Implement the core basics of the per node controller dataset partition manager and writer.
The manager coordinates the whole result distribution process from registering
the result generating node controller with the directory service to creating
the dataset writers and readers as required. There will be only one manager
for the whole node controller and will be started when the node controller
starts but there will be one writer per result writer in a job.
git-svn-id: https://hyracks.googlecode.com/svn/branches/fullstack_hyracks_result_distribution@2469 123451ca-8445-de46-9d55-352943316053
2 files changed