Maestro is a Fabric based toolkit for managing systems. There is also integration with cloud providers via Apache Libcloud.

• pip install maestro

Either download the default fabfile or create your own:

from maestro.core import *
from maestro import system as sys

You can then run fab -l to see the available tasks.

For example, to run a memory report for a host named dev.example.com:

fab -H dev.example.com sys.memory

[dev.example.com] Executing task 'sys.memory'
[dev.example.com]              total       used       free     shared    buffers     cached
[dev.example.com] Mem:          7948       4076       3871          0       1942        295

Maestro has the ability to use a cloud provider (currently EC2 and Rackspace) for a node list.

You can optionally specify a specific region to only show those nodes. By default, all regions will be used.

To see a list of available regions, run:

fab list_available_providers

To use Amazon EC2, you must set the following environment variables:

EC2_ACCESS_ID=<your_aws_access_id>
EC2_SECRET_KEY=<your_aws_secret_key>

To run an uptime report for all nodes in EC2:

fab nodes:ec2 sys.uptime

[1.2.3.4] 07:26:16 up 49 days, 12:34,  1 user,  load average: 0.32, 0.05, 0.01
[4.5.6.7] 07:26:17 up 34 days, 13:15,  1 user,  load average: 0.22, 0.10, 0.01
[4.3.2.1] 07:26:18 up 23 days, 15:25,  1 user,  load average: 0.17, 0.06, 0.01

To run an uptime report for all nodes in EC2 us-east-1 region:

fab nodes:ec2,us-east-1 sys.uptime

[1.2.3.4] 07:26:16 up 49 days, 12:34,  1 user,  load average: 0.32, 0.05, 0.01

To use Rackspace, you must set the following environment variables:

RACKSPACE_ID=<username>
RACKSPACE_KEY=<api_key>

To run an uptime report for all nodes:

fab nodes:rackspace sys.uptime

[50.1.2.3] 00:16:37 up 2 min,  1 user,  load average: 0.00, 0.00, 0.00

You can also filter nodes based upon the machine name (the Name tag in EC2) using a regular expression:

fab nodes:ec2,filter=^puppet sys.uptime

[puppet.example.com] 23:17:19 up 50 days,  4:25,  1 user,  load average: 0.00, 0.00, 0.00

You can put the environment variables in key=value form in a file named .maestrorc in your home directory. Maestro will automatically load it and use the values.

For example, create ~/.maestrorc with the following:

EC2_ACCESS_ID=123456789
EC2_SECRET_KEY=abcdefghijlkmnop
RACKSPACE_ID=rackspaceuser
RACKSPACE_KEY=800012345abcdefg12345

To use Maestro from a custom fabfile simply add the following to your fabfile:

from maestro.core import *
from maestro import system as sys
# optional -- ~/.maestrorc loader
from maestro.utils import load_maestro_rc
load_maestro_rc()

You should then be able to run fab -l and see the Maestro tasks.

There is also experimental support for creating and deploying Python applications (more languages and frameworks will be included in the future).

** Note: this assumes you have the following packages installed (these are for debian/ubuntu):

• build-essential
• python-dev
• python-setuptools
• libxml2-dev
• libxslt-dev
• supervisor
• nginx

** You will also need to install uWSGI:

• sudo easy_install pip
• sudo pip install uwsgi
• sudo pip install virtualenv

fab -H 10.1.2.3 py.create_app:myapp,"myapp.mydomain.com"

This will create the application directory, setup supervisor to run uWSGI, and create an nginx config.

To deploy an application, make sure there is a wsgi.py module in the root of the application. Here is an example for a Django app:

import os
import django.core.handlers.wsgi
os.environ.setdefault("DJANGO_SETTINGS_MODULE", "myapp.settings")

application = django.core.handlers.wsgi.WSGIHandler()

fab -H 10.1.2.3 py.deploy:myapp,/path/to/myapp/source

You can also deploy using git:

fab -H 10.1.2.3 py.deploy:myapp,https://github.com/username/myapp

When deploying, Maestro will upload or pull the latest source, install packages for the virtualenv (from a requirements.txt located in the root directory of your app), and restart the application.

fab -H 10.1.2.3 py.delete_app:myapp

Forking encouraged :) For bugs, features, etc. please use Github issues.

• Many thanks to the Fabric and Apache Libcloud projects for making my job easier :)