Running Combiners in out of core

[Ameya Vilankar <am...@gmail.com>]

Hi there. I am implementing my own algorithm in Giraph using the Latest features from the trunk branch. For every iteration I am setting the Computation class and the combiner class. But the problem is that in one of the iterations, the number of messages sent are too big. So i am using out of core messages by passing the following parameter
-D giraph.useOutOfCoreMessages=true

I have the following questions:
1. Can we use the following parameters:
-D giraph.useOutOfCoreMessages=true and -D giraph.useBigDataIOForMessages=true at the same time.

2. Also whenever I use the -D giraph.useOutOfCoreMessages=true, the combiners don't run.
I am using MasterCompute class to set the combiner and the computation type for every iteration since I am using different in message and out message in every super step. But the combiners don't run even if I set them on the master compute when i use out of core messages.

Please Help me out.
Thanks in advance.
Ameya Vilankar
Software Engineer Intern
Zynga