Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
MapReduce, mail # user - FSDataInputStream: setReadAhead, setDropBehind


Copy link to this message
-
FSDataInputStream: setReadAhead, setDropBehind
John Lilley 2013-11-13, 14:22
I'm assuming that these methods exist to improve performance of certain I/O patterns, but the API doc doesn't explain when one would use them.  Our software's I/O pattern is mostly sequential I/O in large buffers (e.g. 256k).  Would one set these differently for local block I/O vs remote block I/O?
Thanks
John