A 5-minute introduction to pylm-registry

What is the pylm-registry

Assume that you want to run a large PALM cluster, or a bunch of process running in different servers scattered around the world. You may may need some kind of tool to run them in an orchestrated fashion and to centralize logs somehow. You’d probably heard about Kubernetes, Fleet or other similar products. Pylm-registry, or the registry is a pretty simple and hackable tool that can be used to cover the basic needs of running a cluster of servers in a sane way

Important

The registry has been designed with PALM (or pylm) in mind, but you can use it to monitor any process. Many examples of this documentation have nothing to do with PALM. But try PALM, you’ll love it.

Running the whole thing

The first thing you have to do is to install the software.

$> pip install pylm-registry

Important

Pylm, and in consequence the registry, requires a version of Python equal or higher than 3.4, and it is more thoroughly tested with Python 3.5.

That was easy. This step should have installed two programs, the registry and he runner, which is a kind of client.

The registry server

The registry server is a web service that is used to configure and monitor a cluster. It is a complete application with user management, a database, and so on.

To run a local instance of the registry for testing:

.. code-block:: bash

$> pylm-registry –sync

This command starts the registry with volatile storage. This means that the configuration and state of every cluster will be lost once you shut the registry down