileDescriptor;
    at org.apache.hadoop.io.nativeio.NativeIO$Windows.createFileWithMode0(Native Method)
    at org.apache.hadoop.io.nativeio.NativeIO$Windows.createFileOutputStreamWithMode(NativeIO.java:559)
    at org.apache.hadoop.fs.RawLocalFileSystem$LocalFSFileOutputStream.<init>(RawLocalFileSystem.java:219)
    at org.apache.hadoop.fs.RawLocalFileSystem$LocalFSFileOutputStream.<init>(RawLocalFileSystem.java:209)
    at org.apache.hadoop.fs.RawLocalFileSystem.createOutputStreamWithMode(RawLocalFileSystem.java:307)
    at org.apache.hadoop.fs.RawLocalFileSystem.create(RawLocalFileSystem.java:295)
    at org.apache.hadoop.fs.RawLocalFileSystem.create(RawLocalFileSystem.java:328)
    at org.apache.hadoop.fs.ChecksumFileSystem$ChecksumFSOutputSummer.<init>(ChecksumFileSystem.java:388)
    at org.apache.hadoop.fs.ChecksumFileSystem.create(ChecksumFileSystem.java:451)
    at org.apache.hadoop.fs.ChecksumFileSystem.create(ChecksumFileSystem.java:430)
    at org.apache.hadoop.fs.FileSystem.create(FileSystem.java:920)
    at org.apache.hadoop.fs.FileSystem.create(FileSystem.java:901)
    at org.apache.hadoop.fs.FileSystem.create(FileSystem.java:798)
    at org.apache.hadoop.fs.FileUtil.copy(FileUtil.java:368)
    at org.apache.hadoop.fs.FileUtil.copy(FileUtil.java:341)
    at org.apache.hadoop.fs.FileUtil.copy(FileUtil.java:292)
    at org.apache.hadoop.fs.LocalFileSystem.copyFromLocalFile(LocalFileSystem.java:82)
    at org.apache.hadoop.fs.FileSystem.copyFromLocalFile(FileSystem.java:1882)

通过分析异常堆栈可知，

org.apache.hadoop.io.nativeio.NativeIO$Windows.createFileWithMode0方法发生了异常，createFileWithMode0方法的实现如下：

    /** Wrapper around CreateFile() with security descriptor on Windows */
    private static native FileDescriptor createFileWithMode0(String path,
        long desiredAccess, long shareMode, long creationDisposition, int mode)
        throws NativeIOException;

通过代码可知，这个方法是hadoop不支持的方法。那么为什么会调用这个方法，通过异常堆栈继续向上追踪，

是在 org.apache.hadoop.fs.RawLocalFileSystem$LocalFSFileOutputStream.<init> 过程调用了nativeio.NativeIO$Windows类，对应的方法如下：

 1  private LocalFSFileOutputStream(Path f, boolean append,
 2         FsPermission permission) throws IOException {
 3       File file = pathToFile(f);
 4       if (permission == null) {
 5         this.fos = new FileOutputStream(file, append);
 6       } else {
 7         if (Shell.WINDOWS && NativeIO.isAvailable()) {
 8           this.fos = NativeIO.Windows.createFileOutputStreamWithMode(file,
 9               append, permission.toShort());
10         } else {
11           this.fos = new FileOutputStream(file, append);
12           boolean success = false;
13           try {
14             setPermission(f, permission);
15             success = true;
16           } finally {
17             if (!success) {
18               IOUtils.cleanup(LOG, this.fos);
19             }
20           }
21         }
22       }
23     }

通过调用椎栈可知是上述代码第8行调用了NativeIO.Windows类。那么if判断应该是成立的，分析NativeIO.isAvailable方法代码如下：

1   /**
2    * Return true if the JNI-based native IO extensions are available.
3    */
4   public static boolean isAvailable() {
5     return NativeCodeLoader.isNativeCodeLoaded() && nativeLoaded;
6   }

isAvailable方法主要是调用NativeCodeLoader.isNativeCodeLoaded方法

 1  static {
 2     // Try to load native hadoop library and set fallback flag appropriately
 3     if(LOG.isDebugEnabled()) {
 4       LOG.debug("Trying to load the custom-built native-hadoop library...");
 5     }
 6     try {
 7       System.loadLibrary("hadoop");
 8       LOG.debug("Loaded the native-hadoop library");
 9       nativeCodeLoaded = true;
10     } catch (Throwable t) {
11       // Ignore failure to load
12       if(LOG.isDebugEnabled()) {
13         LOG.debug("Failed to load native-hadoop with error: " + t);
14         LOG.debug("java.library.path=" +
15             System.getProperty("java.library.path"));
16       }
17     }
18     
19     if (!nativeCodeLoaded) {
20       LOG.warn("Unable to load native-hadoop library for your platform... " +
21                "using builtin-java classes where applicable");
22     }
23   }
24 
25   /**
26    * Check if native-hadoop code is loaded for this platform.
27    * 
28    * @return <code>true</code> if native-hadoop is loaded, 
29    *         else <code>false</code>
30    */
31   public static boolean isNativeCodeLoaded() {
32     return nativeCodeLoaded;
33   }

通过可以看到，isNativeCodeLoaded方法就是返回一个属性值，那么问题出现在什么地方呢？

经过分析NativeCodeLoaded类的静态构造函数，有一个“System.loadLibrary("hadoop")”方法。是不是这个方法导致的呢？通过在其他同事环境上调试，System.loadLibrary("hadoop") 会异常，从而运行catch部分，但是本人电脑却不会异常，直接继续运行。那么System.loadLibrary方法是什么用途呢，通过分析源码知道，这个方法是加载本地系统和用户的环境变量的。进而分析是因为本人在C:\WindowsSystem32目录下有hadoop.dll文件或环境变量Path中配置了%Hadoop_Home%/bin目录而导致的。

简而言之，是因为配置的系统环境变量Path的任意目录下存在hadoop.dll文件，从而被认为这是一个hadoop集群环境，但是hadoop集群又不支持window环境而产生的异常。处理方法也很简单，检查系统环境变量Path下的每一个目录，确保没有hadoop.dll文件即可。

如果是删除系统环境变量Path的某一个目录，需要重启Intellij Idea后ClassLoader类中的usr_paths或sys_paths才会生效。