hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From jason hadoop <jason.had...@gmail.com>
Subject Re: Need help understanding the source
Date Tue, 07 Jul 2009 08:13:46 GMT
If your constraints are loose enough, you could consider using the chain
mapping that became available in 19, and
have multiple mappers for your jobs.
The extra mappers only receive the output of the prior map in the chain and
if I remember correctly, the combiner is run at the end of the chain of
mappers, when a reduce is scheduled.

The other alternative you may try is simply to write your map outputs to
HDFS [ie: setNumReduces(0)], and have a consumer pick up the map outputs as
they appear. If the life of the files is short and you can withstand data
loss, you may turn down the replication factor, to speed the writes.

The Map/Reduce framework is carefully constructed to provide a completely
sorted input set to the reducer, as that is part of the fundamental

On Tue, Jul 7, 2009 at 12:30 AM, Amr Awadallah <aaa@cloudera.com> wrote:

> To add to Todd/Ted's wise words, the Hadoop (and MapReduce) architects
> didn't impose this limitation just for fun, it is very core to enabling
> Hadoop to be as reliable as it is. If the reducer starts processing mapper
> output immediately and a specific mapper fails then the reducer would have
> to know how to undo the specific pieces of work related to the failed
> mapper, not trivial at all. That said, the combiners do achieve a bit of
> that for you, as they start working immediately on the map out, but on a
> per-mapper basis (not global), so easy to handle failure in that case (you
> just redo that mapper and the combining for it).
> -- amr
> Ted Dunning wrote:
>> I would consider this to be a very delicate optimization with little
>> utility
>> in the real world.  It is very, very rare to reliably know how many
>> records
>> the reducer will see.  Getting this wrong would be a disaster.  Getting it
>> right would be very difficult in almost all cases.
>> Moreover, this assumption is baked all through the map-reduce design and
>> thus doing a change to allow reduce to go ahead is likely to be really
>> tricky (not that I know this for a fact).
>> On Mon, Jul 6, 2009 at 11:14 AM, Naresh Rapolu <
>> nareshreddy.rapolu@gmail.com
>>> wrote:
>>> My aim is to make the reduce move ahead with reduction as and when it
>>> gets
>>> the data required, instead of waiting for all the maps to complete.  If
>>> it
>>> knows how many records it needs and compares it with number of records it
>>> has got until now,  it can move on once they become equal without waiting
>>> for all the maps to finish.

Pro Hadoop, a book to guide you from beginner to hadoop mastery,
www.prohadoopbook.com a community for Hadoop Professionals

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message