final refactoring so that all vcell services can be run locally from an IDE as:
individual java services which can be debugged and use local postgres and mongo and remote activemq
slurm simulation jobs run and send status normally (via the remote activemqsim server running on vcell-node3)
slurm optimization jobs run normally (status will be sent through file system when it is resurrected).
docker-compose (similar to swarm mode) running locally with local postgres - and both mongo and activemq running as containers within same stack). This doesn't receive status messages from slurm simulations
slurm simulation jobs run normally but don't send status (hpc cluster can't talk to local activemqsim running in stack on local dev machine)
slurm optimization jobs run normally (status will be sent through file system when it is resurrected).
For production: docker swarm mode will be tested on vcelltest site after merge.
fix swarm config. latest Alpha deploy is broken.