[EMAIL PROTECTED] wrote:
As far as we understood from MapRed documentation all reduce tasks must be
launched after last map task is finished e.g map and reduce must not work
simultaneously. But often in logs we see such records: map 80%, reduce 10%
and many more records where map is less then
how would I setup mapred for smp machines, I understand it will split up big
jobs like indexing or updating the db into a bunch of chunks to be processed
by separate machines, I have machines that are multiple processor machines
that I want to test this with internally, makes sense to utilize the