1. if you want modify the defalut params, go to params.py
2. two server:namenode and task server, in practice, namenode is persistent progress,and task server is for dispatch a task
3. if you want to deal with a rather big file in this framework, you should do like this, refer to upload.py client = dfsclient.DfsClient() client.upload('local file') after successful upload you can use it in next as s.dfsfile='your file'
4. you can modify the reducer like(task.reducer_num = n), default 1
5. yet you can deal memory data like: s.datasource = your datasource
6. have fun with it

robin1001 2013-8-13