hadoop 入门 启动任务

定义框架接口 
由具体实现类实现 

1. public interface Tool extends
2. int run(String [] args) throws
3. }

ToolRunner 

同一的入口调用 

按配置解析参数，调用接口方法 

1. public static int
2. throws
3. if(conf == null) {  
4. new
5.   }  
6. new
7. //set the configuration back, so that Tool can configure itself
8.   tool.setConf(conf);  
9.     
10. //get the args w/o generic hadoop args
11.   String[] toolArgs = parser.getRemainingArgs();  
12. return
13. }

Mahout 中具体调用示例 

1. public static void main(String[] args) throws
2. new Configuration(), new
3. }

覆盖方法,提取参数，调用核心方法 

1. @Override
2. public int run(String[] args) throws
3.   addInputOption();  
4.   addOutputOption();  
5. //...........
6.   runJob(input,  
7.          output,  
8.          minClusterSize,  
9.          minVectorSize,  
10.          hashType,  
11.          numHashFunctions,  
12.          keyGroups,  
13.          numReduceTasks,  
14.          debugOutput);  
15. return 0;  
16. }

核心方法，配置job，开始map reduce任务 

1. private void
2.                     Path output,  
3. int
4. int
5.                     String hashType,   
6. int
7. int
8. int
9. boolean debugOutput) throws
10.   Configuration conf = getConf();  
11.   
12. //配置参数设置........................    
13. new Job(conf, "MinHash Clustering");  
14. class);  
15.   
16. //Job参数设置.........................
17. true);  
18. }