Replace the per job dataset directory service with a global directory service.

With this commit there will be only one directory service for entire Hyracks
and will be started when the cluster controller is started and all the jobs
share this directory service.

git-svn-id: https://hyracks.googlecode.com/svn/branches/fullstack_hyracks_result_distribution@2481 123451ca-8445-de46-9d55-352943316053
8 files changed
tree: d821eb3d01444e1a126794ef676058c6968802fa
  1. algebricks/
  2. hivesterix/
  3. hyracks/
  4. pom.xml
  5. pregelix/