This guide will walk you through installing pstack
on your machine.
There are a couple of things to take note of before proceeding with the installation.
If you're a member of the POETS project and would
like to use pstack
you can SSH into and use the pstack
deployment on
Coleridge. This does not require setting up anything on your
own computer so it's a very easy way to get started.
This repository contains the three building blocks of pstack
:
psim
: an independent simulator and a compatible enginepcli
: command-line user terminalpd
: the daemon process
All three will become available once you clone this repo but you may only want to use one or two, depending on your needs. Here are possible use cases:
- You want to run simple simulations of local XML files (single-threaded and non-distributed).
- You want to run distributed or multi-threaded simultions on an existing
pstack
deployment. - You want to add your computer as an engine to an existing
pstack
deployment. - You want to deploy
pstack
and make it available to other uses.
For (1), it will be sufficient to use psim
as an independent tool.
(2) and (3) require that you use pcli
and pd
(respectively) while
forwarding local Redis traffic over an SSH tunnel to a remote pstack
deployment as explained here.
(4) requires installing Redis and optionally setting up pd
on one or several
machines.
When setting up Python projects, it's standard practice to install
requirements.txt
dependencies locally within the project. It's therefore
strongly recommended that you use
virtualenv
for step 2.
You may want to place symlinks to the wrappers in bin
in your /usr/bin
to
make pcli
, pd
and psim
available globally. Note that the wrapper require
a virtualenv
environment to be setup up in the repository directory.
If you're deploying pstack
as a service to other users (use case 4), you'll
also need to install Redis. Redis is available as an
Ubuntu package and can also be instantiated as a docker container.