Hi,
Currently the barrier between r(i) and m(i+1) is the Job barrier.
That is, m(i+1) will be blocked until all r(i) finish (until Job i finish).
I'm saying this blocking is not necessary if we can concatenate them
all in a single Job as an endless chain.
Therefore m(i+1) can start immediately ev
Hi,
>>m1 | r1 m2 | r2 m3 | ... | r(K-1) mK | rK m(K+1)
My understanding is it would be something like:
m1|(r1 m2)| m(identity) | r2, if you combine the r(i) and m(i+1), because of
the hard distinction between Rs & Ms.
Amogh
On 2/4/10 1:46 PM, "Felix Halim" wrote:
Talking about barrier, curren
Hi,
AFAIK no. I'm not sure how much of a task it is to write a HOD-like scheduler,
or if its even feasible given the new architecture of single managing JT,
directly talking to TT. Probably someone more familiar with the scheduler
architecture can help you better.
What I was trying to suggest wi
Hi,
Yes the same location is populated with different values ( returned by
iter.next() ) for optimization reasons. There is a new patch which will allow
you to mark() and reset() iterator so that you buffer required values (
equivalently you can do that yourself, its anyways in-mem for the patch
Yeah, my understanding is that the iterator is just giving you a
pointer to the same location each time. This seems to match up with
the behavior we've both observed, but maybe someone more familiar with
the internals can verify.
Also, in case you didn't know, you can use what's called a secondary
Thanks, Ed. I'm copying the values into a list and then sorting them and
then emiting the top 20, so yes they are buffered. I'll try cloning each
item tomorrow and see if that works.
Does this mean the Iterator is returning the same pointer with each call to
next() but with different contents bein
Hi James,
I ran into something similar in the past and suspect the problem may
be in your reduce function. Are you buffering values from the
iterator? If you are, then you need to first clone the value when
taking it from the iterator (implement Cloneable in your custom
Writable). Otherwise they w
Hi,
For a particular project I created a writable for holding a long and a
double called LongDoublePair. My mapper outputs LongDoublePair values and
the reducer receives an Iterable.
The problem is that when I try to use it, whilst I get the right number of
elements in the Iterable, they are all