An OpenStack Orchestration Engine

• Use the monster installer at https://github.com/rcbops-qa/monster-installer.git

• Clone the repo and install

  git clone https://github.com/rcbops-qa/monster.git ~/monster
  virtualenv -p `which python2` ~/monster/.venv
  source ~/monster/.venv/bin/activate
  pip install -r ~/monster/requirements.txt
      

  note: If you’re on a small server I’ve experienced failed installs of gevent because there isn’t enough memory. Adding swap fixed things.

• put your credentials in secret.yaml at the top level of monster

  rackspace:
    user: YOUR_USERNAME
    api_key: YOUR_API_KEY
    auth_url: https://identity.api.rackspacecloud.com/v2.0/
    region: dfw
    plugin: rackspace
      

  if you use cloudfiles your user and password can also be placed in the secret.yaml

  cloudfiles:
    user: YOUR_USER
    password: YOUR_PASSWORD
      

  If you use razor to provision servers add its API endpoint support clean up

  razor:
    url: RAZOR_IP
      

  You will also need to provide the default password for the cluster nodes

  default_pass: DEFAULT_PASSWORD
      

compute.py and storage.py are the CLIs

Deploys an OpenStack cluster

compute.py build -n ubuntu-default --branch master --template ubuntu-default --config config.yaml --secret-path secret.yaml --log log.txt --log-level DEBUG --provisioner rackspace

• name: name of the deployment
• branch: branch rcbops cookbooks
• template: name of template to use
• config: path to config file
• secret-path: path to secret config file
• dry: only build environment (will provision in cloud)
• log: path to log file
• log-level: level to log - INFO, DEBUG
• provisioner: provisioner to use - razor, rackspace, openstack

Shows details about an OpenStack deployment

compute.py show -n ubuntu-default

• name: name of the deployment
• config: path to config file
• secret-path: path to secret config file
• log: path to log file
• log-level: level to log - INFO, DEBUG

Destroys an OpenStack deployment

compute.py destroy -n ubuntu-default --config config.yaml --log log.txt --log-level DEBUG

• name: name of the deployment
• config: path to config file
• secret-path: path to secret config file
• log: path to log file
• log-level: level to log - INFO, DEBUG

Load openrc environment variables into shell. After loaded openstack cli commands will communicate to cluster.

compute.py openrc -n precise-default

nova boot --image cirros-image --flavor 1

• name: name of the deployment
• config: path to config file
• secret-path: path to secret config file
• log: path to log file
• log-level: level to log - INFO, DEBUG

Attempts to open Horizon in browser

compute.py horizon -n ubuntu-default

• name: name of the deployment
• config: path to config file
• secret-path: path to secret config file
• log: path to log file
• log-level: level to log - INFO, DEBUG

Runs tests on a running OpenStack cluster

compute.py test --ha --tempest -n ubuntu-default --config config.yaml --log log.txt --log-level DEBUG --iterations 3

• ha: enables high availability testing
• tempest: enables the tempest testing suite
• name: name of the deployment
• config: path to config file
• secret-path: path to secret config file
• log: path to log file
• log-level: level to log - INFO, DEBUG
• iterations: number of full test runs

Upgrades the deployment to the specified branch

compute.py upgrade -n ubuntu-default -u v4.2.1 --log-level DEBUG

• name: name of the deployment
• upgrade-branch: branch to upgrade to
• config: path to config file
• secret-path: path to secret config file
• log: path to log file
• log-level: level to log - INFO, DEBUG

Opens an new tmux session with each node in a different window

compute.py tmux -n ubuntu-default

Requires tmux version >= 1.8

To add a 12.4 precise tmux 1.8 backport PPA, execute the following:

add-apt-repository -y ppa:kalakris/tmux
apt-get update
apt-get install tmux -y

• name: name of the deployment
• config: path to config file
• secret-path: path to secret config file
• log: path to log file
• log-level: level to log - INFO, DEBUG

To make development of monster easier you can load deployments using iPython and mess around with them:

1. Start ipython in top monster directory
2. Run

   from tools.ipython import load
   deployment = load("yourdep")
       

3. ????
4. profit!!!