Saltstack Example

Trivial Master-Minion Vagrant set up example

Requirements

Developed with:

VirtualBox 5.0.14
Vagrant 1.8.1

Prior versions might work, but are untested.

What is this?

This is a very simple example of a Saltstack master/minion configuration automated with Vagrant, with a Virtualbox provider.

The Salt master machine is called saltmaster, and there is a minion example called webserver: an nginx serving a very simple html page.

Nginx monitoring

In addition, the nginx state uses HttpStubStatusModule for monitoring, in http://192.168.100.3:8080. Requests are restricted to vagrant's subnet (your host should have access to this subnet already). You can browse open, accepted and handled connections, and handled requests. This information will be employed by the custom SaltStack beacon.

How to run

Start the virtual machines

$ vagrant up saltmaster webserver

They should have been automatically provisioned, if you want to reprovision the webserver, just type

$ vagrant provision webserver

Now, just go to a browser and type the webserver IP: http://192.168.100.3/

You should see a very basic webpage provisioned by Saltmaster into the webserver minion.

Custom nginx_connections beacon

The minion config declares a custom nginx_connections beacon (developed under roots/salt/_beacons/nginx_connections.py), which sends nginx connection stats (see ngx_http_stub_status_module) every 10 seconds. This beacon just parses the output given by http://192.168.100.3:8080.

It is possible to see those events in saltmaster running in debug mode:

$ sudo su

$ service salt-master stop

$ salt-master -l debug

The webserver minion will send them periodically:

[DEBUG ] Sending event - data = {'_stamp': '2016-02-22T11:40:52.006223', 'tag': 'salt/beacon/webserver.dev.darioblanco.com/nginx_connections/', 'data': {'accepted': 214, 'handled': 214, 'writing': 1, 'waiting': 0, 'requests': 217, 'reading': 0, 'open': 1, 'id': 'webserver.dev.darioblanco.com'}}

Testing

You can test the nginx_connections beacon with pytest.

First, install the dependencies inside a virtualenv

$ pip install -r requirements -r dev-requirements

Run py.test, it will discover the test case automatically.

$ py.test