* commit '39497051':
  fix windows tests
  get last for broker
  getlast for worker

* commit '83cb6cc8':
  correct behaviour in case of broker restart

* commit '14361e5e': (23 commits)
  change receiver health-check to ps
  change some of producer and reciever messages from info to debug loglevel
  allow override files
  remove journal write concern from broker
  use newer mgo package
  getnext returns on error record number if succedded to update it
  journal write concernt a mongo restart test for broker
  use mongo without nomad for restart test
  journal write concern for receiver's mongo client, test to recover after mongo restart
  add workers to nomad job
  revert to single mongo instance
  sharded mongo
  started mongo sharding
  more nomad jobs, start logging
  dockerimages for go services
  always pull receiver image
  add receivers, Dockerfile for receiver
  fix curl
  fix curl
  more nomad jobs,more static libs, Dockerfile for receiver
  ...