一. 自己搭建开发环境
今天自己搭建了一套Centos5.3 + Hadoop2.2 + Hbase0.96.1.1的开发环境，Win7 Eclipse调试MapReduce成功。可能是版本比较高的原因，出了问题，网上找不到完整解决方案，只能靠自己。
二. Hadoop安装
这个就不啰嗦了，网上大把文章。我下载的是hadoop-2.2.0.tar.gz。
- http://www.cnblogs.com/xia520pi/archive/2012/05/16/2503949.html 很详细的介绍了Hadoop Hdfs的安装。它这个不是Hadoop2.2的，但配置都大同小异。
- MapReduce的配置可以参考http://blog.sina.com.cn/s/blog_546abd9f0101i8b8.html。
安装成功后，能顺利查看以下几个页面，就OK了。我的集群环境是200master，201-203slave。
- dfs.http.address 192.168.1.200:50070
- dfs.secondary.http.address 192.168.1.200:50090
- dfs.datanode.http.address 192.168.1.201:50075
- yarn.resourcemanager.webapp.address 192.168.1.200:50030
- mapreduce.jobhistory.webapp.address 192.168.1.200:19888。这个好像访问不了。需要启动hadoop/sbin/mr-jobhistory-daemon.sh start historyserver才可以访问。
三. Hadoop2.x eclispe-plugin
https://github.com/winghc/hadoop2x-eclipse-plugin
目前这个插件还在开发中，你可以自己下载源码编译，网上也有人发布编译好的jar。
http://blog.csdn.net/zythy/article/details/17397153 这个同学写的很详细了。
需要注意一点的是，Hadoop installation directory里填写Win下的hadoop home地址，其目的在于创建MapReduce Project能从这个地方自动引入MapReduce需要的jar。解压hadoop-2.2.0.tar.gz到本地即可。
四. 各种问题
上面一步完成后，创建一个MapReduce Project，运行时发现出问题了。
-
java.io.IOException: Could not locate executable null\bin\winutils.exe in the Hadoop binaries.
跟代码就去发现是HADOOP_HOME的问题。如果HADOOP_HOME为空，必然fullExeName为null\bin\winutils.exe。解决方法很简单啦，乖乖的配置环境变量吧，不想重启电脑可以在MapReduce程序里加上System.setProperty("hadoop.home.dir", "...");暂时缓缓。org.apache.hadoop.util.Shell.javapublic static final String getQualifiedBinPath(String executable) throws IOException { // construct hadoop bin path to the specified executable String fullExeName = HADOOP_HOME_DIR + File.separator + "bin" + File.separator + executable; File exeFile = new File(fullExeName); if (!exeFile.exists()) { throw new IOException("Could not locate executable " + fullExeName + " in the Hadoop binaries."); } return exeFile.getCanonicalPath(); } private static String HADOOP_HOME_DIR = checkHadoopHome(); private static String checkHadoopHome() { // first check the Dflag hadoop.home.dir with JVM scope String home = System.getProperty("hadoop.home.dir"); // fall back to the system/user-global env variable if (home == null) { home = System.getenv("HADOOP_HOME"); } ... } - 这个时候得到完整的地址fullExeName，我机器上是D:\Hadoop\tar\hadoop-2.2.0\hadoop-2.2.0\bin\winutils.exe。继续执行代码又发现了错误
Could not locate executable D:\Hadoop\tar\hadoop-2.2.0\hadoop-2.2.0\bin\winutils.exe in the Hadoop binaries.
就去一看，没有winutils.exe这个东西。去https://github.com/srccodes/hadoop-common-2.2.0-bin下载一个，放就去即可。 - 继续出问题
at org.apache.hadoop.util.Shell.execCommand(Shell.java:661) at org.apache.hadoop.fs.RawLocalFileSystem.setPermission(RawLocalFileSystem.java:639) at org.apache.hadoop.fs.RawLocalFileSystem.mkdirs(RawLocalFileSystem.java:435)
继续跟代码org.apache.hadoop.util.Shell.javapublic static String[] getSetPermissionCommand(String perm, boolean recursive, String file) { String[] baseCmd = getSetPermissionCommand(perm, recursive); String[] cmdWithFile = Arrays.copyOf(baseCmd, baseCmd.length + 1); cmdWithFile[cmdWithFile.length - 1] = file; return cmdWithFile; } /** Return a command to set permission */ public static String[] getSetPermissionCommand(String perm, boolean recursive) { if (recursive) { return (WINDOWS) ? new String[] { WINUTILS, "chmod", "-R", perm } : new String[] { "chmod", "-R", perm }; } else { return (WINDOWS) ? new String[] { WINUTILS, "chmod", perm } : new String[] { "chmod", perm }; } }cmdWithFile数组的内容为{"D:\Hadoop\tar\hadoop-2.2.0\hadoop-2.2.0\bin\winutils.exe", "chmod", "755", "xxxfile"}，我把这个单独在cmd里执行了一下，发现无法启动此程序，因为计算机中丢失 MSVCR100.dll
那就下载一个呗http://files.cnblogs.com/sirkevin/msvcr100.rar，丢到C:\Windows\System32里面。再次cmd执行，又来了问题应用程序无法正常启动(0xc000007b)
下载http://blog.csdn.net/vbcom/article/details/7245186 DirectX_Repair来解决这个问题吧。记得修复完后要重启电脑。搞定后cmd试一下，很棒。 - 到了这里，已经看到曙光了，但问题又来了
Exception in thread "main" java.lang.UnsatisfiedLinkError: org.apache.hadoop.io.nativeio.NativeIO$Windows.access0(Ljava/lang/String;I)Z
代码就去/** Windows only method used to check if the current process has requested * access rights on the given path. */ private static native boolean access0(String path, int requestedAccess);显然缺少dll文件，还记得https://github.com/srccodes/hadoop-common-2.2.0-bin下载的东西吧，里面就有hadoop.dll，最好的方法就是用hadoop-common-2.2.0-bin-master/bin目录替换本地hadoop的bin目录，并在环境变量里配置PATH=HADOOP_HOME/bin，重启电脑。 - 终于看到了MapReduce的正确输出output99。
五. 总结
- hadoop eclipse插件不是必须的，其作用在我看来就是如下三点(这个是一个错误的认识，具体请参考http://zy19982004.iteye.com/blog/2031172)。study-hadoop是一个普通project，直接运行(不通过Run on Hadoop这只大象)，一样可以调试到MapReduce。
- 对hadoop中的文件可视化。
- 创建MapReduce Project时帮你引入依赖的jar。
- Configuration conf = new Configuration();时就已经包含了所有的配置信息。
- 还是自己下载hadoop2.2的源码编译好，应该是不会有任何问题的(没有亲测)。
六. 其它问题
- 还是
Exception in thread "main" java.lang.UnsatisfiedLinkError: org.apache.hadoop.io.nativeio.NativeIO$Windows.access0(Ljava/lang/String;I)Z
代码跟到org.apache.hadoop.util.NativeCodeLoader.java去看static { // Try to load native hadoop library and set fallback flag appropriately if(LOG.isDebugEnabled()) { LOG.debug("Trying to load the custom-built native-hadoop library..."); } try { System.loadLibrary("hadoop"); LOG.debug("Loaded the native-hadoop library"); nativeCodeLoaded = true; } catch (Throwable t) { // Ignore failure to load if(LOG.isDebugEnabled()) { LOG.debug("Failed to load native-hadoop with error: " + t); LOG.debug("java.library.path=" + System.getProperty("java.library.path")); } } if (!nativeCodeLoaded) { LOG.warn("Unable to load native-hadoop library for your platform... " + "using builtin-java classes where applicable"); } }这里报错如下DEBUG org.apache.hadoop.util.NativeCodeLoader - Failed to load native-hadoop with error: java.lang.UnsatisfiedLinkError: HADOOP_HOME\bin\hadoop.dll: Can't load AMD 64-bit .dll on a IA 32-bit platform
怀疑是32位jdk的问题，替换成64位后，没问题了2014-03-11 19:43:08,805 DEBUG org.apache.hadoop.util.NativeCodeLoader - Trying to load the custom-built native-hadoop library... 2014-03-11 19:43:08,812 DEBUG org.apache.hadoop.util.NativeCodeLoader - Loaded the native-hadoop library
这里也解决了一个常见的警告WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
- http://www.cnblogs.com/xia520pi/archive/2012/05/16/2503949.html 很详细的介绍了Hadoop Hdfs的安装。它这个不是Hadoop2.2的，但配置都大同小异。
- MapReduce的配置可以参考http://blog.sina.com.cn/s/blog_546abd9f0101i8b8.html。
- dfs.http.address 192.168.1.200:50070
- dfs.secondary.http.address 192.168.1.200:50090
- dfs.datanode.http.address 192.168.1.201:50075
- yarn.resourcemanager.webapp.address 192.168.1.200:50030
- mapreduce.jobhistory.webapp.address 192.168.1.200:19888。这个好像访问不了。需要启动hadoop/sbin/mr-jobhistory-daemon.sh start historyserver才可以访问。
-
java.io.IOException: Could not locate executable null\bin\winutils.exe in the Hadoop binaries.
跟代码就去发现是HADOOP_HOME的问题。如果HADOOP_HOME为空，必然fullExeName为null\bin\winutils.exe。解决方法很简单啦，乖乖的配置环境变量吧，不想重启电脑可以在MapReduce程序里加上System.setProperty("hadoop.home.dir", "...");暂时缓缓。org.apache.hadoop.util.Shell.javapublic static final String getQualifiedBinPath(String executable) throws IOException { // construct hadoop bin path to the specified executable String fullExeName = HADOOP_HOME_DIR + File.separator + "bin" + File.separator + executable; File exeFile = new File(fullExeName); if (!exeFile.exists()) { throw new IOException("Could not locate executable " + fullExeName + " in the Hadoop binaries."); } return exeFile.getCanonicalPath(); } private static String HADOOP_HOME_DIR = checkHadoopHome(); private static String checkHadoopHome() { // first check the Dflag hadoop.home.dir with JVM scope String home = System.getProperty("hadoop.home.dir"); // fall back to the system/user-global env variable if (home == null) { home = System.getenv("HADOOP_HOME"); } ... } - 这个时候得到完整的地址fullExeName，我机器上是D:\Hadoop\tar\hadoop-2.2.0\hadoop-2.2.0\bin\winutils.exe。继续执行代码又发现了错误
Could not locate executable D:\Hadoop\tar\hadoop-2.2.0\hadoop-2.2.0\bin\winutils.exe in the Hadoop binaries.
就去一看，没有winutils.exe这个东西。去https://github.com/srccodes/hadoop-common-2.2.0-bin下载一个，放就去即可。 - 继续出问题
at org.apache.hadoop.util.Shell.execCommand(Shell.java:661) at org.apache.hadoop.fs.RawLocalFileSystem.setPermission(RawLocalFileSystem.java:639) at org.apache.hadoop.fs.RawLocalFileSystem.mkdirs(RawLocalFileSystem.java:435)
继续跟代码org.apache.hadoop.util.Shell.javapublic static String[] getSetPermissionCommand(String perm, boolean recursive, String file) { String[] baseCmd = getSetPermissionCommand(perm, recursive); String[] cmdWithFile = Arrays.copyOf(baseCmd, baseCmd.length + 1); cmdWithFile[cmdWithFile.length - 1] = file; return cmdWithFile; } /** Return a command to set permission */ public static String[] getSetPermissionCommand(String perm, boolean recursive) { if (recursive) { return (WINDOWS) ? new String[] { WINUTILS, "chmod", "-R", perm } : new String[] { "chmod", "-R", perm }; } else { return (WINDOWS) ? new String[] { WINUTILS, "chmod", perm } : new String[] { "chmod", perm }; } }cmdWithFile数组的内容为{"D:\Hadoop\tar\hadoop-2.2.0\hadoop-2.2.0\bin\winutils.exe", "chmod", "755", "xxxfile"}，我把这个单独在cmd里执行了一下，发现无法启动此程序，因为计算机中丢失 MSVCR100.dll
那就下载一个呗http://files.cnblogs.com/sirkevin/msvcr100.rar，丢到C:\Windows\System32里面。再次cmd执行，又来了问题应用程序无法正常启动(0xc000007b)
下载http://blog.csdn.net/vbcom/article/details/7245186 DirectX_Repair来解决这个问题吧。记得修复完后要重启电脑。搞定后cmd试一下，很棒。 - 到了这里，已经看到曙光了，但问题又来了
Exception in thread "main" java.lang.UnsatisfiedLinkError: org.apache.hadoop.io.nativeio.NativeIO$Windows.access0(Ljava/lang/String;I)Z
代码就去/** Windows only method used to check if the current process has requested * access rights on the given path. */ private static native boolean access0(String path, int requestedAccess);显然缺少dll文件，还记得https://github.com/srccodes/hadoop-common-2.2.0-bin下载的东西吧，里面就有hadoop.dll，最好的方法就是用hadoop-common-2.2.0-bin-master/bin目录替换本地hadoop的bin目录，并在环境变量里配置PATH=HADOOP_HOME/bin，重启电脑。 - 终于看到了MapReduce的正确输出output99。