Hadoop streaming example:
$ {Hadoop_bin} streaming \ r
-D mapred. Job. Name =$ {job_name} # Task Name
-D stream. Memory. Limit = 1500 # Task memory limit
-D mapred. Map. Capacity. Per. tasktracker = 1 \ r
-D mapred. Reduce. Capacity. Per. tasktracker = 1 \ r
-D mapred. Map. Tasks =$ {map_tasks} # map count
-D mapred. Job. Map. Capacity =$ {map_capacity} # map capacity, which is generally consistent with the number of maps
-D mapred. Reduce. Tasks =$ {red_tasks} # reduce count
-D mapred. Job. Reduce. Capacity =$ {red_capacity} # reduce capacity, which is generally consistent with reduce capacity
-D mapred. Job. Priority =$ {Mapred_priority} # task priority
-D mapred. userlog. Retain. Hours =$ {log_retain_hours} \ r
-D stream. Num. Map. Output. Key. Fields = 3 # The map results are sorted by the first three columns.
-D num. Key. Fields. For . Partition = 1 # The same data is allocated to the same CER In the first column.
-Partitioner org. Apache. hadoop. mapred. Lib. keyfieldbasedpartitioner \ r
-Input/log/1079/cpro_pblog_noah/20110814/*/PB. log * # Input
-Input/log/1079/cpro_pblog_noah/20110815/*/PB. log * # Input
-Input/log/3148/shifen_bdclk_noah/20110814/*/dcharge. BD. *. log * # Input
-Input/log/3148/shifen_bdclk_noah/20110815/*/dcharge. BD. *. log * # Input
-Output $ {performance_out} # output
-Mapper "java6/bin/Java-classpath ad_trade com. Baidu. cm. UFS. mapper testno.txt" # mapper Program
-Reducer "java6/bin/Java-classpath ad_trade com. Baidu. cm. UFS. reducer" # CER Program
-File ad_trade # file to be uploaded and distributed
-File testno.txt \ r
-Cachearchive/APP/ECOM/CM/NOVA. ufs/U-Wangyou/java6.tar.gz # java6 # The compressed package to be distributed on HDFS. The decompressed folder name is java6.