-Re: How / when does On-disk merge work?
Ravi Prakash 2013-10-28, 14:15
Tom White's "Hadoop: The Definitive Guide" is probably the best source for information on this (apart from the code itself ;-) Look at MergeManagerImpl.java btw in case you are so inclined).
On Friday, October 25, 2013 2:36 PM, - <[EMAIL PROTECTED]> wrote:
Can anyone provide documentation regarding how on-disk merge on reduce phase works in detail in Hadoop 2.2.0?
There is an explanation in this page but I am afraid it could be outdated since what I observe in my log files is a bunch of "OnDiskMerger - Thread to merge on-disk map-outputs" work at the end of merge phase.