业务背景

约定五天前的HDFS数据为过期版本号数据。写一个脚本自己主动删除过期版本号数据

$ hadoop fs -ls /user/pms/workspace/ouyangyewei/dataFound 9 itemsdrwxr-xr-x   - pms pms          0 2015-08-11 17:03 /user/pms/workspace/ouyangyewei/data/2015-08-01drwxr-xr-x   - pms pms          0 2015-08-11 17:03 /user/pms/workspace/ouyangyewei/data/2015-08-02drwxr-xr-x   - pms pms          0 2015-08-11 17:03 /user/pms/workspace/ouyangyewei/data/2015-08-03drwxr-xr-x   - pms pms          0 2015-08-11 17:03 /user/pms/workspace/ouyangyewei/data/2015-08-04drwxr-xr-x   - pms pms          0 2015-08-11 17:03 /user/pms/workspace/ouyangyewei/data/2015-08-05drwxr-xr-x   - pms pms          0 2015-08-11 17:03 /user/pms/workspace/ouyangyewei/data/2015-08-06drwxr-xr-x   - pms pms          0 2015-08-11 17:03 /user/pms/workspace/ouyangyewei/data/2015-08-07drwxr-xr-x   - pms pms          0 2015-08-11 17:03 /user/pms/workspace/ouyangyewei/data/2015-08-08drwxr-xr-x   - pms pms          0 2015-08-11 17:03 /user/pms/workspace/ouyangyewei/data/2015-08-09

脚本实现

# ---------------------------------------------------------## 删除历史版本号（五天前的为过期版本号数据）## ---------------------------------------------------------old_version=$(hadoop fs -ls /user/pms/workspace/ouyangyewei/data | awk 'BEGIN{ five_days_ago=strftime("%F", systime()-5*24*3600) }{ split($8,arr,"/"); if(arr[7]

运行以后

$ hadoop fs -ls /user/pms/workspace/ouyangyewei/dataFound 4 itemsdrwxr-xr-x   - pms pms          0 2015-08-11 17:03 /user/pms/workspace/ouyangyewei/data/2015-08-06drwxr-xr-x   - pms pms          0 2015-08-11 17:03 /user/pms/workspace/ouyangyewei/data/2015-08-07drwxr-xr-x   - pms pms          0 2015-08-11 17:03 /user/pms/workspace/ouyangyewei/data/2015-08-08drwxr-xr-x   - pms pms          0 2015-08-11 17:03 /user/pms/workspace/ouyangyewei/data/2015-08-09