mapreduce - why close() method be synchronized while extending RecordReader in Hadoop -


I'm considering the example of a hoop-book of alexhomes

I do not understand Was that why the closest ways to synchronize? And why only closed () and no other method?

  IOException throws a non-public synchronized {} reader.close (); }   

Can anyone explain me? It will be very helpful.

As I mentioned in the book, I used the class of the project on my basis For, as I wanted to write one that did not require LZOP compression, ElephantBird used a syncing closed method in its RecordReader, which resulted in my way in my version as the other mentioned that there is no reason Not that I'm MapReduce I know about syncing the closest method.

Comments