Skip to content

Err chatbot plugin to run Ansible commainds and playbooks

License

Notifications You must be signed in to change notification settings

br0ziliy/err-ansible

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

48 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

err-ansible

Err plugin to run Ansible playbooks/commands asynchronously

Prerequisites

Plugin uses RQ library for jobs/queues management (this way the whole bot is not waiting for a particular long-running playbook to finish), which requires Redis running on the same machine where your bot runs.

Plugin will still work without Redis installed, but this way you will lose asynchronous capabilities, so having Redis is still highly recommended for production use.

Plugin is known to work on following platforms/Err backends:

  • CentOS 7.2 / Telegram

Installation

!repos install https://github.com/br0ziliy/err-ansible.git

You will also need to run the RQ worker in the directory where Errbot installed the plugin (/var/lib/err/plugins/err-ansible by default):

cd /var/lib/err/plugins/err-ansible
bin/start_worker.sh

Couple of considerations:

  • user account under which you run RQ worker should be able to execute ansible-playbook command, and should have read access to the directories/keyfile you supply in the plugin configuration dictionary (see below);
  • running worker from a different user than you run Errbot itself might lead to an unpredictalble results - i.e. if you run Errbot process under chat_bot account, you MUST run the worker under the same chat_bot, otherwise things might not work. This will be fixed in the future, see br0ziliy/err-ansbile#3.
  • since Ansible not yet supporting Python 3 - RQ worker should be running in the Python 2 environment. This also mean you have to ensure rq Python module is installed in that environment.

Example deployment

Follow usual installation instructions for the bot (for CentOS/RHEL you should install Python 3 from Software Collections first).

Then, create Python 2 virtualenv, install rq and (optionally) rq-dashboard modules, activate the Python 2 virtualenv and run the worker from that environment.

This in fact is how I use this plugin.

Configuration

Plugin is configured with the commands below:

  • initial configuration:

!plugin config Ansible {'ANSIBLE_SSH_KEY': u'/home/devops_bot/.ssh/id_rsa', 'INVENTORY_DIR': u'/etc/ansible/inventory', 'PLAYBOOK_DIR': u'/etc/ansible/playbooks', 'ANSIBLE_REMOTE_USER': 'root'}

  • if you want to change some parameter later:

!plugin config Ansible {'ANSIBLE_REMOTE_USER': 'deploy_user'}

Available parameters:

  • ANSIBLE_SSH_KEY - SSH key to use with ansible-playbook command
  • ANSIBLE_REMOTE_USER - SSH user Ansible will use to authenticate to the machines.
  • INVENTORY_DIR - directory where inventory files are stored (subdirectories are not yet supported, see br0ziliy/err-ansbile#2)
  • PLAYBOOK_DIR - directory where playbook files are stored (subdirectories are not yet supported, see br0ziliy/err-ansbile#2)

Usage

To run a playbook:

!ansible playbook.yml my_awesome_servers

To list available inventories/playbooks:

!ansible list

To query task status:

!task info [task UUID]

If you want a nice overview of RQ tasks queue, install rq-dashboard Python package, and run it:

pip install rq-dashboard
rq-dashboard

Licence

GNU GPLv3. See LICENSE file for details.

About

Err chatbot plugin to run Ansible commainds and playbooks

Resources

License

Stars

Watchers

Forks

Packages

No packages published